Downloaded via CENTRO DE INVESTIGACION Y DESARROLLO TECNOLOGICO EN ELECTROQUIMICA SC on November 17, 2024 at [Link] (UTC).

CHELATE COMPLEXES OF 1,10-PHENANTHROLINE AND

RELATED COMPOUNDS
WARREN W. BRANDT
Department of Chemistry, Purdue University, Lafayette, Indiana
AND

FRANCIS P. DWYER and ELEANORA C. GY ARFAS

See [Link] for options on how to legitimately share published articles.

Department of Chemistry, University of Sydney, Sydney, Australia

Received July 81, 1954

CONTENTS

I. Introduction............................................................... 960
II. Coordination compounds with metal ions.................................... 960
A. Group 1 metals.......................................................... 962
B. Group 2 metals.......................................................... 963
C. Group 3 metals.......................................................... 964
D. Group 4 metals.......................................................... 965
E. Group 5 metals.......................................................... 965
F. Group 6 metals.......................................................... 965
G. Group 7 metals.......................................................... 966
H. Group 8 metals.......................................................... 967
III. The polypyridines and their metal complexes................................ 971
A. 2,2',2"-Terpyridine...................................................... 971
B. 2,2',2",2"'-Tetrapyridine................................................ 974
C. Higher polypyridines.................................................... 975
IV. The nature of the bonds.................................................... 975
V. Chelate molecules similar to bipyridine...................................... 978
VI. Stability of metal-chelate complexes........................................ 981
A. The iron-phenanthroline system......................................... 981
B. The iron-bipyridine system.............................................. 986
C. The iron-terpyridine system............................................. 988
D. A comparison of the iron systems........................................ 989
E. Other metal systems..................................................... 990
VII. Substitution and steric considerations....................................... 991
VIII. Optical activity of chelate complexes........................................ 997
IX. Redox reactions............................................................ 1000
X. Analytical applications..................................................... 1004
A. As oxidation-reduction indicators........................................ 1004
B. As colorimetric reagents................................................. 1004
1. Iron.................................................................. 1004
2. Copper............................................................... 1005
3. Other metals.......................................................... 1006
C. As agents in precipitation................................................ 1006
D. Miscellaneous applications............................................... 1006
XI. Biological activity.......................................................... 1007
A. Effect of the organic reagents............................................ 1007
B. Effect of the metal complexes............................................ 1008
XII. References.................................................................. 1010
959
960 BRANDT, DWYER, AND GYARFAS

I. Introduction
The chemistry of coordination compounds has become a highly specialized
branch of the science. Several review articles and books have dealt with various
broad phases of the problem (12, 26, 27, 158, 192, 196, 220, 242, 276), but there
has been no correlated consideration of one family of reagents other than Bjer-
rum’s excellent monograph on the ammine complexes (26).
The extent of the recent studies in the chemistry of the chelate-forming mole-
cules containing the aromatic =N—C—C—N= grouping has made a review of
this work both desirable and necessary. The present work is an effort to present
the chemistry of a related series of organic reagents in their reactions with
metal ions. It is hoped that this critical evaluation will prompt correlations of
other systems and further investigation of the one described. Although investi-
gations now known to be in the process of completion or publication will make
some of these considerations out of date, they should provide present and future
investigators with a coordinated and critical background for interpretation.
The discussion includes the information published prior to January 1, 1954.
In sections dealing with observed reactions and stabilities the coverage is in-
tended to be essentially complete. The applications cited are merely representa-
tive of the types of analysis possible. No attempt has been made to include the
extensive literature in this section.
The terms 1,10-phenanthroline, 2,2'-bipyridine, and 2,2',2"-terpyridine
are the nomenclature recommended by Chemical Abstracts and the International
Union of Pure and Applied Chemistry. The prefix numbers will be omitted in
the text. The abbreviations “phen,” “bipy,” and “trpy” seem consistent with
the nomenclature and should not conflict with common abbreviations for other
organic reagents. The letter “B” will be used in the writing of equations to repre-
sent the general type of molecule being discussed. The trivial names “ferroin”
and “ferriin” are utilized to denote the tris-1,10-phenanthroline ferrous and
ferric complexes, respectively.

II. Coordination Compounds with Metal Ions

The reaction between iron(II) salts and 2,2'-bipyridine with the formation
of intensely red substances was first observed by Blau (29, 32), who isolated a
series of salts of the formula Fe(bipy)<X2 and demonstrated that the color re-
sided in the common cation. The analogously constituted 1,10-phenanthroline,
named -phenanthroline by Blau, who first synthesized the substance, was
shown to form similarly colored salts of the same type. When the red iron (II)
compounds were treated with chlorine or potassium permanganate in acid solu-
tion the color changed to blue; from the amount of reagent consumed it was
concluded that the blue substances contained tervalent iron. This was con-
firmed by the isolation of the chloroplatinates [Fe(bipy)3]2[PtCl6L· ·5.5H20 and
[Fe(phen)3J2[PtCle]3 18H20. These substances could not be prepared directly
·

from iron(III) salts and the parent bases, a reaction which it was felt should
give the true “ferri-salts”; hence they were designated as “ ferri-salts.” By
ingenious, if somewhat subjective, arguments, it was concluded that the red
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 961

ferro- and the blue ferri-salts could be assigned to the “lúteo” or hexammine
class in Werner’s new classification. According to the Werner theory (292) the
formation of the compounds and their reactions were simple and natural, but
by the then current Blomstrand-Jorgenson structural theory they were inex-
plicable. A number of complex compounds were prepared from copper(II),
nickel(II), and cobalt(II) salts, and the bases were stated to react also with
zinc, cadmium, and chromium(III) salts, but not with manganese, lead, or alu-
minum salts (32).
The resolution of the tris-2,2'-bipyridine iron(II) cation into the opti-
cally active forms was achieved by Werner (293) in 1912, and interest in these
chelate molecules then lapsed for nearly twenty years. It was revived by the
discovery of Hammett, Walden, and Chapman (129) that the iron(II) com-
pounds could be used as valuable redox indicators and by the preparative re-
searches of Morgan and Bur stall, Barbieri, Pfeiffer, and Jaeger.
It is now recognized that the bipyridine type of chelate molecule rivals in
versatility the well-known ethylenediamine, not only in the stability of many
of the metal derivatives and in their frequently intense colors, but also in the
number of metals that can be chelated.
With the exception of the rare earths, numerous examples are known from
each periodic group, including the alkali and alkaline earth elements. In the field
of biology, bipyridine and phenanthroline have been shown to be useful anthel-
mintics (7) and to have some, but not pronounced, bacteriocidal value (2).
Bipyridine (I) and phenanthroline (II) function as typical bidentate chelate
molecules through the nitrogen atoms with the formation of five-membered
rings.
5 e

I II
2,2 '-Bipyridine 1,10-Phenanthroline
Attachment of the isomeric bipyridines such as 2,3'-, 2,4'-, or 4,4'-bipyridine
to metal atoms would involve many-membered rings and these have been shown
(30, 298) not to form stable iron complexes, though they may well attach to
other metal atoms as monodentate ligands like pyridine itself. In the same way
the structurally isomeric 1,7- and 4,7-phenanthrolines1 do not act as chelate
molecules (268).
When crystallized from -water, phenanthroline and most of its substituted
derivatives form well-defined monohydrates with characteristic melting points
and low solubility in benzene. This property is of value in purification, since
the addition of a few drops of water to solutions of the anhydrous substances
in benzene precipitates the monohydrate. The molecule of water is hydrogen-
bonded to the nitrogen atoms (118).
1
Smith and Richter (268) have referred to confusion in the early literature on the nomen-
clature of these substances.
962 BRANDT, DWYER, AND GYARFAS

A hydrate of bipyridine has not been described; this is in accordance with

crystal structure (46) and dipole moment studies (115, 123), which show that
the rings are coplanar but in the trans position. However, it is certain that in
the metal complexes the pyridine rings are in the cis position, and as in phenan-
throline, which necessarily has the cis structure, coordination occurs in such a
way that the chelate ring containing the metal atom is coplanar with the rest
of the molecule.
Both chelate molecules have been attached to metal carbonyls and nitrosyls,
but in such adducts the properties remain substantially those of carbonyls or
nitrosyls. Therefore these compounds will not be discussed in this review.
A. GROUP 1 METALS

Coordination compounds of the alkali metals are relatively rare, as can be

anticipated from their large ionic sizes and small electronegativities. Usually
the attachment is to oxygen and non-electrolytes result, as when the alkali
metal salt of a diketone is treated with one molecule more of the diketone (38,
64, 252). Monophenanthroline lithium perchlorate and bisphenanthroline so-
dium perchlorate are reported to be preparable in methanol solution. The color-
less needles are soluble in hot water, nitrobenzene, and dioxane (228). Similarly,
solutions of sodium iodide in acetone yield crystalline precipitates on the addi-
tion of phenanthroline or bipyridine (132). No evidence has yet been advanced
to show that the base is actually chelated.
The colored copper® chelates were first reported by Tartarini (275), who
reduced mixtures of phenanthroline or bipyridine and various copper(II) salts
in ammoniacal solution with hydrazine, sodium bisulfite, or hydroxylamine. In
some instances they were prepared directly from the copper(I) salt and the base.
The following were reported: CuSCN-phen, Cul-2phen, CuSCN-bipy, and
Cul bipy. Subsequently, the red color developed by the reduction of copper(II)
·

salts in the presence of phenanthroline was recommended for the microcolori-

metric estimation of copper (213). Since 2-covalent copper® has a linear con-
figuration, it is likely that these substances are all 4-covalent. This view is sup-
ported by the isolation of red-brown salts (132) of the form [Cu(phen)2]C10*
and [Cu(bipy)2]ClC>4. These are conductors in nitrobenzene solution, as are
[CuX-B] (X = I- or SCN-; B =
phen or bipy). In this solvent spectrophoto-
metric measurements showed the presence of the common ion CuB2+; hence
they should be formulated as [CuB2][CuX2]. There was, however, some evidence
for a small amount of the bridged dimer (III) in equilibrium:
X

B>Cu Cu<B
X
III
This proposal had been previously advanced (189) for Cul-bipy. Copper(II)
complex salts containing one, two, or three molecules of either chelate have
 CHELATE COMPLEXES OF 1 ,10-PHENANTHROLINE 963

been described. Most are hydrated and of a blue or violet-blue color (32, 153,

155, 235). Apart from the isolation of a large number of trischelate metal salts,
the existence of the ions Cu(phen)3++ and Cu(bipy)3++ has been demonstrated
polarographically (223). Attempts at the resolution of these octahedral com-
plexes have not succeeded, but the phenomenon of “induced asymmetry” has
been observed (64).
Bisphenanthroline silver® nitrate monohydrate has been obtained as pale
yellow needles (201, 228), and there is also a yellow perchlorate, [Ag(phen)3]C104
(235). It is unlikely that the metal is 6-covalent in this substance. The large
flat molecule of chelated phenanthroline may have the capacity to hold addi-
tional unchelated molecules by van der Waals forces in the solid lattice. It is
not inconceivable that this could occur also to some extent in solution. In a calcium
compound to which reference is made later no less than six molecules of phe-
nanthroline are present in the solid phase. It is perhaps significant that these
anomalously high coordination numbers have not been obtained with bipyridine.
Complexes with the silver halides do not appear to have been described.
Very pronounced stabilization of bivalent silver has been achieved with both
chelate molecules (9, 136, 201). Usually the silver(I) salt is simply treated with
the base and then oxidized anodically or with ammonium peroxydisulfate. The
bivalent complexes range in color from orange to red-brown. They are usually
sparingly soluble (57) bis derivatives. Trisbipyridine silver(II) nitrate has been
described, and presumably here the metal is octahedrally coordinated (201).
The usual condition of non-oxidizability of the anion applies, and peroxydi-
sulfates, nitrates, bisulfates, chlorates, and perchlorates have been prepared.
The potential of the system AgB2+/A.gB2++ has not been determined, and would
be difficult in any event, owing to the low solubility of the oxidant.
Oxidation to the tervalent state has not been reported, though lead dioxide,
ozone, bismuth dioxide, and cerium(IV) salts have been tried as oxidizing
agents (185).
Univalent gold can be expected to behave similarly to copper®, but no com-
plexes with either base have been reported. In the course of their work on organo-
gold compounds, Foss and Gibson (117) prepared [AuBr2 bipy] [AuBr4] inci-
·

dentally; more recently Block and Bailar (33) prepared a series of compounds
with both chelate molecules. The Au(III) cation contained two covalently
bonded halogens and one molecule of the chelate, the salts being of the form
[Au(B)X2]X and [Au(B)X2][AuX4] (X =
Cl, Br). Compounds containing more
than one molecule of base per gold atom were not obtained.
It appears from conductometric and spectrophotometric studies that the sub-
stances [Au(B)X2]X rearrange in nitrobenzene solution (132).

2[Au(B)X2]X ± [Au(B)X2]+ + [AuX«]- + B

B. GROUP 2 METALS

By the reaction between phenanthroline and the alkaline earth perchlorates

in methanol solution, Pfeiffer (228, 235) obtained colorless crystalline solids
,[M(phen)3](C104)2 (M =
Be, Mg) and [M(phen)4](C104)* (M Ca, Sr, Ba). A
=
964 BRANDT, DWYER, AND GY ARFAS

crystalline substance of the formula [Ca(phen)6](C104)2 was also isolated, but

this was thought to be an inclusion compound, because it decomposed in 5 per
cent methanol in ether to the tetrakis compound. The molecular weight of the
tetrakis calcium compound in nitrobenzene was found to be 435 compared with
the theoretical value of 959. Though Pfeiffer cited the ammines MC12 -6(or 8)NH3
as evidence, high coordination numbers of six for beryllium and eight for cal-
cium, strontium, and barium are most improbable. As pointed out earlier, addi-
tional molecules of unchelated base may well be attached by van der Waals
forces. No bipyridine compounds have been described.
The colorless zinc and cadmium complexes with both bases have been investi-
gated by a number of workers (38, 153, 173, 231, 232, 235, 252) and appear to
be 4- or 6-covalent, containing one, two, or three molecules of the base. In
aqueous solution the zinc ion may attach one, two, or three molecules of phe-
nanthroline (168), and polarographic studies have demonstrated that in the
presence of at least one hundred fold excess of the base, the ions Cd(bipy)3++
and Cd(phen)3++ exist (68).
The mercury complexes appear to have been little investigated. There is a
report of a sparingly soluble Hgl2 bipy that can be used for the analysis of bi-
·

pyridine (201), and trisphenanthroline mercuric perchlorate has been prepared

(228, 235). The latter substance may be only 4-covalent.

C. GROUP 3 METALS

Aluminum compounds with these chelates have not been described, although
it is probable that they could be prepared under anhydrous conditions. Rare
earth complexes are not known.
In aqueous solution indium salts yield the sparingly soluble colorless com-
plexes [In(bipy)2]X3 and [In(phen)3]X3 (X =
Cl, Br, I, SCN). These substances
do not appear to hydrolyze very easily, but they have not been resolved into the
enantiomeric forms (72, 272). Preliminary investigations show that gallium can
coordinate three molecules of either chelate, but the complex cations are hy-
drolyzed to some extent in aqueous solution with the separation of gallium
hydroxide (73).
A bisbipyridine thallium(I) nitrate has been prepared (228), and a bisphenan-
throline thallium (I) iodide has been obtained by pyrolysis of the tervalent com-
plex (273).
In aqueous or dilute alcoholic solution thallium(III) is coordinated by both
bases to yield sparingly soluble complexes of the form [T1(B)]X3 (X =
Cl, Br,
SCN) (273). The evidence available does not allow one to decide whether these
involve dimeric bridged structures or should be formulated as [Tl^XjX.
The orange-red iodides, T1(B)2I3, do not react with starch-potassium iodide
in dilute acetone solution and the existence of the I3~ ion thus seems to be pre-
cluded, so that these substances may well be [T1(B)2I2]I. Consequently, they
provide an interesting addition to the list of valence-stabilized compounds, since
the simple iodide T1I3 does not exist. A similar chelate structure has been pro-
posed for the bisethylenediamine thallium (III) iodide (137).
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 965

D. GROUP 4 METALS
Phenanthroline chelates with tin(II) and lead(II). [Sn(phen)3](C104)2 and
[Pb(phen)4](C104)2 have been described (235). Although the lead compound has
been quoted (192) as an example of 8-covalency, the evidence available is purely
analytical and this substance may contain free phenanthroline. Tetrachlorobi-
pyridinetin(IV) has been obtained by reaction of stannic chloride and the base
in alcohol. The corresponding bromide, Sn(bipy)Br4, has been isolated as un-
stable, highly insoluble, brown needles (3).
E. GROUP 5 METALS

Recent work (210) shows that vanadium coordinates with phenanthroline in

the oxidation states of +2, +3, and +4, giving deeply colored complex salts.
Vanadium(II) solutions react to give the intensely blue-violet V(phen)3++.
Most salts of this cation are extremely soluble in water, but the iodide can be
obtained as unstable blue-black crystals. The solutions are oxidized rapidly even
in the absence of air to give the tervalent complex cation. It appears that little
stabilization of the bivalent state is achieved by coordination.
Aqueous solutions of vanadium(III) salts react to give deep violet-purple
colors, due to the ion [V(OH)(phen)2]++. This is similar to the brown iron(III)
complex cation and is also presumably dimeric. The sulfate and perchlorate are
sparingly soluble and stable in air, but the solutions are slowly oxidized to a
tetravalent vanadium complex cation. In non-aqueous solvents or by fusion
with phenanthroline, the purple crystalline [V(phen)2Cl2]Cl can be obtained, but
it is easily hydrolyzed.
From solutions of vanadyl salts the greenish-yellow [V0(phen)2]S04-5H20
and [VO(phen)2] (C104)2 ·4H20 can be obtained. The structures of these salts are
unknown. However, from solutions of vanadium(IV) chloride in carbon tetra-
chloride V(phen)2Cl4 can be isolated as a brown precipitate, which can be formu-
lated as [V(phen)2Cl2]Cl2.
The only other Group 5 element which coordinates is bismuth. It forms the
salt [Bi(phen)3](C104)2 (235), which is not necessarily 6-covalent.
F. GROUP 6 METALS

The chelation of bipyridine with chromium(II) salts was first reported by

Barbieri and Tettamanzi (10), who isolated the black bromide, Cr(bipy)3Br2·
6H20. The dark brown solution rapidly lightens in color in the air with the for-
mation of the Cr(III) complex, but the solid appears to be stable in dry air
(8, 44). The extent of the stabilization achieved by coordination can only be
speculated upon, since data as to the redox potential are unavailable, although
it is believed (8) to be appreciable. However, it might be mentioned that the
sparingly soluble chromium (II) salt addenda with hydrazine are also quite stable
in the solid state. This can be ascribed to either their insolubility or the reducing
power of the hydrazine molecule (163, 278). The magnetic moment of the bro-
mide (3.27 B.M.) indicated 3d24s4p3 bonds, and it is suggested that the strong
bonds and electron pairing are due to pronounced double bonding here as with
966 BRANDT, DWYER, AND GYARFAS

iron(II) (44). An attempted resolution through the tartrate was not successful
(72), owing to rapid oxidation to the Cr(III) complex cation. If the latter were·
optically labile, as anticipated, then rapid electron transfer would also cause
racemization of the Cr(II) compound (85).
When solutions of [Cr(bipy)3]Cl2 are treated with magnesium, reduction to the
Cr(I) complex has been claimed to occur and a blue perchlorate isolated. The
magnetic moment (2.1 B.M.) is slightly high for the required one unpaired elec-
tron with 3d24s4p3 bonding. In solutions of Cr(bipy)3++ the following equilibrium
has been proposed (135):

2Cr(bipy)3++ ?=* Cr(bipy)3+ + Cr(bipy)3+++

Since one might expect Cr(I) to be 4-covalent, confirmation of this work is
desirable.
Hammett, Walden, and Edmonds (130) have reported the preparation of
Cr(phen)3++ and the determination of the potential of the reaction:
Cr(phen)3++ <=» Cr(phen)3+++ + e~

However, it now seems possible that Cr(II)-phenanthroline complexes may

not exist. Instead, during the course of the preparation the phenanthroline is
reduced to an orange substance of unknown composition. This is the substance
that is reversibly oxidized (100).
With tervalent chromium, the yellow [Cr(bipy)3](C104)3 is made from the
bivalent complex (10) and [Cr(bipy)2Cl2]Cl from anhydrous chromium (III)
chloride in methanol solution in the presence of a trace of zinc dust (44). The
phenanthroline analog of the latter substance has been prepared by fusion
of anhydrous chromium chloride and the base (236), and the tris compound
[Cr(phen)3] [Cr(SCN)6] from potassium hexathiocyanatochromate(III) and
the base in boiling alcoholic solution.
It has been observed that solutions of reduced molybdenum give a red color
with bipyridine, but no definite compound has been isolated (169).
No definite compound results from the interaction of uranium(IV) chloride
and bipyridine in alcoholic solution, but on the addition of potassium thio-
cyanate, a small amount of yellow [U(bipy)3](SCN)e is precipitated. This is
probably the result of oxidation (3).
G. GROUP 7 METALS

With 2,2'-bipyridine and Mn(II) salts the yellow trisbipyridine compounds

Mn(bipy)5X2 result (204), but one molecule of the chelate is easily lost in aqueous
solution and the diaquo salts result. Pfeiffer (235) has prepared [Mn(phen)4l-
(0104)2 and by distillation of this substance obtained the bis derivative. There
is evidence, however, that trisphenanthroline complexes exist in aqueous solu-
tion (230). For reasons advanced pre\fiously, the tetrakis derivative is almost
certainly not 8-covalent. The Mn(II) compounds apparently are not capable of
being oxidized to the tervalent compounds, but by operating with ammonium
peroxydisulfate it is possible to obtain the octahedrally coordinated
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 967

[Mn02(bipy)2]S2O8 (221). The magnetic moment of [Mn(bipy)3]Br2 indicates

five unpaired electrons; hence electron pairing or significant double bonding does
not occur (44).
With the exception of the simple salts of bipyridine with perrhenate, hexa-
chlororhenate(IV), and tetracyanodioxorhenate(V), no rhenium derivatives of
either base have been described (209).

H. GROUP 8 METALS

The elements of the iron triad, iron, ruthenium, and osmium, form complex
cations of great stability in the bi- and tervalent states. The red iron(II) tris-
chelates are so well known that they need no further comment, but the evidence
for the existence of bis and mono derivatives is of special interest in order to
locate the point at which the paramagnetic hydrated ferrous ion changes to the
diamagnetic cation with strong covalent bonds. Spectrophotometric evidence
for the existence of the ion Fe(bipy)++ has been obtained by Krumholz (170)
and Baxendale and George (19). This is in disagreement with the findings of
Ferrari (113) and Gould and Vosburgh (125). Jaeger (155) has claimed to have
obtained FeS04-bipyasa crystalline solid, but when mixed -with barium chloride,
only the tris compound [Fe(bipy)3]Cl2 resulted. The existence of the pale yellow
Fe(phen)++ has been demonstrated in sulfuric acid solutions of the base in the
presence of a large excess of ferrous salt (167). There does not appear to be any
reliable evidence, however, for the existence of either Fe(bipy)2++ or Fe(phen)2++
in aqueous solution. In the orange solutions of 2-methylphenanthroline and
iron(II) salts, there appears to be an equilibrium between mono-, bis-, and tris-
chelated cations. Negligible amounts of the tris compound are present, very
much of the bis compound, and a small amount of the mono compound. The
lability of the system is demonstrated by the complete precipitation of the tris
complex as the perchlorate (145, 147).
Recent work (12a) has shown that [Fe(bipy)Cl2], [Fe(bipy)2Cl2], and
[Fe(phen)2Cl2] can be obtained by the same method used to prepare [Ni(bipy)2Cl2]
(234). When trisbipyridine iron(II) chloride hexahydrate is heated in vacuo at
100°C. loss of water and one molecule of bipyridine results, with the formation
of violet [Fe(bipy)2Cl2], This substance is paramagnetic (5.1 B.M.) and can be
considered to have four unpaired electrons, the excess moment being due to
Fe(III) impurity or insufficient diamagnetic correction. With small amounts of
water the substance becomes red, and in alcoholic solution spectrophotometric
studies show only the presence of the usual tris compound. At 150°C. the violet
substance, with further loss of bipyridine, changes to the red [Fe(bipy)Cl2]. Pre-
liminary studies show that this is paramagnetic (1.8 B.M.). Whether the para-
magnetism is due to impurity or the structure is a planar dimer, with metal-
metal interaction, remains to be determined.
At 150°C. the trisphenanthroline compound changes to the blue paramag-
netic (5.2 B.M.) [Fe(phen)2Cl2], which does not undergo further change at 200°C.
(12a). The dark red dibromo and diiodo bisphenanthroline compounds have
been reported (163a) as being obtained from the tris compounds in vacuo at
968 BRANDT, DWYER, AND GYARFAS

100°C. and as being slightly paramagnetic and diamagnetic, respectively. How-

ever, the tris compounds are merely dehydrated at this temperature (12a), but
at 255°C. the blue dibromo bis compound, which is paramagnetic (4.9 B.M.),
results. Loss of phenanthroline from the iodide does not occur even at this tem-
perature. Though the situation is not identical, it can be reasonably inferred
that the diaquo ion [Fe(phen)2(H20) 2]++ is also paramagnetic, as required by the
hypothesis of orbital stabilization.
The orange-red trisbipyridine ruthenium (II) chloride was first prepared by
heating a mixture of ruthenium (III) chloride with bipyridine at 250°C, Tetra-
pyridine is formed in the reaction (41). The trisbipyridine and trisphenanthro-
line complexes are more conveniently prepared by refluxing an aqueous alcoholic
mixture of potassium pentachlorohydroxyruthenate (IV), the organic base, and
sodium hypophosphite (71, 95). The cations Ru(bipy)3++ and Ru(phen)3++ are
remarkably stable and are unattacked by boiling concentrated hydrochloric acid
or caustic alkalis. The free bases can be obtained as crystalline solids and are
stable, unlike [Fe(bipy)3](OH)2 (32).
The intensely dark-green Os(bipy)3++ is obtained by heating mixtures of po-
tassium hexachloroosmate (IV), osmium metal, and bipyridine at 250°C. The
use of ammonium hexabromoosmate(IV) obviates the necessity for the presence
of free osmium (43). The dark greenish-brown Os (phen) 3++ is best prepared
from the base and ammonium bromoosmate at 200°C. in glycerol solution (84).
Both of these complex cations are comparable in stability with the ruthenium
analogs, and the free bases can be isolated (43, 84).
The oxidation of Fe(bipy)3Cl2 and Fe(phen)3Cl2 was first reported by Blau
(32), who isolated the green chloroplatinates. The blue perchlorates are also
easily isolated (44, 97). The magnetic moment of [Fe(bipy)3](C104)3 (2.1-2.4
B.M.) shows that strong 3d24s4p3 bonds are present (44). In the light these blue
substances become red, presumably as a result of reduction, but the reaction
has not been studied.
Addition of the bases to solutions of iron(III) salts does not yield the foregoing
trischelated compounds (32); instead, brown binuclear salts, [Fe(B)2-OH]2X4,
result (119). All attempts to prepare the blue compounds directly have failed,
and in ether or acetone solution, for example, brown water-soluble substances,
[FeCls-bipy] and [FeClrphen], result (253, 254). The binuclear substance, IV,
has a moment of 1.4 B.M., much smaller than the theoretical value calculated
for spin only and the partial neutralization of the moment has been advanced
in support of the binuclear structure (119). Michaelis and Granick (199) in a
reexamination of the compound found the values of 2.1 B.M. for the almost
saturated solution and 2.4 B.M. for the solid and concluded that there is no ap-
preciable metal-metal interaction.
OH 4+
z \
(bipy)2Fe Fe(bipy)2
OH
IV
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 969

Irving (146) has compared the structure of this substance with the diol-iV,iV'-
diethylethylenediamine copper(II) salts and proposed that the diol-bridged
structure is adopted to relieve steric strain, in this case due to the three large
molecules attached to the iron(III) ion. It should be pointed out, however, that
there is no significant difference in the octahedral covalent radius of iron(III)
compared with iron(II) (225) to suggest that iron(III) salts should behave dif-
ferently from iron(II) salts. It is known that when phenanthroline is added to
very weakly acid solutions of iron(III) salts, a basic salt is precipitated (268), and
further, that at the pH at which the binuclear complex is formed the iron(III)
ion is already hydrolyzed, i.e., has at least one OH- group already attached. In
the same way, the very stable potassium ferricyanide cannot be made directly
from iron (III) salts and potassium cyanide. The blue tris ruthenium (III) com-
pounds also cannot be prepared directly from ruthenium(III) salts and the bases
in either neutral or acid solution (90).
Oxidation of the cation Ru(B)3++ is easily effected anodically or with ce-
rium (IV) in nitric acid solution, but pure crystalline salts of the blue cation
Ru(B)3+++ have not been isolated, as they rapidly undergo reduction. Some re-
sults are available on the reaction between phenanthroline and potassium penta-
chloroaquoruthenate(III). In neutral or weakly acid solutions, green or brown
complex cations, [Ru(phen)2X2]+ (X =
Cl, OH), are conveniently isolated as
the perchlorates. With hydrochloric acid replacement of the hydroxyl groups
occurs, and this reaction is reversed with sodium hydroxide. Oxidation with
ceric nitrate leads to deep blue ruthenium(IV) compounds, and reduction to
orange or brown ruthenium(II) compounds with some disproportionation to the
tris complex. There is some evidence that cis and trans forms may exist, but no
evidence has been obtained for diol-bridged structures as with iron (III) (90).
After oxidation with chlorine, salts of the red cation Os(B)3+++ are easily iso-
lated (86). Dichloro-bisbipyridine osmium (III) chloride has been reported among
the products of pyrolysis of mixtures of potassium hexachloroósmate (IV) and
bipyridine (43).
The tervalent cations M(B)3+++ (M Fe, Ru, Os) undergo a remarkable re-
=

action when the solutions are made alkaline. Reduction occurs instantly; if acid
is added at once, reoxidation can be effected. The reaction was first reported by
Blau (32), who observed the odor of ozone following the addition of alkali.
Hydrogen peroxide can also be detected (86). In weakly acid solutions of the
iron and ruthenium compounds the reduction also occurs slowly (95, 97).
The following reaction is said to occur (283):

Fe(Ci2H8N2)3+++ + OH- -* Fe(Ci2H3N2)3++ + OH

The ozone and hydrogen peroxide doubtless arise from reactions of the hy-
droxyl radical. It is noteworthy that oxidation of the bivalent complexes is not
effected by ozone or dilute hydrogen peroxide at room temperature.
In the cobalt triad, compounds of cobalt(II), cobalt(III), and rhodium(III)
have been described, but no iridium compounds.
Brown tris cobalt(II) salts with a wide variety of anions are readily prepared
in the absence of oxygen by direct addition of three molecules of the base to the
970 BRANDT, DWYER, AND GYARFAS

simple cobalt(II) salt. Mono and bis compounds are also known in which the
metal can be either 4- or 6-covalent, with water molecules occupying some of
the coordination positions (32, 44, 154, 235, 236). The moment of the perchlorate
Co(bipy)3(C104)2 (4.85 B.M.) indicates ionic bonds (44).
Oxidation of the cobalt(II) complexes can be effected with oxygen, hydrogen
peroxide, or chlorine. The yellow tris compounds are diamagnetic like most
cobalt(III) compounds (44). One molecule of the base is easily disengaged and
replaced by aquo groups. Though it has been stated (236) that Co(phen)3Cl3
yields a red color with ferrous sulfate only on long heating of the solution, this
is at variance with the method of preparation of czs-dichloro-bisphenanthroline
cobalt(III) chloride by removing one molecule of the base simply by refluxing
in chloroform solution. Green irans-dichloro-bisphenanthroline cobalt(III) chlo-
ride, though it dissolves in nitrobenzene to a green solution, yields a rose-colored
solution in water because of rapid aquation. This is confirmed by the isolation
of [Co(phen)2(H20)2][Cr(NH3)2(SCN)4] from the aqueous solution (235, 301).
Rhodium(III) salts can attach one, two, or three molecules of bipyridine, but
only the brown compound Rh(bipy)3Cl3 has been obtained pure (152). The sub-
stance [Rh(bipy)2Cl2]Cl has been isolated, and the substance Rh(bipy) Cl3 4H20
·

with varying amounts of ionizable chlorine. The lability of these substances and
also of cobalt(III) complexes contrasts with those containing ethylenediamine.
In many experiments a sparingly soluble pink substance, Rh2(bipy)3Cle, was ob-
tained. This can be formulated in various ways in which rhodium is in both the
cation and anion, such as [Rh(bipy) 3] [RhCU. Compounds containing rhodium
and phenanthroline have not been described.
When mixtures of rhodium (III) chloride and bipyridine are treated in metha-
nol solution with zinc amalgam, intensely colored violet solutions result. These
rapidly change to yellow in the air and presumably contain chelated derivatives
of rhodium(II) or rhodium (I) (72).
All of the metals of the nickel triad can be coordinated with either base, but
nickel(II) alone attaches three molecules. The bright pink tris compounds are
of the same order of stability as the iron(II) compounds, and the phenanthroline
complex is the more resistant to acids. This can be heated in 10 N sulfuric acid
for a short time without decomposition (82). Their magnetic moments, the na-
ture of the bonds, and the failure to oxidize them to nickel(III) complexes are
discussed subsequently (page 977). Salts of the cation Ni(B)3++ with a wide
variety of anions are known (32, 202, 229, 234).
The bright green Ni(bipy)2Cl2 can be prepared by heating the tris compound
in vacuo at 140°C. for 2 hr. (234). The substance very rapidly disproportionates
when added to water, to yield the pink tris cation; but when it is added to a
solution of potassium iodide, the blue [Ni(bipy)2(H20)2]I2 immediately separates
(12a). It is of interest to mention that when d-Ni(bipy)3Cl2 is subjected to py-
rolysis, the optical activity is lost (13). This does not indicate, however, whether
the bis compound has the cis or the trans octahedral structure.
The bisphenanthroline compounds are much more stable, though they also
undergo disproportionation. When phenanthroline is added to nickel salt solu-
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 971

tions, the formation of the bis compound is evidenced by the resulting blue color,
but even when the ratio of nickel salt to base is less than 1:2, the addition of
sodium perchlorate instantly precipitates a substantial portion of the metal as
the pink trischelated salt. On standing for a short time, only the pale green
color of the hydrated nickel ion remains (82). The blue and blue-violet
[Ni(phen)2(H20)2]Cl2 3H20 and [Ni(phen)2(H20)2] (C104)2 have been obtained,
·

however, in crystalline form (229, 234).

Jaeger and van Dijk (151a) claim to have obtained the mono derivative
[Ni(bipy) (H20)2]S04 in two crystalline forms, blue and also indigo needles, by
the reaction between the tris compound and excess nickel sulfate solution. The
same substance could be obtained by direct synthesis in alcoholic solution. The
substance [Ni(phen) (H20)J (004)2 has been described as being formed when
nickel perchlorate and phenanthroline in the ratio of 5:1 are heated in 30 per
cent methanol (229). The bis- and monochelated compounds in the resulting
mixture were then separated by crystallization. The mono compound gave dark
blue solutions in methanol. The evidence for the isolation of these mono com-
plexes does not appear entirely satisfactory. The stabilities favor the trischelates
so strongly, and the rates of dissociation are so high (13) for the monochelates,
that in a solution where equilibrium is rapidly attained, they must be present in
very low concentration. Spectrophotometric evidence, however, does suggest
that the nickel ion can form mono complexes (284) in aqueous solution.
Bivalent palladium is almost invariably 4-covalent and attaches usually one
molecule of either base to form the sparingly soluble [Pd(B) · 2] (X =
Cl, Br, I,
SCN, N02). These substances are more soluble in the presence of excess of the
base, and the existence of the ions Pd(B)2++ in such solutions is evidenced by
the isolation of the perchlorate, [Pd(phen)2](C104)2, and the 2-naphthol-6-sul-
fonate. Mixed tetramines containing ethylenediamine and pyridine can be iso-
lated and dichloro-phenanthroline palladium(II) can be oxidized to the easily
reducible tetrachloro-phenanthroline palladium(IV) (181, 182, 247).
Platinum (II) behaves in a similar manner to palladium, one molecule of bi-
pyridine coordinating firmly and the second weakly. The substance [Pt(bipy)Cl2]
exists in yellow and red dimorphous forms. The bis chloride cannot be isolated,
but [Pt(bipy)2][PtCl4] and [Pt(bipy)2]I2 2H20 exist. The latter has red and black
forms in the solid state. Mixed tetramines containing ammonia, pyridine, and
ethylenediamine can be prepared and also triamines, [Pt(bipy) (NH3)X]X. Treat-
ment of the tetramines and also the triamines with acids causes loss of the am-
monia, ethylenediamine, etc. and leaves the monobipyridine compound.
Oxidation to platinum(IV) complexes can be effected with chlorine. The sub-
stances [Pt(bipy)2Cl2][PtCUl, [Pt(bipy)Cl4], and [Pt(bipy)(Py)ClJCl result (205).

III. The Polypyridines and their Metal Complexes

a. 2,2',2#-terpyridine [2,6-di(2-pyridyl)pyridine]
This substance (V) was isolated (203) among the products of the dehydrogena-
tion of pyridine with anhydrous ferric chloride in an autoclave at 500°C. A sub-
972 BRANDT, DWYER, AND GYARFAS

stance described originally as 2,2', X"-terpyridine, which also gave an intense

color with iron(II) salt solutions, was subsequently shown to be a dimorphous
form of 2,2', 2"-terpyridine (207). Terpyridine (V) acts as a tridentate planar
chelate molecule, and the formation of the platinum derivative [Pt(trpy)Cl]Cl
(VI) has been advanced as evidence for the planar distribution of the bonds about
bivalent platinum (206).

AAwx/\
Vn nV1 Vn—Pt—nV
ci
VI
Vapor-phase bromination gives a mixture of the 6-bromo and 6,6"-dibromo
derivatives (42), the former of which produces a weak magenta color with iron(II)
salts, while the latter is insoluble in acids and no color reaction occurs.
The chelate complexes derived from Cu(II), Ag(I), Ag(II), Zn(II), Cd(II),
Hg(II), Pd(II), Pt(II), and Ir(III) contain one molecule of the base, while those
from Fe(II), Co(II), Co(III), Ni(II), Ru(II), Os(II), Cr(III), and In(III)
contain two. Since the bonds from the nitrogen atoms to the metal must lie in
the plane of the pyridine ring and resonance requires coplanarity of the three
rings, it follows that all the bonds from the chelate must be planar. Though the
molecule can encompass three positions in a plane in the planar or octahedral
coordinations, it is apparent that there is some strain involved, the bonds from
the nitrogen atoms being at a little less than 90°. This would suggest, as actually
is the case, that the metal complexes would be somewhat less stable than would
be anticipated for a tridentate chelate molecule. From models it can also be
seen that even if the stabilizing resonance in the system is sacrificed and the
pyridine rings are moved out of coplanarity, it is not possible for the molecule to
span three positions in tetrahedral coordination. It has been suggested, therefore,
that the zinc, cadmium, and mercury complexes have the planar configuration
(207). If this is so, these complex compounds are further examples of forced bond
deformation analogous to some metal derivatives of phthalocyanine (180). The
steric conditions and the bond strengths do not seem, however, to be so favorable
here. The zinc and cadmium complexes formulated as [M(trpy)Cl]Cl could well
be of the form [M(trpy)ClJ, with one of the donor nitrogens remaining un-
attached. The formulation of the mercury derivative Hg(trpy)(N03)2 is difficult
in the solid state, though in solution the complex ion could be [Hg(trpy) (H20)2]++.
Difficulties also arise in the formulation of the colorless silver(I) compounds
Ag(trpy)NOs and Ag(trpy)ClC>4. It is possible that the silver could be linear 2-
covalent when either the center nitrogen atom is not attached or is attached by
a weak bond, or the substance is polymeric, utilizing single nitrogen-metal bonds.
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 973

If the metal is tetrahedral 4-covalent, like monovalent copper, the same diffi-
culties arise as with zinc and cadmium.
The silver(I) compound can be oxidized anodically to the brown terpyridine
silver(II) nitrate, formulated as [Ag(trpy)N03] NOs, and a chlorate and
perchlorate of the same type have been isolated. It is noteworthy that a second
molecule of the chelate could not be attached. According to the results of Cox,
Webster, and Wardlaw (62), bispicolinic acid silver(II) is planar.
With copper(II) chloride, the green substance [Cu(trpy)Cl]Cl ·2H20 is formed,
and from the increase in solubility in the presence of excess base, there is some
evidence for the existence of the octahedral ion, Cu(trpy)2++, in solution. The
pale yellow palladium compound [Pd (trpy)Cl] Cl is obviously similar to the red
platinum(II) compound (VI).
The stable bivalent platinum compounds [Pt(trpy)X]X (X =
Cl, Br, OH)
can be oxidized with chlorine to trichloro-terpyridine platinum(IV) chloride,
analogous to the iridium compound [Ir(trpy)Cl3]Cl.
With octahedral coordination, Morgan and Burstall considered three ar-
rangements of the two chelate molecules. Only one of these possible arrangements
satisfies the condition of coplanarity of the pyridine rings, that in which the
molecules of the base are arranged in two equatorial planes at right angles (VII).

VII
This has a plane of symmetry, and attempts to resolve the stable ruthenium(II)
and nickel compounds, [M(trpy)2]++, were not successful. It might be mentioned
that a substituent in one of the outside pyridine rings can confer asymmetry. A
substance of such a type is the planar tridentate molecule 8-(2-pyridylmethylene-
amino)quinoline(VIII), the iron(II) compound of which has been obtained
in enantiomeric forms (77).

VN NV
VIII
974 BRANDT, DWYER, AND GYARFAS

Bisterpyridine iron (II) salts are characterized by an intense purple color, so

intense that a definite color reaction for iron (II) is observed at concentrations of
one part in two million. As an analytical reagent, it has been used for the de-
termination of iron in sea water and in marine plankton (61). Oxidizing agents,
such as chlorine and cerium(IV) salts, change the color to green (or sometimes
yellow), but though the ion Fe(trpy)2+++ undoubtedly exists in these oxidized
solutions, attempts to isolate crystalline salts have failed. This appears to be due
not only to its tendency to undergo reduction to the iron(II) compound, but
also to loss of one of the tridentate ligands (89).
The red ruthenium (II) and dark brown osmium (II) complex compounds have
been prepared (207) by pyrolysis of mixtures of the base with ruthenium (III)
chloride and ruthenium metal, and with potassium hexachloroosmate(IV) and
osmium, respectively. These substances show a stability towards acids com-
parable to that of the analogous bipyridine and phenanthroline derivatives.
Oxidizing agents convert them to the green tervalent derivatives. Bisterpyridine
ruthenium(III) salts have been obtained only in solution, but bisterpyridine
osmium(III) perchlorate can be isolated as a crystalline solid (207). The redox
potentials of the complexes of the iron triad are discussed subsequently.
Cobalt(II) salts react with the organic base in aqueous solution with the forma-
tion of the deep brown ion [Co(trpy)2]++, which can be isolated as the almost
black bromide [Co (trpy) 2] Br2 3,5H20. The formation of this ion has been used
·

as a colorimetric method for the estimation of cobalt (213).

Oxidation of a mixture of cobalt(II) chloride and terpyridine with chlorine
yields the yellow substance [Co(trpy)2]Cl3 7H20, analogous to the chromium (III)
·

complex [Cr (trpy) Cl3 2H20. The latter is obtained from anhydrous chro-
2]
·

mium(III) chloride and the base in alcoholic solution in the presence of a trace
of zinc dust. The bischelate nickel compounds are light brown and are stated to
be comparable in stability to the trisbipyridine compounds (207).
The indium complex salts separate as precipitates on mixing an alcoholic solu-
tion of the base with the halides in aqueous solution. They can be crystallized
from acetone, but are decomposed by dilute acid or alkali (273).
b. 2,2' ,2", 2'"-tetrapyridine [6,6'-di (2-pyridyl)-2,2 '-bipyridine]
2,2',2",2'"-Tetrapyridine (IX) has been obtained by heating bipyridine with
either ferric chloride or ruthenium (III) chloride, or by the Ullmann synthesis
from 2-bromopyridine and 6,6'-dibromo-2,2'-bipyridine (42).

The coordination compounds have been described by Morgan and Burstall

(208), who consider that the molecule acts as a planar quadridentate base. From
a model, however, this does not appear at all likely, since the fourth nitrogen
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 975

atom is quite far from the required position for square coordination. At best it
could only function as a tridentate molecule, and even here steric difficulties
may arise from the fourth pyridine ring. From superficial inspection this molecule
might be considered analogous to the porphyrin or phthalocyanine molecules,
but the latter form a closed system with pyrrole rings which form six-membered
chelate rings, whereas with tetrapyridine they would all be five-membered.
All of the metal chelates contain one molecule of the substance and do not
appear to be notably stable. The silver® compound, Ag(tetrpy)N03, obtained
as yellow needles, significantly is not oxidized by peroxydisulfate ion to the
Ag(II) complex. The latter is usually 4-covalent and probably planar, and this
suggests strongly that the base is not functioning as a quadridentate. Similar
difficulties of formulation of the Ag® complex arise as with Ag(trpy)NOs.
With iron(II) salts, a greenish-yellow color reaction of relatively low intensity
occurs. The sparingly soluble reddish-brown Fe(tetrpy) S04 4H20 and the
·

greenish-brown bromide and iodide have been isolated. Examination of the

published properties of these substances suggests that they have not the robust-
ness of the bipyridine complexes
Hydrated salts [M(tetrpy)]X2 ( =
Co (pink), Ni and Cu (green), Zn and
Cd (pale yellow)) also have been prepared. It has been suggested that in all
of these substances the complex cation is planar, but it seems more probable that
in all of them molecules of water are included in the coordination sphere of the
metal and the usual stereochemistry applies. The formulation of the cobalt(III)
compound as [Co(tetrpy)Cl2]Cl 3H20 implies quadridentate chelate function,
but there is not good evidence that the third chlorine is ionized or, alternatively,
that some of the water molecules are not really coSrdinated aquo groups.
C. HIGHER POLYPYRIDINES: 2,2', 2" 2"', 2""-PENTAPYRIDINE [2,6-BIs(2,2'-
,

bipyridyl-6-yl)pyridine] and 2,2', 2" ,2'" ,2"" ,2'""-hexapyridine

[6,6'-bis (2,2'-bipyridyl-6-yl) -2,2'-bipyridine]
These compounds have also been prepared (42) by the Ullmann reaction. No
metal chelates from the pentapyridine have been reported. Hexapyridine has
been stated (42) to form a tetrachloride and to yield a series of complex cations
with the transition metals in which one mole of the base is united with one
atomic proportion of the metal ion. From the stereochemistry of the substance
it does not seem possible that it could act as a sexadentate molecule even if the
coordination of the metal atom involved a group of planar hexagonal orbitals.

IV. The Nature of the Bonds

Bipyridine and phenanthroline are both relatively weak bases, yet they chelate
with such a wide variety of metal ions and the resulting complexes are generally
of such a high order of stability that it is evident that much more is involved
than the simple functional grouping —C=N—C—C—N=C—. A comparison
with ethylenediamine, a much stronger base with the same atomic grouping,
is instructive. Trisethy enediamine ferrous chloride, like the hexammine, is a
1

weakly colored substance, easily hydrolyzed in water. It is paramagnetic, with

976 BRANDT, DWYER, AND GYARFAS

ionic or 4d24s4p3 bonds (134, 159). The deeply colored bipyridine and phe-
nanthroline compounds, on the other hand, are decomposed slowly by acids, and
their diamagnetism (44, 49, 50) indicates strong covalent 3d24s4p3 bonds. The
trisethylenediamine ruthenium (I I) and osmium (I I) salts have not yet been
prepared; it can be inferred that they will be very easily oxidized in the air, but
will probably be more stable than Fe(en)3Cl2. The diamagnetic (116) ruthenium-
(II) and osmium (I I) compounds with bipyridine and phenanthroline resist boil-
ing concentrated acids and alkalis and like the iron(II) compounds are easily
obtained in the enantiomeric forms. In the same way the stability and the exist-
ence of optical forms of the nickel compounds contrast sharply with the tris-
ethylenediamine nickel salts.
As far as the iron triad of Group 8 is concerned, there is also the remarkable
stabilization of the bivalent state evidenced by the redox potentials. In the bi-
valent iron, ruthenium, and osmium compounds, all of the available d, s, and p
orbitals are filled and this in itself favors strongly the bivalent state because oxida-
tion would involve the removal of one electron. Nyholm (44, 219), following the
original suggestion of Pauling (161, 225), has proposed that the major factor
responsible for electron pairing and strong covalent binding in many of these
compounds is the formation of double bonds, and suggested that structures X
and XI contribute to the resonance hydrid.

X XI
Similar structures can be written for the phenanthroline compounds. Since for
octahedral coordination a maximum of only three double bonds is permissible,
structures such as XII and XIII would be preferable.

V—
N N= N-
\ / \ /
M M
XII XIII
Similar structures have been proposed for copper phthalocyanine in order to
account for the extraordinarily high stability considered in relation to the low
proton affinity of the parent quadridentate molecule (192).
All of these double-bond structures involve the placing of a negative charge
on the para carbon atom of the pyridine ring, and it has been suggested (219)
that electron-attracting substituents in this position should permit further
charge delocalization from the nitrogen atom and enhance the amount of double
bonding. It should be remembered that such substituents lower the base strength
and hence for this reason tend to lower the stability of the metal complex.
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 977

No quantitative data are available on substituted bipyridines of this kind,

but the stabilities of the iron compounds with 5-nitro- and 5-chloro-
phenanthrolines do not show any departure from the usual relationship between
strength of the base and stability of the metal complex (34).
The values of the force constants of the carbon-oxygen bonds in tetracarbonyl
nickel, bipyridinedicarbonyl nickel, and o-phenylenebis(dimethylarsine)di-
carbonyl nickel have been compared; the weakening of the carbon-oxygen bond
has been ascribed to double bonding by the chelate molecule (222). It is inferred
that the nickel-bipyridine bond has some double-bond character. Evidence of
this kind is not entirely conclusive. If one compares the infrared spectra of
Ni(CO)4 with that of Fe(CO)s and Cr(CO)e, it is readily apparent that the
metal-carbon stretching frequency increases in the order nickel, iron, chromium,
and that the carbon-oxygen frequency decreases in the order nickel, iron, chro-
mium (38, 63, 255). If one assumes that the available d electrons are used in
double-bond formation, it would be expected that the double bondedness should
decrease in the order nickel, iron, chromium. This is inconsistent with the ob-
served trends of stretching frequencies for metal-carbon and carbon-oxygen.
Though some double-bond character might be expected to be present in the
pyridine-metal link, the experimental evidence is unfavorable. The placing of
electron-attracting substituents in the appropriate positions in pyridine and
the comparing of the first dissociation constants of the silver ion adducts with
Ng(py)z+ suggests no double bonding (215). Similarly, comparison of the dis-
sociation constants of bispyridine silver ion and of bispiperidine silver ion leads
to the same conclusion (27).
It thus appears that though double bonding is indicated to account for the
stabilities of some phenanthroline and bipyridine metal complexes, there is as
yet no reliable evidence to suggest that it does occur.
Some interest has been shown in the fact that the trischelate nickel salts can-
not be oxidized to the tervalent or quadrivalent states. Though this has been
discussed in a recent review (220), some of the implications are worth clarifying.
The nickel compounds are paramagnetic, with “ionic” or upper-orbital binding,
and the unpaired electrons are thus considered to be in the lower 3d orbital. It
is proposed that the extent of the double bonding is insufficient for strong covalent
(lower-orbital) binding and hence that the electron promotion favorable for oxida-
tion has not occurred. This is consistent with the magnetic data on the assump-
tion that the promoted electrons would be paired in the 5s orbital. Oxidation,
however, would form a complex cation carrying three or four positive charges.
The situation with dichloro-bis-o-phenylene(bisdimethylarsine) nickel(III)
chloride and dichloro-bis-o-phenylene(bisdimethylarsine) nickel (IV) perchlorate
(220) is much more favorable, involving complex cations with one and two
positive charges.
One can agree with the view (44) that in metal complex compounds where
electron promotion could conceivably occur, it is far less common than generally
supposed, and add that far too little attention seems to be paid to the large
energy requirement for promotion. It cannot be accepted as a corollary that
978 BRANDT, DWYER, AND GYARFAS

where promotion has not occurred, the lower valence states should be preferred,
and hence ignore the more favorable orbital hybridization that could well exist
in an oxidized state. A few examples which also illustrate the effect of the ligand
and of the ionic charge of the complex are pertinent. Trisethylenediamine
cobalt(II) chloride is paramagnetic (197) with ionic or 4d24s4p3 bonds having
the unpaired electrons presumably in the 3d orbital. Promotion of an electron,
as may occur in the hexacyanocobaltate(II) or the hexanitrocobaltate(II) ions
(225), has not occurred, yet the substance is oxidized very rapidly in the air.
The cobalt(II) and nickel(II) compounds of the sexadentate molecule 1,8-bis-
(salicylideneamino)-3,6-dithiaoctane are both paramagnetic with ionic bonds
(96). The cobalt compound is oxidized spontaneously even in the solid state to
the diamagnetic Co(III) compound, but the nickel compound cannot be oxidized.
On the other hand, if pyridine-2-aldehyde is substituted for salicylaldehyde in
the above sexadentate, the resulting cobalt(II) bivalent cation is not oxidized
by dilute hydrogen peroxide in boiling solution (78). It is evident that, in general,
valence stabilization by coordination involves a great many factors, and as yet
much of our knowledge is incomplete and empirical.

V. Chelate Molecules Similar to Bipyridine

A number of chelate molecules which contain the functional grouping
—C=N—C—C—N=C— have been tested for the iron(II) color reaction. As
criteria of similarity, we shall consider the intensity of color, the stability towards
acids, and the preferred stabilization of the bivalent state. It will be apparent
from the subsequent discussion that other chelates which can be included in the
bipyridine category also permit the writing of double-bonded formulas. Emmert
(103, 104, 105) showed that 2-(2-pyridyl)pyrrole (XIV) and 2-pyridylhydrazine
(XV) formed the red iron(II) compounds, XVI and XVII.
H

H
XIV XV
H

XVI XVII
Although the pyridine nitrogens in XIV and XV permit double-bonded formulas,
as with bipyridine, the iron compounds were quite unstable and in air were
rapidly oxidized to the iron(III) compounds.
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 979

2,2'-Dipiperidyl was prepared by Blau (31) and the purity of the substance
has been confirmed by Smith (257). It gave a violet color with iron(II) salts,
and though no information is available on the composition of the iron compound,
it is presumably a trischelate derivative. This compound is of particular interest
because double-bond structures are not possible; hence stability and magnetic
data would be of great value.
2,2'-Diindoyl (XVIII) has been reported to give a red color reaction with
iron(II) (268). One would, however, expect this substance to behave like the
sterically hindered 2,2'-biquinoyl (see Section VII).
h2 h2,- A I—s s-

J_1x;
1
1

w
XVIII XIX
2,2'-Dithiazole (XIX) was found to yield a red solution with iron(II) salts on
heating, and a coral-red bis-2,2'-dithiazole ferrous bromide dihydrate could be
isolated. The corresponding substance from 4,4'-dithiazole (XX) was colorless.
Erlenmeyer and Ueberwasser (106) concluded that these substances were not
analogous to bipyridine and ascribed the difference to the bond angles in the
=N—C—C—N= systems. These bond angles are all 120° for bipyridine, but
are 110°, 125°, 125°, and 110° in XIX and 125°, 110°, 110° and 125° in XX.
There is some similarity between the dioximes and bipyridine, and double-
bond structures can also be written for these metal complexes. The intense
purplish-red iron (II) compound with dime thy lgly oxime is easily prepared in
the presence of bases (ammonia or pyridine), but only two molecules of the
oxime are coordinated. The remaining positions are filled by molecules of the
added base. The complex compound formed (XXI) is easily decomposed by
acids, and ferric hydroxide separates from the solutions on standing (279, 280).
R R
c—c
/ %
•N N- O

S—
H3N-Fe-NH. H
X/ W
0 O

R R
XX XXI
A number of substances, such as 2-(2-pyridine)quinoline, which do not give
iron(II) color reactions are discussed in Section VII.
980 BRANDT, DWYER, AND GYARFAS

Recently (77, 78) the Schiff bases (XXII) from pyridine-2-aldehyde and
primary aliphatic and aromatic amines have been found to form very intensely
colored iron(II) complex salts very similar to those formed with bipyridine.

%
X -N NR
NR \ /
M
XXII XXIII

Considerable resonance is possible in these chelate molecules and a double-

bonded structure (XXIII) can be written. 2-Pyridylmethyleneaminomethane
(XXII: R CHS) yields a tris iron(II) compound which is much more intensely
=

colored and qualitatively is more stable than the trisbipyridine compound.

Reversible oxidation to the blue Fe(III) compound is possible. Substitution of
the more weakly basic aromatic amines, as can be anticipated, lowered the
stability of the Fe(II) compound. 2-Pyridylmethyleneaminobenzene (XXII:
R =
C6H5), for instance, gave an iron(II) compound which was sensitive to
acids, and which was decolorized on boiling in dilute aqueous solution. The
color returned on cooling.
Reference has been made already to 8- ( -pyridylmethyleneamino) quinoline,
which is analogous to terpyridine. The Schiff bases from ethylenediamine (XXIV)
and trimethylenediamine behave as typical quadridentates toward palladium(II),
nickel(II), and copper(II). Unlike 2,2',2",2'"-tetrapyridine, these form such
intensely permanganate-violet iron (II) derivatives that less than one part of
iron in five million parts of water can be detected by the color reactions. The iron
compounds can be reversibly oxidized, but are sensitive to acids and have not
been obtained analytically pure. The two vacant coordination positions appear
to be associated with the relatively low stability.
Especially noteworthy is the tris Schiff base (XXV) from pyridine-2-aldehyde
and 2-aminomethyl-l ,3-diaminopropane. This acts as a sexadentate molecule
and can be considered as three linked-molecules of bipyridine. The permanganate-
colored iron(II) compound is stable to acid and is not oxidized by cerium(IV)
ion in nitric acid solution. A structure (XXVI) involving three double bonds at
right angles (nitrogen atoms with asterisk in XXVI) can be written; furthermore,
very considerable resonance stabilization is probable.

H2 C—ch2 NX
HC
HC=N N—CH N
\
c
/ \
c
XV
L H2 H
XXIV XXV
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 981

VI. Stability of Metal Chelate Complexes

The past few years have witnessed an increasing interest in the stability of
metal chelate complexes. As an important property of such compounds, it has
been hoped that this type of data will lead to an increased understanding of their
formation and behavior. The stabilities of the metal complexes of a particular
reagent are also important to the analytical chemist, as they affect the applica-
tion of organic reagents to a multitude of problems in separation and
measurement.
The determination of the stabilities of the phenanthroline complexes of metal
ions has been part of the overall program. The experimental approaches employed
have varied considerably. The familiar technique developed by Bjerrum (26),
which has provided the bulk of this type of information in other complex systems
(192), has been little used. Instead, the more standard approaches of equilibrium
and kinetic measurements have been the main methods of attack. These tech-
niques undoubtedly have been made applicable by the intense color of the ferrous
complexes. This has provided a convenient method for the determination of
small concentrations. The polarographic method for the study of complex sta-
bility has been used, but not extensively. The polarographic behavior of the
organic molecules concerned is very complex and has not yet been clarified.
Potentiometric measurements have been used with some success also. The
basicity of the organic molecules has generally been determined by the usual pH
or conductometric titration or some modification thereof.

A. THE IRON-PHENANTHROLINE SYSTEM

In order to provide a means of calculating the ferrous-ion concentration in

biological fluids containing phenanthroline, Dwyer and Nyholm first determined
an instability constant for the system in 1946 (99). Earlier work had established
the existence of the trisphenanthroline ferrous ion as the intense red species
which had been used analytically. No evidence had been reported for other
species involving ferrous iron and phenanthroline. Assuming no interaction
between ferric ion and phenanthroline they attributed the change in potential,
caused by the addition of phenanthroline to a ferric-ferrous system, to the
982 BRANDT, DWYER, AND GYARFAS

formation of the trisphenanthroline ferrous ion. The assumption appeared

reasonable in the light of the well-known fact that the blue trisphenanthroline
ferric ion could not be formed by direct mixing, but had to be prepared by oxida-
tion of the corresponding ferrous complex. Assuming constant ferric-ion concen-
tration, the Nernst expressions for the system prior to and following the addition
of phenanthroline were combined. In the resulting equation the instability con-
stant (Kr) for the reaction

[Fe(phen)3]++ <=> Fe++ + 3 phen (1)

is given by

Kr =
£/0·059 .(73ß (2)
where AE is the change in potential of the system due to the reduction of the
ferrous-ion concentration resulting from the addition of the phenanthroline, and
CB is the concentration of the free phenanthroline base. The experiments were
carried out around pH 2, necessitating the determination of the acid dissociation
constant, Ka, for the reaction
phenH+ phen + H+ (3)
as given by
&phen
*
( *
(4)
_

(ZphenH+

Potentiometric titration established the fact that the phenanthroline molecule

reacts with only one hydrogen ion. Measurement of the pH of mixtures of phen-
anthroline and hydrochloric acid gave a value for pKa of 5.2 at 25°C. Concentra-
tions were substituted for activities. This value is comparable to a value of 4.8
previously obtained by Albert and Goldacre (1). Neglecting activities in the
calculation of the instability constant an average value of pKr 17.1 ± 0.8
=

was obtained.
Lee, Kolthoff, and Leussing later conducted a thorough investigation of the
iron-phenanthroline system (166, 178, 179). They used both the equilibrium
and the kinetic approaches to the problem. Potentiometric titration confirmed
the 1:1 ratio of phenanthroline to hydrogen ion. The two nitrogen atoms are
apparently too close together (approximately 2.5 A.) to permit the presence of
two hydrogen ions in solutions of moderate acidity. The decomposition pressures
of phenanthroline hydrate give a AH for each hydrogen bond to a nitrogen of
7.15 kcal. (118). No evidence has been presented concerning the possibility of
another proton being added in strongly acidic solution. The extrapolation to
infinite dilution of a plot of paH+ (in a 1:1 mixture of phenanthroline
and phenanthrolium ion) versus the square root of ionic strength gave a value of
4.77 for pKa at 25°C. From this value the activity coefficients of the
phenanthrolium ion were obtained at various ionic strengths. The calculation of
the degree of hydrolysis of the phenanthrolium ion by conductometric titration
with hydrochloric acid gave an independent check of pKa 4.96. This value was
=

used in subsequent calculations.

 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 983

In solutions acidic than pH 3, the phenanthroline is essentially present

more
as phenanthrolium ion, and the equilibrium
[Fe(phen)3]++ + 3H+ <=± Fe++ + 3phenH+ (5)
can be followed by using known amounts of iron, acid, and phenanthroline and
measuring the color of the [Fe(phen)3]++. The activity constant
3
1 _
CtFe++ 'QphenH+
OFe(phen)3 + +
·

may then be evaluated, using known and estimated activity coefficients and the
assumption that aFe++ ctFe(Phen),++. Wide variations of the concentrations of
=

the three starting materials gave a series of constants averaging 4 X 10-7 at

25°C. Calculation of K; for reaction 1 is given by

aFe++'a"hen
Kt =

UFe(phen)g+ +
=
K\Kaf (7)

Using the average value of Kl, a value of pKt 21.3 at 25°C. was obtained (178).
=

This constant was confirmed by an entirely independent method involving the

determination of rate constants for the formation and dissociation of ferroin
(179). The addition of ferroin to acid solutions 0.01 N and stronger results in
dissociation of the complex. The rate of dissociation can be followed by the
decrease in color intensity, and a plot made of log co/c versus t, where co is the
initial concentration and c is the concentration at a given time, t. The results
from 0.05 and 0.5 M sulfuric acid solutions give the same straight line, demon-
strating a first-order reaction independent of hydrogen-ion concentration. The
use of 0.005 M sulfuric acid gives the same straight line during the first portion
of the decomposition. However, the appreciable formation of ferroin at this
acidity causes the plot to deviate markedly as equilibrium is approached. The
acid-independent rate constant of dissociation is 0.0045 min.-1, which agrees
with the value calculated from the published exchange data (246).
The formation of ferroin is quite slow in solutions more acidic than pH 3.
By proper choice of conditions its formation is readily followed by measurement
of the ferroin color. By utilizing concentrations of iron, phenanthroline, and
hydrogen ion which are large relative to the amount of ferroin formed, it is
possible to obtain a forced zero-order reaction. The kinetics may then be treated
as a zero-order formation followed by a first-order dissociation. The integrated
rate equation for this type of situation is
(1 —
e
A-r--
.

c =
(8)
f^diss.

where A is a constant, c is the concentration of ferroin, and t is the time. The

data obtained by following the formation spectrophotometrically gave a series
of straight lines when c was plotted versus (1 é~kd>“-*)/kdiM.· The slope, A, —

of the lines was found to be proportional to [Fe++] and [phen]3. This relationship
is most readily interpretable on the basis of the assumption of three equations,
984 BRANDT, DWYER, AND GYARFAS

Fe++ + phen Fe(phen)++ (9)

Fe(phen)++ + phen *=* Fe(phen)2++ (10)
Fe(phen)2++ + phen —>
Fe(phen)3++ (11)
in which equations 9 and 10 represent rapidly established equilibria, and equation
11 is the rate-determining step. The equation representing the rate of formation
then becomes
A =
fc'[Fe(phen)2++][phen] (12)
Having assumed that equations 9 and 10 reach equilibrium rapidly, the equi-
librium constant for the formation of Fe(phen)2++ from Fe++ and phen can be
expressed by

Fe(phen)2++ =
Aeq.[Fe++][phen]2 (13)
Combining equations 12 and 13 yields
A =
&f0rm.[Fe][phen]3 (14)
from which fcf0rm. =
1.3 X 1019 min.-1 at 25°C. Activity coefficients were used
only in the calculation of the concentration of free phenanthroline from the
acid dissociation constant. Calculation of pK¡ from the mass action law

K, =
^
Morm.
(15)

where fcdiBs. and ¿form, are the rate constants of dissociation and formation,
gave a value of 21.5 (179), which compares very favorably with the value of
21.3 calculated from equilibrium data (178). Both values contain a factor of
Kl These results differ considerably from previous work (298), but the data
have been confirmed (34).
The stepwise equations (9, 10, 11) for the formation of ferroin indicate two
additional iron-phenanthroline species: namely, Fe(phen)++ and Fe(phen)2++.
The existence of such intermediate complexes would certainly be expected in the
light of the work of Bjerrum on stepwise formation of ammine complexes. Early
colorimetric investigation of the ferrous-phenanthroline reaction had indicated
only Fe(phen)3++ in solution under the usual analytical conditions. This is es-
sentially true; however, the existence of Fe(phen)"1-1" under special conditions
was observed (178, 179) and its characteristics carefully evaluated (167). The
ion Fe(phen)2++ apparently does not exist in appreciable concentration in solution.
In the study of the pseudo zero-order formation of ferroin (179), the straight-
line plots of c versus (1 e~kdiM'i)/kdi»s. did not pass through the origin, indi-
—

cating the existence of some colored species formed immediately upon mixing
the reactants. The color is yellow upon observation. By working in solutions in
which the acidity and the ferrous-ion concentration are high relative to
phenanthroline, the 1:1 complex can be prepared and is stable for several hours.
The absorption spectrum has a broad maximum between 400 and 450 µ, with
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 985

a possible peak at 440 µ (167, 179). The molar absorbance index is approxi-
mately 370 (167). Variations of absorbance with changing concentration of the
reactants confirmed the existence of the mono complex.
In solutions of pH 4.5-5.5 it can be demonstrated that if excess ferrous ion is
present, the theoretical amount of ferroin for the amount of phenanthroline
present is not completely formed. Assuming that the incomplete ferroin forma-
tion is due to the presence of [Fe(phen)]++, the instability constant of the latter,
Kh, of reaction 9
(Fe^Xphen)
n (16)
[Fe(phen)-^]
can be determined using the known instability constant Ki for the tris complex.
The combination of equations 7 and 16 gives
Fe(phen)++ _
\ß (Fe++) T3
Fe(phen)3++ K,, L[Fe(phen)3++]J
Substitution of experimental values in equation 17 gave pKh 5.9 (167). =

Utilization of the instability constants for the mono- and trischelates permits a
calculation of the relative amounts of the different species involved in the ferrous-
phenanthroline system (167). The graphical data indicate the absence of any
appreciable quantities of [Fe(phen)2]++. This makes the ferrous-phenanthroline
system quite different from the corresponding ammine and ethylenediamine com-
plex systems (26). In the latter two, a stepwise formation is observed in which
each intermediate is the predominant species at some point during the course of
adding increasing amounts of the complexing agent to the solution. The
Fe(phen)44" is never very predominant (not over 3 mole per cent) and
the Fe(phen)2++ has yet to be characterized. The difference between these systems
cannot be attributed to a stability due to chelation, yet some property of the
arrangement of three phenanthrolines around the ferrous ion gives an abnormally
enhanced stability of the 3:1 species over the 2:1 and 1:1.
The ferric-phenanthroline system is somewhat more difficult to study. The
reaction products resulting from direct mixing of the reagents have thus far
been somewhat incompletely characterized (119, 133, 199, 289). The trisphenan-
throline ferric ion is formed only by oxidation of the corresponding ferrous
complex. Knowing the instability constant for ferroin and the oxidation-reduc-
tion potentials of the ferrous-ferric and ferroin-ferriin systems, it is possible
to calculate values for the instability constant of ferriin. Another approach
involves the measurement by potentiometric means of the equilibrium repre-
sented by

Fe4-4 + [Féchenla4-44"] Fe+++ + [Fefphenls]"44" (18)

These data may then be utilized in the equation

'

O'Fe'l· 3 Ctphen ttre+S &Fe(phen)3 + +

'

K iferriin •
K ^ferroin (19)
OFe(phen)3 + + + CtFe++ OFe(phen)3 + ++
*
986 BRANDT, DWYER, AND GYARFAS

The calculation gave values of the instability constant of ferriin which decrease
with increasing hydrogen-ion concentration (178).

P^iferriin HiSO. PKl ferriin H,SO«

M M
14.1 0.05 15.5 2.0
14.6 0.5 20.9 8.0
15.0 1.0

The hydrogen-ion dependence is evidenced also in the determination of the rate

of dissociation of ferriin (179). The rate was followed potentiometrically in a
cell containing Fe++, Fe4^, Fe(phen)3++, and Fe(phen)3+++. The specific rate
constant for the first-order dissociation varied from 0.0137 min.-1 in 0.05 M
sulfuric acid to less than 0.002 min.-1 in 2 M sulfuric acid. The direction of the
hydrogen-ion dependence of the instability constant is opposite to that which
would be predicted if a ferric sulfate complex were forming or if the activity
coefficient of ferriin were larger than that of ferroin. The data can be explained by
assuming a stabilization of ferriin by the reaction
[Fe(phen)3]+++ + H+ [FeH(phen)3]++++ (20)

However, this species has not been demonstrated conclusively, and, owing to
lack of information concerning the mechanics of the formation of such an ion
and the driving force for it, the possibility of the existence of another explanation
must be allowed.
A yellow species formed by the decomposition of the blue Fe(phen)3+++ ion
has been identified as having the empirical ratio of Fe2(phen)3. The structure has
not been explained (133). The possibility of a phenanthroline bridge would seem
quite unreasonable.
The trisphenanthroline ferrous ion undergoes exchange with ferrous ion slowly,
although the complex is diamagnetic and such behavior would not be predicted
(246).
B. THE IRON-BIPYRIDINE SYSTEM

The iron-bipyridine system is, in general, quite similar to that of iron-

phenanthroline. It has been studied by the same techniques: namely, equilibrium
and kinetics. However, the results of different investigators do not agree in
several details. The system also possesses a remarkable difference in hydrogen-ion
dependence compared to phenanthroline. The validity of the difference seems
well established; however, the lack of explanation of the observed phenomenon
in either case prevents a sound judgment of the credibility of the difference.
The instability constant for Fe(bipy)3++ was first determined by Dwyer and
McKenzie (98), using the previously described potentiometric method (equa-
tion 2), to be K¡ =16.4 ± 0.4. This value agrees reasonably well with later
values. The equilibrium reaction analogous to equation 5 can be measured and
the instability constant KrP,(bipy,,++ may be calculated from
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 987

(Fe++) (bipy)8
(21)
Fe(bipy)s++
by utilization of the acid dissociation constant for the conjugate acid of the
base. The calculated value of P-K/re(bipy)l++ is 17.1, using pKa =
4.33 at 25°C.
(µ =
0.025) (19). The heat of formation is calculated to be —24.3 ±1.3 kcal.
Other closely agreeing values of these constants have been reported (17, 18,
170, 171, 172).
The kinetic aspects of the iron-bipyridine system resemble the iron-
phenanthroline system in that the formation is rapid above pH 5, but in more
acidic solution (pH 2-3) the rate is slower and permits easy measurement. In
the presence of large excesses of bipyridine the reaction is first order with respect
to ferrous ion, and the formation constants are proportional to (bipy)5 or to
1/[H+]5 (17). This confirms the fact that the free base is reacting and not its
conjugate acid. By using comparable concentrations of ferrous ion and bipyridine
the rate of formation may be measured at pH 2-6. The data fit a fourth-order
reaction expression

-[Fe(^y)8++1
=
fc0[Fe++][bipy]8 (22)
d<

where fa is the observed velocity constant (pH dependent). The fourth-order

kinetics fit a stepwise formation
K!
Fe++ + bipy ± Fe(bipy)++ (23)

Fe(bipy)++ + bipy Fe(bipy)2++ (24)

*/
Fe(bipy)2++ + bipy <=> Fe(bipy)3++ (25)
kb

where equations 23 and 24 are rapidly established equilibria with constants Ki

and K2, while equation 25 is the slow reaction having rate constants kf and fa
for the forward and back reactions. The rate of formation is apparently inde-
pendent of temperature.
In solutions more alkaline than pH 1.3, the fourth-order rate equation for the
formation of Fe(bipy)3++ involves an acid-dependent term and an acid-
independent term. The rate-determining steps are assumed to be the reaction of
Fe(bipy)2++ withbipyH+ and bipy, respectively (171, 172). The rate of formation
goes through a maximum with increasing concentration of ferrous ion. If it is
assumed that this inversion is due to a rapid formation of a species such as
Feibipy)4-1", which effectively lowers the ferrous-ion concentration, it is possible
to get satisfactory representation of the kinetics. Attempted explanation of the
data by the formation of Fe(bipy)2++ gives unsatisfactory agreement.
The dissociation of Fe(bipy)3++ is first order and is dependent upon the hydro-
gen-ion concentration. It also varies with the type of acid used (17), being dif-
ferent in perchloric acid than in sulfuric or hydrochloric acid. For the considera-
988 BRANDT, DWYER, AND GYARFAS

tion of the dissociation reaction, Krumholz (171) developed an equation which

involves the hydrogen-ion dependence in solutions of [H+] =
X 10~2 M.

d[Fe(b!^y)s++]
at
=
K + fcdl[H+]][Fe(bipy)3++] (26)

This indicates the presence of an acid-dependent dissociation combined with

the usual acid-independent mechanism. Since ki1 decreases with increasing
[H+], it is assumed that the rate-determining step becomes a monomolecular
decomposition of some intermediate species, such as FeH(bipy)3+++, rather
than a bimolecular reaction between H+ and Fe(bipy)3++. Equation 26 may be
expressed as

d[Fe-^-y--3+--
=
WFe(bipy)3++] + fcá2[FeH(bipy)3+++] (27)

From this equation the instability constant for the protonated ion, FeH(bipy)3+++,
going to Fe(bipy)a++ is K¡ 0.6. This reaction is apparently independent of
=

temperature, while the activation energies of the acid-dependent and acid-

independent reactions are 26 and 29.5 kcal., respectively.
The rate constants for the formation and dissociation give a value of pK¡
= 17.5 (µ 0.33) for the instability constant of Fe^ipy)^ (171). Values from
=

equilibrium measurement agree wrell.

The intermediate species required for satisfactory explanation of the kinetics
has been investigated separately. Thus, this yellow species has been shown to be
Fe(bipy)++, and it has been spectrophotometrically characterized (Xmax. 440 =

µ) (19). The instability constant has been determined by equilibrium measure-

ments and by detailed interpretation of the kinetics of formation of Fe(bipy)3++.
The values obtained for the pK¡ of Fe(bipy)++ are 4.2 (18, 19) and 4.4 (170,
171, 172). The value, KiTe(bipy)!++ > 10-6, listed for the instability constant of
Fe(bipy)2++, was obtained by approximation and represents a minimum value
attainable (18).
The instability constant for the Fe(bipy)3+++ ion, calculated from oxidation-
reduction data, is approximately pKT 12 (19). The rate of dissociation of the
—

complex appears to be acid-dependent.

From the foregoing discussion it is apparent that a protonated complex,
FeH(bipy)3++-+, quantitatively fits the available experimental data. However,
the explanation does not fit the observed rate constants in high acidities (1.0 N
hydrochloric acid) nor does it offer an explanation of the dependence of the rate
constants of dissociation on the type of acid (18). It may be unreasonable to
expect it to explain the latter fact.
The trisbipyridine ferrous ion undergoes slow exchange with Fe++ in a similar
manner to the phenanthroline complex.

C. THE IRON-TERPYRIDINE SYSTEM

The instability constant of the bisterpyridine ferrous ion determined by equi-

librium measurements is pKT 18 (36). The terpyridine molecule reacts with
=

two hydrogen ions with an acid dissociation constant for the reaction
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 989

trpyH2++ +± trpy + 2H+

of 7.1. The stepwise constants for the dissociation have not been determined. An
estimation of the constant for Fe(trpy)2+++ from the oxidation-reduction po-
tential of 1.10 in 0.1 F sulfuric acid gives pK¡=
11.

D. A COMPARISON OF THE IRON SYSTEMS

The ferrous-phenanthroline, bipyridine, and terpyridine complexes exhibit a

number of similarities. More important, they have points of non-similarity which
demonstrate the need for considerable careful investigation directed towTard an
explanation of some of these observed phenomena. The terpyridine system has
not received an amount of study comparable to the other two reagents. This
fact, coupled with the different type of coordinating system involved, removes it
from much of the discussion.
The bipyridine and phenanthroline systems are very similar in the stepwise
formation of metal complexes. In each case, for example, a yellow 1:1 complex
has been found and studied under very similar conditions. However, in both
cases the 2:1 complex is apparently unstable with respect to the other two and
never exists in appreciable concentration in solution. By comparison, a 1:1 spe-
cies has not been reported in the terpyridine system.
The explanation of the effect of hydrogen ion on the phenanthroline and bi-
pyridine systems seems to present the greatest challenge. It appears to be in-
volved in the dissociation of the ferric complexes of both, but only with the
ferrous complexes of bipyridine. The complete explanation of this phenomenon
should prove valuable. The existence of species such as FeH(phen)3+4'++ intro-
duces many questions concerning the position and means of attachment of the
proton and its effect upon the properties of the molecule. Isolation of salts of
both the protonated and non-protonated complex should be possible. The ex-
tent to which this type of behavior is typical of metal-phenanthroline and
metal-amine complexes in general should be explored.
One point which is apparent from the compilation of data is the discrepancy
of the instability constant of ferroin based on potentiometric measurements.
This presumably is due to an uncorrected interaction between ferric ion and phe-
nanthroline. However, the corresponding value for the bipyridine complex agrees
quite well with the results obtained by other methods. Thus the need for further
study of the reaction of ferric ion with these molecules is indicated. The state
of knowledge of the ferric complexes in general leaves considerable to be desired.
The products of direct mixing and the explanation of the failure of completely
complexed ions such as Fe(bipy)3+++ to form directly will require further study.
The nature of the compounds formed upon dissociation of the trisferric com-
plexes should prove interesting. Such information might well fit into the prob-
lem of the photochemical reduction and x-radiation of the ferric complexes (240).
Two differently colored species of the ferric-terpyridine complex have been
demonstrated in different acidity (36) and point out the desirability of knowing
the influence of hydrogen ion on these species. Only as these factors become
known will it be possible to draw sound and thorough conclusions regarding the
significance of these important, but yet quite isolated, pieces of data.
990 BRANDT, DWYER, AND GYARFAS

E. OTHER METAL SYSTEMS

The investigation of the stability of complexes other than those of iron has
progressed slowly. In the case of other metals, the existence of colored species
is often completely lacking; even when available, such colors seldom approach
the intensity of that of ferroin. Although complexes of the phenanthroline type
of compounds are known to form with a large number of metal ions (see Sec-
tion II), only a few have received quantitative measurement of their stability.
The system zinc-phenanthroline was investigated by Kolthoff (168). The
formation of a zinc complex had been indicated by the interference of zinc
in the colorimetric determination of iron, and several salts of complexes had
been reported in the literature. Zinc forms three well-defined species. The mono-
phenanthroline zinc(II) ion is the main zinc species formed when zinc ion is
added to solutions of ferroin and equilibrium is established. The instability
constant calculated using the known values for the iron system gives
pA/Zn[Pkear-
=
6.43. The corresponding constants for the 2:1 and 3:1 com-
plexes were obtained from measurements of the amount of phenanthroline which
could be extracted from solutions of zinc and phenanthroline by chloroform.
In this manner the average number of phenanthroline molecules retained per
zinc ion is obtained. Using a modified Bjerrum-type calculation the values
P-Krzn(Phen>,++
=
12.15 and P-K/Zn(phen)!++ = 17.0 are obtained. Confirmation was
obtained for these values by conductometric titration of zinc with phenan-
throline (168) and equilibrium determinations using ultraviolet absorption meas-
urements (173).
Polarographic measurements of the cadmium-phenanthroline and cadmium-
bipyridine complexes give the following values for pKT: Cd(bipy)3++, 10.5;
Cd(bipy)3++ in 28.5 per cent alcohol, 10.0; Cd(phen)s"H", 15.2; Cd(phen)2++ in
28.5 per cent alcohol, 13.2 (68). The apparent loss of one phenanthroline in
28.5 per cent alcohol may possibly be due to the effect of the ethanol on the
availability of the metal ion. The quoted values would require an extreme dif-
ference in the attraction of the cadmium for the first two organic molecules as
compared to the third. The proportional decrease of stability in alcohol of the
Cd(bipy)3++ system is only slightly less than the change in going from Cd(phen)3++
in water to Cd(phen)2++ in 28.5 per cent alcohol.
A Bjerrum-type approach has been reported for the cupric-phenanthroline
complex, indicating the existence of the trisphenanthroline cupric ion. The in-
stability constants of the three stepwise additions of phenanthroline to cupric
ion are 6.30, 6.15, and 5.50. The overall instability constant pK¡ =
18.0 (237).
This compares closely with the value of pK¡ =
17.0 obtained for the trisbi-
pyridine cupric ion from polarographic measurements (223). The proximity of
the bipyridine and phenanthroline values does not correspond to the iron or
cadmium system, but since values are not available for the complexes of both
compounds with any other metals, the significance of this fact is purely con-
jecture. The polarographic reduction of the trisbipyridine cupric ion goes first
to the bisbipyridine cuprous ion and then to the metal. Both reduction steps
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 991

are reversible, and the bisbipyridine cuprous ion has a value of p= 14.2
(223). Thus the individual stepwise stability would seem to be greater for the
cuprous complex in contrast to the cupric. It is unfortunate that the copper(II)-
phenanthroline complex is not suitable for similar investigation.
The nickel systems with bipyridine and phenanthroline have received some
attention (13, 284). This has been directed mainly toward demonstrating the
existence of the 1:1, 2:1, and 3:1 complexes and determining the kinetics of the
decomposition and racemization. The instability constant for the trisphenan-
throline nickel(II) complex is pKi = 18.3 ± 1.1 (65).
Exchange reactions have been studied between trisbipyridine nickel (II) and
nickel(II) (157) and between trisbipyridine vanadium(II) and vanadium(III)
(162). In the case of nickel, exchange is quite slow. With vanadium the reaction
goes to the extent of 40 per cent in zero time and then stops.
These entries demonstrate the difference in the phenanthroline type of com-
plexes from other series already investigated, e.g., acetylacetone and ethylene-
diamine. In such systems the usual order or stability of the metal complexes is
Pd > Cu > Ni > Pb > Co > Zn > Cd > Fe > > Mg. Although the data are too in-
complete to permit conclusive statements, this order is obviously quite different
for the phenanthroline-type systems. Iron is greater than copper, nickel, and
zinc, and the latter three are about equal. Zinc and cadmium are in proper
sequence, although separated more widely than is common. The apparently
marked deviation of this system from the regular order makes the determina-
tion of other constants in the series of interest. The correlation of the stabilities
with physical properties such as the second ionization constant will become
quite difficult. The existence of a major deviation from the expected regular be-
havior will necessitate an expansion of the present theories. As data become
available, it will become necessary to determine whether the system will fit
into the comparison pattern established by Fernelius (112).
VII. Substitution and Steric Considerations
Substitution in the molecules of the parent bases might well be expected to
modify somewhat the stabilities and colors of the iron(II) coordination com-
pounds, and also to shift the redox potentials, by alteration of the relative sta-
bilities of the compounds in the bivalent and tervalent states, but not lead to
any profound changes. Much the same kind of modification would be anticipated
with other metal derivatives. It is now well established that for a given metal
there is a relationship between the proton-binding capacity of a ligand and the
stabilities of the complexes (27, 192). For example, substitution of the electro-
philic bromo or nitro groups in the 4,4'-positions in bipyridine or the 4,7-posi-
tions in phenanthroline should merely lower the base strengths and hence reduce
the stabilities of the metal complexes.
In the same way, methyl substituents in these positions should increase the
stabilities. This property of several 5-substituted phenanthrolines has been de-
termined. The stability of the corresponding ferrous complexes has also been
evaluated and a straight-line relationship between pKt and pKa demonstrated.
992 BRANDT, DWYER, AND GYARFAS

The stabilities were determined by kinetic and equilibrium measurements (34).

The variation of the stability of the hydrogen bond in the hydrates of two sub-
stituted derivatives has been determined (118).
These results confirm the direct relationship between basicity and stability
in a series of substituted phenanthroline complexes, but it is not possible to
include bipyridine and terpyridine complexes in such a direct comparison. In
the latter case a different number of molecules and hydrogen ions is involved.
In order to include terpyridine in the comparison, a concept of the “average
basicity per nitrogen” was introduced (36). This value, obtained by dividing the
total pKa of the molecule by the number of nitrogen atoms, permits a direct
comparison.
This comparison indicates the failure of the basicity concept to explain more
than the variation of stability with substitution in a given molecule with the in-
formation now available. The more fundamental variation, that between similar
molecules, is not explained and will require further investigation. The variation
of double-bonded character in the chelate ring or geometric strain effects might
be logical points of attack.
Following the synthesis of a large number of variously substituted phenan-
throlines (51, 52, 53, 54, 55, 56), qualitative investigation of the effect of sub-
stitution of methyl groups in the phenanthroline ring upon the stability of the
corresponding ferrous (35) and cuprous (194) complexes has been made. In the
ferrous system, the pH of precipitation of the hydroxide was used as the criterion
of stability. The results indicated that a methyl group in the 4- and 7-positions
caused the greatest increase in stability, with the 5,6-positions somewhat less
effective. The 3,8-positions caused some apparent decrease in stability. These
effects are compatible with the -electron distribution in the molecule (183, 241).
The fact that the 2,9-positions cause effects which are not consistent with an
explanation due to basicity will be discussed later.
The cuprous system was studied by comparing the pH at which extraction
into isoamyl alcohol first occurred as the acidity of the solution is decreased.
Here the 2,9-positions cause extraction at by far the lowest pH (1.0). The 3,8-
and 5,6-positions were without much effect (pH 3-4), and the 4,7-substitutions
caused marked increase in the pH of extraction (pH 7) (194). Since these results
are directly opposite to those obtained with the ferrous system it would appear
that the two investigations were not both sensitive to stability. Since the factors
causing extraction of a precipitate are not thoroughly understood, it might ap-
pear that the extraction phenomena were not based primarily upon stability.
In the main, these general principles have been found to apply (vide infra),
but substitution in the nitrogen-adjacent positions in either molecule and in the
3,3'-positions in bipyridine apparently makes such profound changes in the
chelating abilities of the molecules that, as far as iron is concerned, either no
complex is formed or the complex is of very low stability. Loss of function is
related to the position of the substituents and their sizes and not to their electro-
philic properties. It thus appears that neither the base strength nor the partial
double-bond character of the nitrogen-metal bond is involved, though few quan-
titative data are yet available.
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 993

TABLE 1
Effect of substitution on thecolor intensity of chelate complexes

Base Absorption aM Base Absorption aM

Maximum Maximum

µ µ

3,3'-Dimethylbipyridine .. . 526 1770 5, S'-Dimethylbipyridine. . . 510 8300

4,4'-Dimethylbipyridine .. . 529 8470 2,2'-Bipyridine............ 522 8650

The orange-colored tris-3,3'-dimethyl-2,2'-bipyridine iron(II) cation is stable

in the limited pH range of 4-6 as compared with the pH range of 3-8 for bipy-
ridine itself and various dimethyl-substituted derivatives (48). The effect of
substitution on the color intensity is shown in table 1, in which the molar ab-
sorptivities, aM, at the wave length of maximum absorption are compared (48).
With the larger substituents present in 3 ,3'-dicarboxy-2,2'-bipyridine, no
color reaction occurs and presumably no iron complex is formed (30, 295). It
can readily be seen that substituents in these positions, unless quite small, as
with the hydrogen atoms in the parent base, interfere sterically with each other
and tend to rotate the pyridine rings about the linking single bond in the 2,2'-
position. The chelating ability is thereby greatly reduced by the decreased sta-
bilizing resonance in the donor, which requires coplanarity of the rings. More
important, probably, is the introduction of strain in the metal-nitrogen bonds,
which are forced out of their normal direction in the plane of the pyridine
rings (48).
Buckling of the benzenoid ring in 1,10-phenanthroline should occur with hy-
drogenation in the 5,6-positions and this would move the pyridine rings out of
coplanarity. As a result, the iron chelate compound would not be expected to
form (45).
A large number of nitrogen-adjacent substituted bipyridines have been pre-
pared (42) and tested for the formation of a red color with iron(II) salts. A single
substituent in the 6-position, such as CH3, Br, CN, N02, NH2, or COOH, ap-
pears to lower the color intensity of the reaction considerably. No information
is available upon base strengths or the relative stability constants. Disubstitu-
tion (formula XXVII) with one exception completely inhibits the color reaction.
6,6'-Dicarboxy-2,2'-bipyridine gives an orange color, but this is undoubtedly
due to chelation through the carboxyl group and nitrogen, as occurs in picolinic
acid.

R R R R
XXVII XXVIII
R =
CH3, Br, CN, COOH, NOa, NH2.
Since reagents normally attack phenanthroline in the benzenoid ring, 5 (or 6)-
derivatives are the easiest to obtain and a comparable series of 2- and 2,9-
994 BRANDT, DWYER, AND GYARFAS

derivatives (XXVIII) has not yet been tested. However, the tris-2-methylphe-
nanthroline iron(II) salts are of an orange color (147, 227) and recently have
been shown to be of low stability. Though 2-methylphenanthroline is a stronger
base (pKa =
5.42) than phenanthroline itself (pKa=
4.92), it forms much weaker
complexes, pKi =
10.8, as compared with pXj- =
21.4 for Fe(phen)3++. The
presence of the second methyl group in 2,9-dimethyl-l, 10-phenanthroline com-
pletely inhibits the iron color reaction (52, 147, 194). The only other nitrogen-
adjacent substituted derivative prepared, 2-chloro-l, 10-phenanthroline, has
been found to yield an orange-yellow iron (II) compound, but quantitative sta-
bility data are still unavailable (127, 147).
The following substances that can be regarded as nitrogen-adjacent substi-
tuted bipyridines have also been reported (139, 256) not to give a color with
iron(II) salts: 2-(2-pyridine)quinoline (XXIX); 2-(2-quinoyl)quinoline (2,2'-bi-
quinoline) (XXX); 3-methyl-2,2'-biquinoline (XXXI); 3-methyl-2-(2-pyridine)-
quinoline (XXXII); and l-(2-quinoyl)isoquinoline (XXXIII).

XXIX XXX
ch3 ch3

N- / % %-x X-
XXXI XXXII

Smirnoff (256) suggested that in XXIX and XXX the benzene ring with its
own system of conjugated double bonds interfered with the donor properties
of the nitrogen atoms, and Feigl (109), in support of this view, cited the weaker
base strength and the smaller addition capacity of quinoline as compared with
pyridine. It was concluded that such considerations were invalid for the 6,6'-di-
substituted bipyridines and that here a configuration of the pyridine rings other
than in the parent base must be involved.
Since 2,9-disubstitution in the phenanthroline system, in which the pyridine
rings are fixed, also apparently inhibits the color reaction, this explanation is
untenable.
It is evident that whatever explanation is adduced to explain the inhibition
of the iron(II) reaction must apply in some measure to the other metals, and
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 995

it becomes of interest to examine how general is the lack of reactivity of these

substances. The 6,6'-disubstituted bipyridines and also substances XXIX to
XXXIII apparently do not lack all coordinating ability. It has been found that
biquinoline (XXX) gives the complexes Co(biqn)Cl2 and Cu(biqn)Cl2, which
have been formulated as typical monomeric 4-covalent compounds (37, 138).
Chelate derivatives containing more than one molecule of the base were not
isolated, but when dichloro-biquinoline copper(II) was reduced in the presence
of excess of the base, a deep purple solution resulted, which from spectrophoto-
metric studies was concluded to contain the bisbiquinoline copper(I) ion. All
of the substances XXVII (R =
CHS) and XXIX to XXXIII have also been
found to form deeply colored copper(I) derivatives, sometimes stable to dilute
acid and considered to contain two molecules of the chelate (139). Qualita-
tive drop tests were applied with XXVII (R =
CH3) and with XXIX
to XXIII, using most of the metal ions in dilute aqueous solution; no color
change was observed. This cannot be taken, however, as good evidence that
metals other than copper(I) do not form chelate compounds. Many if formed
would be colorless, as with zinc or cadmium, or weakly colored, as with copper(II)
or nickel. Ruthenium and osmium, which form the most stable of all of the com-

plex cations with bipyridine and phenanthroline, would not have reacted even
with these bases under the experimental conditions used, owing to the slow rate
of complex formation.
This lack of precise information necessarily makes much of the discussion that
follows somewhat speculative. Hoste was of the opinion that the high position
of copper in the sequence of stabilities of metal complexes could account for the
apparently unique copper(I) reaction, but it should be pointed out that this
series refers to copper(II), and as yet insufficient data are available to fix the
position of copper(I).
For the bis-2,2'-bipyridine copper(I) ion (223), pX7 =
14.2, and for the
tris-2,2,-bipyridine copper(II) ion (223), pKr=
17.8. Since Kz is probably quite
small in accordance with the usual weak octahedral coordination of copper(II),
it may be inferred that for Cu(bipy)2++ pKt =
15. Comparison of these values
with p =
13 for the Fe(bipy)2++ ion (19) suggests, if anything, that the
chelate molecules that give copper(I) complexes should also react with copper(II)
and iron(II).
If it is assumed from the incomplete evidence available that the copper(I)
reaction is, in fact, unique, this anomaly can be explained on the basis of the
tetrahedral coordination of copper(I) in contrast to the planar or octahedral
coordination of the transition elements, and the steric factors introduced by
nitrogen-adjacent substitution.
The importance of steric factors in the stabilities of metal complexes has been
emphasized (146) and attention drawn to the failure of 2-methyl-8-hydroxy-
quinoline (XXXIV) to give a precipitate of the aluminum trischelate derivative
at pH values below that necessary to precipitate aluminum hydroxide (198).
A series of 2-substituted derivatives of 8-hydroxyquinoline and also 1-hydroxy-
acridine (XXXV) behaved similarly.
996 BRANDT, DWYER, AND GYARFAS

AAA
CH3
OH
VZZxy
OH
XXXIV XXXV
The similarity in the steric considerations between substances XXVII to
XXXIII, and XXXIV and XXXV is obvious. Substitution in 8-hydroxyquinoline
adjacent to the hydroxyl group, as might be anticipated, has little effect on the
chelating tendency (291).
Even large, strongly basic substituents in the 7-position, such as dimethyl-
aminomethyl, diethylaminomethyl, and piperidylmethyl, while they raise the
pH of complex precipitation do not otherwise modify the reaction (238). Sub-
stantial support of a more quantitative nature (14, 15) for the effect of substi-
tution comes from studies with di-X-substituted ethylenediamines. It will be
apparent from table 2 that the addition of even the first ligand to copper or
nickel is less readily effected than with ethylenediamine itself and the difficulty
increases, as reflected in the smaller values of log K\, with increasing size of the
substituent. Addition of the second and, with nickel, of the third ligand is even
more difficult. That these are essentially steric effects is demonstrated by the
relatively small effect of substitution on the proton-binding capacity of the
molecule (27).
Essentially the same factors are involved in the reactions between pyridine
and 2,6-lutidine with the Lewis acids boron trifluoride and trimethylboron.
Though the latter base is the stronger, it forms the weaker addition compounds
with boron trifluoride and no compound results with trimethylboron (40). It
can be inferred that the apparently lowered chelating ability of the nitrogen-
adjacent substituted bipyridines and phenanthrolines—and the molecules with
quinoline rings can be included in this category—is largely, if not entirely, due
to steric hindrance. It can be demonstrated easily with a model or a scale draw-
ing that the attachment of a single chelate molecule would interfere with the

TABLE 2
Chelation of N, '-dialky lethylenediamines
en di-Me-en di-Et-en di-n-Pr-en

Copper compounds; 25°C.

pXi of base.............. 10.08 10.29 10.46 10.27

log X,................... 10.55 10.47 9.30 8.79
log X2................... 9.05 7.63 6.32 4.84

Nickel compounds; 25°C.

log Xi................... 7.66 7.11 5.62 5.52

log X,................... 6.40 4.73 3.3 2.5
log X,................... 4.4 1.5
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 997

aquo groups remaining on the metal ion when the coordination is planar or
octahedral but not to any significant extent when it is tetrahedral. With the ad-
dition of the second and the third ligand, interference would occur between the
substituents themselves (14). This is strikingly confirmed by the data on the
iron(II) compound of 2-methylphenanthroline, where the successive constants
are log K\ —

4.2, log K2 =
3.7, and log Kz =
2.9 following the normal pattern,
as against log K\ =
5.9, log K% =
7, and log Kz =
8 for Fe(phen)3++.
The more open nature of tetrahedral coordination readily permits the addi-
tion of two chelate molecules without difficulty. Data on the phenanthroline
and bipyridine complexes of zinc (168) and cadmium (68) indicate strong coor-
dination, and it would not be surprising if these metals formed tetrahedral com-
plex cations of appreciable stability with these sterically hindered bipyridines
and provided yet another instance of the inversion of the order of stability of
metal complexes (148, 186, 187, 188).
VIII. Optical Activity of Chelate Complexes
All of the phenanthroline and bipyridine metal complexes obtained in the
enantiomeric forms have been tris derivatives of the Group 8 elements:
M(phen)3++ (M Fe (293), Ru (80), Os (75), Ni (82)); M(bipy)s++
=

(M =
Fe (293), Ru (41), Os (43), Ni (202)). The resolution of bisethylenedi-
aminephenanthroline cobalt(III) chloride has been reported (150).
Resolution of the phenanthroline series is best effected through the antimonyl
d-tartrates. One of the antipodes forms a very insoluble precipitate on the addi-
tion of potassium antimonyl d-tartrate, whereas the other diastereoisomer is
very soluble. A complete separation is thus achieved in a single precipitation,
and the antimonyl tartrate radical is easily eliminated with dilute alkali or
sulfuric acid.
Werner (293) resolved Fe(bipy)3Cl2 by addition of a large amount of ammo-
nium d-tartrate to the aqueous solution, and the same method has been used
for Ru(bipy)sCl2 (41) and Ni(bipy)3Cl2. Unless both ammonium d-tartrate and
ammonium Z-tartrate are employed, only one enantiomer can be prepared in
this way. Complete separation can be effected, however, in one operation by the
formation of the iodide antimonyl tartrate. When both potassium iodide and
potassium antimonyl tartrate are added to solutions of M(bipy)3Cl2,
[M(bipy)3]3l2 (SbO tart.)4· 18H20 is precipitated and contains only one antipode
·

of the complex (81, 83).

Oxidation of the optical forms with retention of configuration can be effected
with the iron(II), ruthenium(II), and osmium(II) complexes of both chelates.
The crystalline antipodes of Fe(bipy)3(C104)3, 0s(bipy)3(C104)3, and Os(phen)3-
(C104)3 have been isolated (86, 88), and active solutions of Fe(phen)3+++,
Ru(bipy)3+++, and Ru(phen)3+++ have been prepared (79, 80, 81, 83). The
chemically and optically stable osmium bipyridine compounds have been used
to demonstrate the identity of potentials of enantiomeric redox systems (6).
In a recent review Basolo (12) has discussed the oxidation of enantiomeric forms
and drawn attention to the value of an experimental method for the study of
998 BRANDT, DWYER, AND GYARFAS

TABLE 3

Specific rotations of M(phen)3(C104)2,s

Metal Wd W«ei Metal Me Méííl

Fe(II)........... 1432 Os (II)......... 0 3670*

Ru(II)........... 1340 2560 Os (III)......... 0 400
Ru(III).......... 409 1630 Ni(II).......... 1460 1740

*
[M]e46l =
34,800.

the kinetics of electron transfer based upon theuse of optical forms in two oxida-
tion states (85).
Work on the kinetics of racemization of these substances has recently been
reviewed critically (12), and though the nickel compounds undergo racemization
through a rate-determining dissociation step in aqueous solution, it has now
been shown that both dissociation and intramolecular change are involved in
aqueous alcoholic solution (16, 67). However, the rates of racemization of the
iron (III) compounds, as for the iron (II) compounds, are very much greater
than the rates of dissociation; hence intramolecular change seems to be indi-
cated (66). The osmium and ruthenium compounds are optically stable even
in boiling aqueous solution (41, 43, 80).
It will be apparent from table 3 that these substances have extremely high
specific rotations, especially when it is realized that the molecular weights are
usually in excess of 700. Though the complete curves have not been prepared,
the rotatory dispersion appears to be anomalous and very large.
The relative generic configurations are of interest. The sign of the rotation in
a particular wave length of light is, of course, of no significance. Though the
assumption of identity of the least soluble diastereoisomers has been considered
dubious (151), the large differences in solubility of the antimonyl d-tartrates
perhaps makes this simple criterion carry more weight. However, considerable
support for the principle comes from the biological experiments, discussed later,
in which the stereospecificity was shown to be very pronounced. The mouse
intraperitoneal toxicity was found to be greatest for that enantiomer in the
phenanthroline series which gave the least soluble antimonyl tartrate. This was
(+)-Ru(phen)3++, (+)-Ni(phen)3++, ( —)-Fe(phen)3++ in the Na0 line; and
Os(phen)3++ with no rotation in the Na¿ line but a negative rotation in the
HgMei line. A similar correlation existed between the solubilities of the iodide
antimonyl tartrates or the d-tartrates in the bipyridine series and the inhibitory
action in the enzyme acetylcholinesterase.
Since the biological activity of optically active metal complex compounds has
not been studied previously, this is the first instance where biological evidence
can be quoted for generic configuration.

Attempts by Pfeiffer and coworkers (230, 231, 233) to resolve the trisphenan-
throline and trisbipyridine zinc and cadmium ions were fruitless, but instead
provided the first examples of the so-called “first-order” asymmetric trans-
formation among inorganic complex compounds. This term refers to the shift
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 999

in the d l equilibrium of a labile optically active substance to favor one enan-

tiomer, and is brought about by the optical isomer of another optically stable
substance. The phenomenon is designated “first order” when it occurs without
the separation of a second phase (crystallization of one diastereoisomer), and is
to be ascribed to the free-energy difference between the diastereoisomer mole-
cules in solution. This subject, chiefly from the standpoint of organic substances,
has been discussed in an excellent review by Turner and Harris (282). With
organic compounds the phenomenon seems to be confined to substances in which
the labile and the stable optical centers are part of the same molecule, or with
salts only in non-dissociating media such as chloroform. The effect is observed
with inorganic complex salts in dilute aqueous solution and even where the
labile and stable substances are both cations. It has been found that when dl-
trisacetylacetone cobalt(III)—a neutral complex—is allowed to reach equilibrium
in the presence of d-trisethylenediamine cobalt(III) chloride one enantiomer is
favored, and a partial resolution can be effected in this way (87).
Pfeiffer found that when Zn(phen)3SC>4 was added to a solution of ammonium
bromocamphorsulfonate a very large change in rotation occurred (+0.92° —*
+0.09°). This could not be brought about by addition of zinc sulfate or phenan-
throline separately, or by the addition of hexammine zinc ion (which has a plane
of symmetry). Other optically active anions or cations could be substituted for
bromocamphorsulfonate. Similar results have been obtained (64, 230, 231, 233)
in the series MB3++ (M =
Zn, Cd, Mn, Co, Ni, Cu; B =
phen or bipy). Only
the nickel compounds have been obtained in the optically active forms and here
the movement of the equilibrium to favor one antipode was slow. Recently, the
rates of racemization of the d- and Z-forms of Ni(phen)8I2 have been studied in
the presence of bromocamphorsulfonate and cinchoninium ions (66). The rates
are different: K =
kd/ki =
0.895 in 2 per cent ammonium bromocamphor-
sulfonate at 25°C.
The term “configuration activity” (93) has been applied to the general phe-
nomenon of the interaction between optically active species in solution. This in-
volves the assignment of different activity coefficients to the dextro and levo
forms of a racemic mixture in the presence of an extraneous optically active sub-
stance. Pfeiffer’s results may then be accounted for in the following way: In the
racemic mixture of Zn(phen)8++, the activities (and concentrations) of the el-
and Z-ions are equal, but on the addition of the bromocamphorsulfonate ion the
activity coefficients become unequal. Racemization then proceeds until equi-
librium is reached, when the activities are equal but the concentrations are now
unequal; hence the solution shows a rotation due to the excess enantiomer. Pre-
cipitation of the iodide, for instance, at this stage must yield an optically inac-
tive salt because the activities of the two enantiomers in solution are the same.
The small excess of either optical form will equilibrate if sufficiently labile or
will be found in the last fractions of the precipitation. This has been confirmed
in the following manner (93). The Ni(bipy)3++ ion is optically labile with a half-
life of about 15 min. at 17°C. When the cZZ-chloride was dissolved in the presence
of either ammonium bromocamphorsulfonate, quinine bisulfate, or d-trisethyl-
 1000 BRANDT, DWYER, AND GYARFAS

enediamine cobalt chloride at 5°C. and immediately precipitated with iodide

ion, the first fractions were the d-iodide, subsequent fractions were the racemate,
and finally the Z-iodide was obtained. If the mixture was allowed to stand, it
underwent mutarotation in the levo direction, and on precipitation, as before,
gave first a large amount of racemate and finally a little of the Z-iodide.
Much still remains to be elucidated, however. The separation of Ru(bipy)3I2
by the method used for the nickel compound is only just detectable. Though
cinchoninium and ZV-methylcinchoninium ions produce a large shift in the d <=* Z
equilibrium of Zn(phen)3++, di-ZV-methylcinchoninium ion is ineffective (233).
The nature of the interaction between optically active ions in solution is a
matter for speculation. Most of the ions discussed are rather large and ion-pairs
could well exist even in the dilute aqueous solutions by the operation of van der
Waals dispersion interaction forces. Such association would be a spatial phe-
nomenon and would be expected to be favored by one antipode more than the
other. On the other hand, the differential interaction between multicharged ionic
species of the same sign cannot occur in this way. Such an effect must be long
range and could be due to van der Waals interaction of charge, if it is postulated
that associated with an antipode is an asymmetric distribution of charge. In
metal complexes such a charge configuration could be due to the distribution of
the central metal ionic charge over the ligands. With the asymmetric carbon
atom, the substitution of the concept of four different dipoles for four different
groups attached to carbon leads to a similar picture of charge configuration (93).
Probably in many systems both mechanisms of interaction are involved.
IX. Redox Reactions
The introduction of the ferroin-ferriin system as an oxidation-reduction in-
dicator of considerable merit (287) was the impetus which brought the chemis-
try of the metal-phenanthroline chelates into prominence. However, in some
respects the endeavor in the oxidation-reduction field has not been as vigorous
as in other categories. Nevertheless, several systems have been proposed as indi-
cators (74, 260, 270).
In addition to the interest in indicators, the potentials of many couples have
been investigated. The values for iron, ruthenium, and osmium demonstrate
dramatically the stabilization of valence states by chelation. The shift in oxida-
tion-reduction potential as a result of complexation is very large with all three
metals. It reaches a value of 1.3 v. with ruthenium. However, complete com-
parison is not possible, owing to the lack of formal potentials for the Os(II) —»
Os (III) system. It is certain that the shift is considerably greater with ruthenium
and osmium than with iron. This would seem to correlate with the observed
greater stability of these two (176).
The measurement of the potentials has required different methods, depending
upon the relative stability of the system. Ideally, the potential is that measured
in a system prepared from an equimolar mixture of the salts of the oxidized and
reduced forms (method I). This has been possible in very few cases, but it has
been used in the osmium-bipyridine system to study the application of the
 CHELATE COMPLEXES OF 1 ,10-PHENANTHROLINE 1001

Debye-Hiickel theory to redox systems (11). More often a potentiometric titra-

tion has been carried out and the value determined graphically (method II).
With the iron systems, it is common to include excess ferrous ion in the original
mixture. This provides a differential titration curve in which the potential of
the ferrous-ferric couple also serves as a reference point. In many cases, how-
ever, neither of the above approaches is possible, and the usual procedure is to
utilize a measured amount of the complex in the lower oxidation state and
rapidly add to it sufficient standard oxidizing agent to react with half of the
amount present (method III). The maximum value attained after mixing is gen-
erally assumed to be the potential of the couple. The values reported are col-
lected in table 4.
The data indicate some points of comparison and differences between the iron,

TABLE 4
Oxidation-reduction potentials
Iron (Il)-iron(III)
Tris Complexes
Acid Normality Bistrpy Method Reference
Phen 5-Meth- 5-Meth- 5-Bro- 5-Chlo-
Bipy 5-^°- ylphen trophen mopben rophen

None added 1.096 1.120 I (97)

0.01 1.084 1.112 I (97)
0.1 1.071 1.102 I (97)
1.10 1.06 II (35)
1.076 III (89)
0.2 1.060 III (89)
0.5 1.086 I (97)
1.054 III (89)
1.0 0.97 1.26 1.13 III (267)
1.073 I (97)
0.927 III (89)
2.0 1.057 III (97)
1.06 II (35)
1.06 II (142)
1.06 1.02 II (267)
1.25 1.23 1.11 III (267)
4.0 1.028 III (97)
1.03 1.00 II (267)
1.22 1.10 III (267)
5.0 1.015 III (97)
6.0 0.996 I III (97)
1.00 0.96 1 II (267)
7.0 0.977 III (97)
8.0 0.92 1.12 1.04 III (267)
0.96 0.93 II (267)
12.00 0.89 0.86 II (267)
1.12 0.97 III (267)
16.0 0.76 0.70 II (267)
1.11 III (267)
 TABLE 4—Continued
Ruthenium (Il)-ruthenium (III)
Tris Complexes
Acid Normality Bistrpy Method Reference
4,4-Dimeth- 5,5-Dimeth- Phen 5-Meth- 5-Bro-
Bipy ylbipy ylbipy ylphen mophen

0.002 1.303 1.152 1.240 1.31 1.29 1.41 III (71)

0.1 1.288 1.115 1.208 1.30 1.28 1.36 III (71)
1.27 II (35)
1.281 III (89)
0.2 1.263 III (89)
0.3 1.279 1.096 1.185 1.29 III (71)
0.5 1.270 1.087 1.171 1.28 III (71)
1.219 III (89)
1.0 1.257 1.070 1.154 1.26 III (71)
1.29 III (95)
1.25 II (35)
1.202 III (89)
2.0 1.240 1.24 III (71)
1.175 III (89)
2.1 1.26 III (95)
3.0 1.222 1.22 III (71)
4.0 1.21 II (35)
5.0 1.19 III (71)
1.22 III (71)
1.22 III (95)
7.1 1.205 III (95)
8.0 1.16 II (35)
10.0 1.16 III (95)
12.0 1.15 III (95)
1.07 II (35)
13.5 1.105 III (95)
16.0 0.98 II (35)

Osmium (Il)-osmium (III)

Complexes
Acid Normality Method Reference
Trisbipy Trisphen Bistrpy

None added 0.877 I (76)

0.877 III (76)
0.05 0.863 I (76)
0.863 III (76)
0.1 0.855 I (76)
0.859 III (76)
0.9512 I (89)
0.2 0.847 I (76)
0.856 III (76)
0.944 I (89)
0.5 0.833 I (76)
0.842 III (76)
0.9250 I (89)
1.0 0.819 I (76)
0.822 III (76)
0.9070 I (89)
2.0 0.802 I (76)
0.8838 I (89)
3.0 0.775 I (76)
5.0 0.727 I (76)

1002
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1003

ruthenium, and osmium systems. In all, increasing acid concentration causes

lowered potentials. The change is in the direction required by consideration of
the effect of ionic strength on the cationic complex. However, the shift in poten-
tial is greater than can be explained in this manner. The comparison of the bi-
pyridine, terpyridine, and phenanthroline systems involves both similarities and
differences. The bipyridine potentials are generally less negative than those of
the phenanthroline systems. However, a definite trend is noticed from iron to
osmium. The iron-bipyridine system lies appreciably below that of phenan-
throline. With ruthenium and osmium the difference is not significant. The ter-
pyridine potentials in the iron and ruthenium systems are comparable at low
acidities to those for the phenanthroline and bipyridine complexes. However,
with osmium the terpyridine values are significantly above the other two. Thus
for phenanthroline and bipyridine, ruthenium and osmium are alike and differ-
ent from iron, whereas with terpyridine, iron and ruthenium are similar to each
other but different from osmium. It is noteworthy that the potentials of the ter-
pyridine systems are considerably more sensitive to hydrogen-ion concentration.
With all three metals the stability of the reduced form is best with phenan-
throline and that of the oxidized form is best with bipyridine (76). The difference
in potentials would not seem to be sufficient to explain this property. The oxida-
tion of the phenanthroline molecule in solution has been offered as an explana-
tion (95). However, this has not been definitely demonstrated. In view of the
extreme conditions necessary for such oxidation (268), it would seem that an-
other explantation might be sought.
The oxidation-reduction potentials of a large number of methyl-substituted
(35) and two hydroxy-substituted (128) phenanthroline complexes with iron(II)
have been determined. In the course of the investigation of the methyl deriva-
tives it became apparent that the effects of substitution were regular and could
be predicted with reasonable accuracy. Thus a methyl substitution in the 4- or
7- position lowers the potential by 0.11 v., while similar substitution in the 3- or
8- position lowers it by 0.03 v. and in the 5- or 6-position by 0.04 v. The addi-
tivity of these values has been demonstrated (35). This is the first system in
which properties of analytical interest are so readily predictable. Further ad-
vances along the same line should aid in the obtaining of molecules having spe-
cifically desired analytical properties with a minimum of synthetic work.
The same predictability was found to hold for the wave length of maximum
absorption of both the ferrous and the ferric complexes.
The first correlation of the physical properties of these compounds with sub-
stitution ivas demonstrated by Ewens (107), who found that the potentials of
the ferrous-phenanthroline complexes with nitro, methyl, bromo, or chloro sub-
stituents in the 5-position gave a straight-line relationship with the pKa of the
correspondingly p-substituted benzoic acids. The determination of the pKa of
the substituted phenanthrolines themselves permitted direct correlation (34).
Thus the variation in potential can be related directly to the basicity of the nitro-
gens in the substituted phenanthrolines. From the points available (71) it ap-
pears that the same correlation is realized with the ruthenium complexes. These
1004 BRANDT, DWYER, AND GY ARFAS

relationships appear reasonable on the basis of the pKa being a measure of the
availability of the electron cloud around the nitrogens and the resultant effect
upon the electrons around the central metallic ion. Thus, electrophilic sub-
stitution lowers the stability of the complex in both oxidation states, but the
stability of the tervalent complex is lowered more than that of the bivalent and
E° becomes more negative. The converse is true with nucleophilic substituents.

X. Analytical Applications
A. AS OXIDATION-REDUCTION INDICATORS

The proposal by Walden, Hammett, and Edmonds in 1931 (287) that the
iron-phenanthroline system might serve as an excellent oxidation-reduction
indicator was well received, and ferroin is now a commonly used indicator. Its
success has been due to the need for high-potential indicators for titrations with
quadrivalent cerium in particular. High precision and accuracy are obtainable.
It has been used in titrations with cerium in the determination of iron (130, 175,
268, 286, 287, 288), ferrocyanide (296), niobium (165), thallium (296), uranium
(193, 296), antimony (296), chromium and vanadium (297), hydrogen peroxide
(296), and arsenic (122).
The 5-nitroferroin complex has also been used in analysis (258, 262, 263, 264,
268), particularly where exceptionally high end-point indication is needed. How-
ever, the system is not as stable as ferroin and therefore finds limited usage. The
iron-bipyridine complexes suffer from the same limitation (47, 218, 287).
Ferroin has been used in titrations with brómate as an oxidizing agent (274).
The interference of free halogen in the titration was obviated by the use of
mercuric oxide. It has also been used in reactions utilizing potassium perman-
ganate (165, 297) and potassium dichromate (287). Ferroin is oxidized by di-
chromate in strongly acid solutions, whereas the reverse reaction can take place
in weak acid (174). A more suitable indicator for direct titrations involving
potassium dichromate is the 5,6-dimethylferroin (260). The latter indicator pro-
vides a more suitable potential for dichromate titrations (see table 4). Exchange
studies demonstrated fast electron-transfer reaction between 5,6-dimethylferroin
and 5,6-dimethylferriin (101). The oxidation of ferroin has been utilized in the
qualitative detection of cerium(IV) (59).
The trisbipyridine ruthenous ion has been proposed as an oxidation-reduction
indicator of high stability in hot, concentrated acid solutions and in solutions
of high potential (270). The trisphenanthroline osmium(II) complex has also
been suggested as having exceptional stability (76). The latter system has a
somewhat lower potential (see table 4). The advantages of these systems do not
appear to have been utilized yet to any great extent.

B. AS COLORIMETRIC REAGENTS

1. Iron
The intense red color formed by compounds of the phenanthroline type is
their primary characteristic. As colorimetric reagents for the determination of
iron they have attained international usage. Their application to this determi-
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1005

nation in all types of materials has produced a large number of incidental refer-
ences. These are not included in this survey. Phenanthroline has been used in
the determination of iron in biological materials (143, 184, 251) and alloys (250),
among others (268). In a similar capacity, bipyridine has found application with
samples of soil (144), biological materials (102, 160, 164), and miscellaneous
materials (268). Terpyridine has been used mostly for the determination of iron
in water samples (211).
The wave lengths of maximum absorption of the iron complexes of phenan-
throline, bipyridine, and terpyridine are 510, 522, and 552 µ, respectively
(268). Investigation of the series of methyl-substituted phenanthroline deriva-
tives previously described demonstrated the predictability of the change in wave
length of maximum absorption with substitution (35). Thus a methyl group in
the 4- or 7-position raises the value 1 µ, in the 5- or 6-position it is raised 5 µ,
and in the 3- or 8-position it is lowered 8 m¿u. This sequence is not the same as
with the oxidation-reduction potentials, where all substitution lowered the value,
with the 4- and 7-positions being the most effective. An attempt was made to
find the same numerical predictability with the molar absorptivity. However,
no such definite relation was found (35). A very useful concept did develop
however. It was shown that the 4- and 7-substituted molecules were the only
ones showing significant increases in the intensity of the color. This led to a
series of 4,7-substituted phenanthrolines which have since been utilized to more
than double the molar absorptivity with iron. This is a dramatic demonstration
of the usefulness of generalizations and predictability in the application of or-
ganic reagents.
To date the most sensitive reagent described is the 4,7-diphenylphenan-
throline (bathophenanthroline), which gives a complex with iron which can be
extracted into n-hexyl alcohol. The resultant molar absorptivity is 22,400. The
determination of iron in water is considerably improved by this increased sensi-
tivity (266). Three compounds with substituted amines in the 3- and 4-positions
have been investigated in a similar manner (295). The complexes of these rea-
gents give molar absorptivities of 16,000-19,000 but exhibit peculiar color
changes in acid, presumably due to the presence of the amino group (259).
Substitution in general appears to enhance the solubility of the complexes in
non-aqueous solvents (101, 140, 190, 266, 295). Rather striking effects are ob-
served with variations in the anion used (101, 190). The separation of iron from
vanadium, nickel, and chromium is readily accomplished (190). It also suggests
possible applications for the colorimetric determination of anions.
Further application of these generalizations would presumably be possible
with substituted bipyridines and terpyridines. The success of this approach
might be enhanced by a systematic study of the effect of different substituents.
Also, it has never been demonstrated that substitution in the 4- or 7-position
always produces the greatest increase in sensitivity, regardless of substituent.
2. Copper
The intense colors formed by compounds of the phenanthroline type with
cuprous ion had been reported (275) and analytical application unsuccessfully
1006 BRANDT, DWYER, AND GYARFAS

attempted (213). However, the introduction of a less familiar bipyridine type of

reagent, 2,2'-biquinoline, by Breckenridge, Lewis, and Quick (37) called atten-
tion to the cuprous reaction. Following a thorough investigation of many simi-
lar compounds by Hoste (139) and the establishment of the remarkable selec-
tibity of the reaction, the determination of copper with biquinoline has been
applied to several types of materials (60, 121, 126, 141).
The obvious steric factor which seemed to lead to the extreme selectivity of
the copper reaction is also present in 2,9-dimethylphenanthroline (265). The
increased sensitivity expected for 4,7-substitution was demonstrated with the
2,9-dimethyl-4,7-diphenylphenanthroline complex (269).
S. Other metals

The application of phenanthroline compounds to the colorimetric determina-

tion of metals other than iron and copper has been very limited. Terpyridine
has been proposed as a reagent for cobalt (212), and phenanthroline for molyb-
denum (214). An attempt to utilize the strong ultraviolet absorption prompted
proposals for the determination of cadmium (285) and zinc (173). These meth-
ods have not been utilized to any appreciable extent as yet. The first two sys-
tems are relatively unstable, and the last two present formidable experimental
difficulties in the form of interferences in the ultraviolet region. Data from spec-
tra for many phenanthroline and bipyridine complexes have been published
(300, 301), but no methods proposed.

C. AS AGENTS IN PRECIPITATION

The trisphenanthroline ferrous ion is a good precipitant for many anions. Use
of this fact has been made in the precipitation of the fluoride complex of tan-
talum (120). A good separation from niobium is possible. The qualitative solu-
bilities of many other ions in hydrofluoric acid solution have been determined
(120). The insolubility of the ferroin precipitates can be utilized to determine
mercury by precipitation of the tribromomercuriate anion (128a) and for the
precipitation of the hexaiodostannic (IV) ion (5). The iron-bipyridine complex
also is a sensitive precipitant for many anions (111, 239), and can be used in a
qualitative test for tannin in wine and leather (110). The bipyridine complex
has been utilized in a scheme for microscopic identification of anions (191). The
effect of substitution on solubility has been investigated with the 5-nitro and
5-bromo derivatives (268, 277). The solubility of the perchlorate salt of the fer-
rous complex increases in the order nitro derivative, bromo derivative, parent
molecule.
The insoluble complex Pd(phen)Cl2 has been used for the separation and de-
termination of the metal (247, 248, 249). The stability of this complex in con-
centrated sulfuric acid is utilized in a unique recrystallization which enhances
the value of the method (248).
D. MISCELLANEOUS APPLICATIONS

Various properties of complexes of the bipyridine type have been utilized to

analytical advantage. Their stability has been used to remove the interference
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1007

of iron in the determination of other metals. In this manner aluminum can be

precipitated as the hydrous oxide while iron is held in solution as the bipyridine
complex (261). Similarly the removal of iron interferences facilitates determina-
tions of magnesium, beryllium, and titanium (114). The stability of the nickel
and cobalt complexes permits a back-titration to determine the excess bipyridine
after its precipitation as the silver complex (58). The oxidation-reduction po-
tential of ferroin permits a colorimetric determination of vanadium (124, 156).
High-frequency titrations of many metals can be successfully carried out (131).
Phenanthroline complexes have been utilized in the drying of paint (217, 218,
294, 299) to compare the effects of complexed and uncomplexed metals as pro-
moters of the drying of paints and ink (302).

XI. Biological Activity

A wide variety of biological systems are affected by the bases themselves and
also by their metal complexes. The effects, however, are usually so distinct that
it is evident that entirely different principles are involved in the mode of bio-
logical action.
A. EFFECT OF THE ORGANIC REAGENTS

When phenanthroline or bipyridine (or for that matter any substance capable
of complex formation) is found to disturb the functioning of a biological system,
it is usually assumed that the sequestration of trace metals is involved. These
may be present as hydrated ions in the biological fluid medium or attached by
some coordinate bonds to protein by sulfur and/or nitrogen linkages. Thus the
metal ion may be rendered unavailable to the system if the stability of the
complex formed is high. On the other hand, in a situation where the metal ion
is bound to a protein surface, for instance, the reagent may attach itself to
vacant coordination positions by displacement of aquo groups, thus altering the
nature of the enzyme function. Inhibition of enzymes by complex-forming mole-
cules is often used as an indication of a heavy metal constituent. The presence
of a particular metal is often accepted when inhibition is caused by its specific
analytical reagent, without realizing that true specificity is rare, and that the
reaction tendencies may be very different when the metal ion is partly bound to
a protein.
The organic molecule itself may be the effective biological agent without the
intervention of a metal ion or even the free metal complex. It is unfortunate that
all of these possible modes of action are not usually recognized by workers in
this field. Generally one explanation is adduced, but the experimental work does
not exclude the alternatives. The large amount of relevant work is only briefly
surveyed here, and one hesitates to ascribe the biological effects to one particu-
lar mode of action.
The following systems were found to be affected by phenanthroline or bipyri-
dine: succinic oxidase (4); triphosphopyridine nucleotide nitrate reductase (216);
zymase (303); the reduction of fumarate and malate by molecular hydrogen in
the presence of washed suspensions of Escherichia coli (176). The latter inhibi-
tion could be partially reversed by addition of metal ions. The action of liver
1008 BRANDT, DWYER, AND GYARFAS

and intestine phosphomonoesterases on ¿3-glycerophosphate and of intestine

pyrophosphatase on pyrophosphate (69, 70) was considerably accelerated by
5 X 10~4 M solutions of the bases.
The endogenous respiration of anabaema (a blue-green alga) (290) and the
oxygen uptake (271) of green leaves were inhibited, but that of brain tissue was
accelerated (224) by both reagents as well as by their iron(II) complexes. Bac-
teriostatic action of phenanthroline has been demonstrated on several tubercle
bacteria (281), on a number of Gram positive organisms (28, 108), and on rumen
bacteria (195). Several of the phenanthrolines inhibited the growth of many
fungi (28). m- and p-Phenanthrolines were less effective than 1,10-phenanthro-
line, but 2,9-dimethylphenanthroline was the most potent inhibitor. Crude bi-
pyridine oil, obtained from the pyridine sodium reaction, has been reported as
toxic to aphides (245), but it has been shown (244) that the most active con-
stituents were a,¿3-bipyridine and a,¿3-bipiperidine. A 33 per cent saturated solu-
tion of phenanthroline suppressed the chemotactic power of “guinea pig” leuco-
cytes but did not destroy the cells (177). Highly potent anthelmintic action has
been observed with both substances (7).

B. EFFECT OF THE METAL COMPLEXES

Metal complexes of phenanthroline, bipyridine, and terpyridine have high bio-

logical activity. Toxicity to mice and rats (91) and to frogs and rabbits (20),
bacteriostatic and bacteriocidal properties (91), and action on enzymes (91, 92,
200, 224) and on neuromuscular transmission (25, 94) have been demon-
strated. The metal complexes used in these investigations were all coordinately
saturated and highly stable chemically and had no specific active groups or cen-
ters. The redox potentials were out of the biological redox range generally.
Whereas chelating molecules themselves can enter into various chemical reac-
tions in a biological system, these stable complexes, in which metal and chelate
are held in the firm grip of each other, should be unaffected. The highly stable
ruthenium(II) and osmium(II) complexes are ideal for this type of investigation.
Beccari (21) has demonstrated that the trisbipyridine iron(II) cation is excreted
unchanged through the kidneys of frogs and rabbits. The use of trisphenanthro-
line ruthenium(II) salts containing radioactive R,u106 and a chromatographic
technique showed that this compound is not metabolized in rats and mice
(91, 92).
Changing the metal in a series of complex cations M(B)£t (where B = bipy,
phen, or trpy, and M =
Fe, Ru, Os, Co, Zn, In) left the biological reaction es-
sentially unaffected. But if the ligand is altered—for instance, ethylenediamine
is substituted for bipyridine—the biological effect is quite different. These ob-
servations indicate that the activity is due to the complex cation as a whole and
not to its dissociated fragments. Since no chemical change is involved, the mode
of action must be purely physical, such as adsorption. The substances are
strongly adsorbed on protein and most likely on negative sites. As a result of
the positive charge they may alter the isoelectric point of proteins and may de-
polarize charged surfaces. Such changes in the zeta potential often lead to
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1009

changes of dispersity of colloids. By occupying negative sites they may block

active spots on enzymes. Whatever the effect, it is caused by virtue of their
positive charge or by their mere presence on an active site. The simple mode of
action of these complexes offers a unique tool for the investigation of biochemical
and physiological processes, and stable metal complexes will probably be used
more widely in the future. The desirable features of chemical and biological sta-
bility are by no means restricted to complexes of the bipyridine type but apply
to many others beyond the scope of this review.
The stereochemical specificity of biological systems towards these substances
is noteworthy. In general, this difference is greater between the optical isomers
than between complexes of the same configuration but containing different
metals. The lethal intraperitoneal dose for a mouse is 5.8 mg./kg. for (+)-
Ni(phen)3(C104)2 but more than 11 mg./kg. for the ( —)-form. At a concen-
tration of 10-4 M the inhibition of the enzyme acetylcholinesterase is 70 per cent
by ( —)-Ru(bipy)3Cl2 but only 20 per cent by the (+)-form (91, 92).
The results obtained by the detailed investigations of Beccari (20, 21, 22, 23,
24) with Fe(bipy)3++ on rabbits, and by Dwyer, Rogers, and coworkers (91, 92)
with Ru(phen)3++ on rats and mice, using radioactive tracer methods, show a
close resemblance. Very small amounts were lethal when introduced intraperi-
toneally, intravenously, or subcutaneously, but even large oral doses did not
cause death. Absorption through the intestine was slow (this has also been ob-
served on humans) (243) and, owing to rapid elimination through the kidney,
it was not possible to build up the lethal concentration in the blood. Some indi-
cations were obtained of the iron compound being split by intestinal bacteria
(20), but no changes were observed with the much more stable ruthenium com-
pound. The (+)-form of the ruthenium-phenanthroline compound is more
rapidly absorbed from the intraperitoneal cavity than the ( —)-form, and also
excreted more rapidly through the kidney. A dose of previously injected ( )- —

compound suppresses the adsorption of the (+)-compound. Thus an animal

which had previously been injected with half the toxic dose of the ( )-compound
—

survived a lethal dose of the (+)-form.

The symptoms caused by the complexes, i.e., torpor, tremor, paralysis of the
limbs, clonic convulsions, an unaffected heart, and death due to respiratory
failure, indicated an attack on the central nervous system. This has been sup-
ported by the finding that these substances are potent inhibitors of acetylcho-
linesterase (91, 92), and that they have a curare-like action on the neuromuscu-
lar transmission (25, 94). Acetylcholinesterase is thought to have two active
sites—a negatively charged one, and a so-called esteractic site. The inhibition
is competitive and presumably due to adsorption of the positive ions on the
anionic sites (92).
Neuromuscular transmission studies have been carried out by Beccari (25)
and Dwyer and coworkers (94) on rather similar lines. It is unfortunate that a
detailed account of Beccari’s work is unavailable. Nervous conduction does not
seem to be affected (25), but curare-like effects were obtained on the endplate
potentials at synaptic junctions (25). Neuromuscular block was observed at low
1010 BRANDT, DWYER, AND GYARFAS

concentrations with rat diaphragm preparations (94); the response of the frog
rectus abdominis preparation to acetylcholine was inhibited (94). Normal ac-
tivity could be restored by washing out the complexes.
Similar behavior of compounds of such different chemical constitution as
curare, decamethonium salts, and the metal complex cations indicates that the
common features of low chemical reactivity and the positive charge are the im-
portant ones.
It is generally believed that the positive charge associated with a complex
cation is not localized on the central metal atom but is fractionally distributed
over the complex (226). Since the size of the cations in a series M(phen)3++ is
probably very similar, quantitative differences in their biological action should
be due to differences in the peripheral charge. These seem to follow a certain
order:
Action on rat diaphragm preparation (bipyridine series, dl): Fe > Ni > Ru, Os
> Zn
Frog rectus abdominis preparation (phenanthroline series, dl): Ni > Fe > Ru > Zn
Inhibition of acetylcholinesterase (terpyridyl series): Fe < Ni < Ru < Os. Simi-
larly, Ni(5-N02-phen) 3++ < Ni^hen^ < Ni(5-Cl-phen)3++ < Ni(5-CH3-
phen)8++

The last series of substituted phenanthrolines does not follow the same order
as the pK of the bases. At the moment it is not possible to predict even qualita-
tively the order of increasing peripheral charge in such a series of complexes
with different metals. Apart from considerations of electronegativities, such fac-
tors as amount of double bonding and the use of higher or lower d-orbitals could
affect the charge distribution.

The authors are indebted to Dr. J. Koch, McMaster Laboratory, C.S.I.R.O.,

Sydney, for assistance with the biological section of this review. The work of
one of us (F. P. D.) has been written during his tenure as Du Pont Visiting
Professor of Chemistry at Northwestern University, Evanston, Illinois.
XII. References
(1) Albert, A., and Gold acre, R. J.: Private communication.
(2) Albert, A., Rubbo, S. D., Goldacre, R. J., and Balfour, B. G.: Brit. J. Exptl.
Pathol. 28, 69 (1947).
(3) Allison, J. A. C., and Mann, F. G.: J. Chem. Soc. 1949, 2915.
(4) Ames, S. R., Ziegenhagen, H. J., and Elvehjem, C. A.: J. Biol. Chem. 165, 81
(1946).
(5) Anderson, J. R., Livingstone, S. E., and Plowman, R. A.: J. Proc. Roy. Soc.
N. S. Wales 84, 184 (1950).
(6) Backhouse, R., Barnes, G. T., Dwyer, F. P., and Gyarfas, E. C.: Unpublished
work.
(7) Baldwin, E.: Brit. J. Pharmacol. 3, No. 2, 91 (1948).
(8) Balthis, J. H., and Bailar, J. C.: J. Am. Chem. Soc. 55, 1474 (1936).
(9) Barbieri, G. A., and Malaguiti, A.: Atti accad. nazi. Lincei 8, 619 (1950).
(10) Barbieri, G. A., and Tettamanzi, A.: Atti reale accad. nazi. Lincei 15, 877 (1932).
(11) Barnes, G. T., Dwyer, F. P., and Gyarfas, E. C.: Trans. Faraday Soc. 48, 269
(1952).
 CHELATE COMPLEXES OF 1,10-PHENANTHEOLINE 1011

(12) Basolo, F.: Chem. Revs. 52, 459 (1953).

(12a) Basolo, F., and Dwyer, F. P.: J. Am. Chem. Soc. 76, 1454 (1954).
(13) Basolo, F., Hayes, J. C., and Neumann, . M.: J. Am. Chem. Soc. 75, 5102 (1953).
(14) Basolo, F., and Murmann, R. K.: J. Am. Chem. Soc. 76, 211 (1954).
(15) Basolo, F., Murmann, R. K., and Yun Ti Chen: J. Am. Chem. Soc. 75, 1478 (1953).
(16) Basolo, F., and Neumann, . .: Private communication.
(17) Baxendale, J. H., and George, P.: Nature 162, 777 (1948).
(18) Baxendale, J. H., and George, P.: Nature 163, 725 (1949).
(19) Baxendale, J. H., and George, P.: Trans. Faraday Soc. 46, 55 (1950).
(20) Beccari, E.: Boll. soc. ital. biol. sper. 13, 6 (1938).
(21) Beccari, E.: Boll. soc. ital. biol. sper. 13, 8 (1938).
(22) Beccari, E.: Boll. soc. ital. biol. sper. 16, 214 (1941).
(23) Beccari, E.: Boll. soc. ital. biol. sper. 16, 216 (1941).
(24) Beccari, E.: Arch. sci. biol. (Italy) 27 , 204 (1941).
(25) Beccari, E.: Arch. sci. physiol. 3, 611 (1949).
(26) Bjerrum, J.: Metal Ammine Formation in Aqueous Solution. P. Haase and Son,
Copenhagen (1941).
(27) Bjerrum, J.: Chem. Revs. 46, 381 (1950).
(28) Blank, F.: Nature 168, 516 (1951).
(29) Blau, F.: Ber. 21, 1077 (1888).
(30) Blau, F.: Monatsh. Chem. 10, 367 (1889).
(31) Blau, F.: Monatsh. Chem. 10, 375 (1889).
(32) Blau, F.: Monatsh. Chem. 19, 647 (1898).
(33) Block, B. P., and Bailar, J. C.: J. Am. Chem. Soc. 73, 4722 (1951).
(34) Brandt, W. W., and Gullstrom, D. K.: J. Am. Chem. Soc. 74, 3532 (1952).
(35) Brandt, W. W., and Smith, G. F.: Anal. Chem. 21, 1313 (1949).
(36) Brandt, W. W., and Wright, J. P.: J. Am. Chem. Soc. 76, 3083 (1954).
(37) Breckenridge, J. G., Lewis, R. W. J., and Quick, L. A.: Can. J. Research 17B,
259 (1939).
(38) Brewer, F. M.: J. Chem. Soc. 1931, 361.
(39) Brimm, E. O.: Private communication.
(40) Brown, H. C., Schlesinger, . I., and Cardón, S. Z.: J. Am. Chem. Soc. 64, 325
(1942).
(41) Burstall, F. H.: J. Chem. Soc. 1938, 172.
(42) Burstall, F. H.: J. Chem. Soc. 1938, 1664.
(43) Burstall, F. H., Dwyer, F. P., and Gyarfas, E. C.: J. Chem. Soc. 1950, 953.
(44) Burstall, F. H., and Nyholm, R. S.: J. Chem. Soc. 1952, 3570.
(45) Burwell, R. W.: Private communication.
(46) Cagle, F. W.: Acta Cryst. 1, 658 (1948).
(47) Cagle, F. W., and Smith, G. F.: Anal. Chem. 19, 384 (1947).
(48) Cagle, F. W., and Smith, G. F.: J. Am. Chem. Soc. 69, 1860 (1947).
(49) Cambi, L., and Cagnasso, A.: Gazz. chim. ital. 63, 767 (1933).
(50) Cambi, L., and Cagnasso, A.: Gazz. chim. ital. 64, 772 (1935).
(51) Case, F. H.: J. Am. Chem. Soc. 68, 2574 (1946).
(52) Case, F. H.: J. Am. Chem. Soc. 70, 3994 (1948).
(53) Case, F. H.: J. Am. Chem. Soc. 71, 821 (1949).
(54) Case, F. H.: J. Am. Chem. Soc. 71, 1828 (1949).
(55) Case, F. H.: J. Org. Chem. 16, 941 (1951).
(56) Case, F. H.: J. Org. Chem. 16, 1541 (1951).
(57) Cavicchi, G.: Ann. chim. (Rome) 40, 149 (1950).
(58) Cavicchi, G.: Ann. chim. (Rome) 40, 158 (1950).
(59) Charlot, G.: Anal. Chim. Acta 1, 309 (1947).
(60) Cheng, K. Lu, and Bray, R. H.: Anal. Chem. 25, 655 (1953).
(60a) Cook, C. M., Jr., and Long, F. A.: J. Am. Chem. Soc. 73, 4119 (1951).
1012 BRANDT, DWYER, AND GYARFAS

(61) Cooper, L. . N.: Proc. Roy. Soc. (London) 118, 419 (1935).
(62) Cox, E. G., Webster, K. C., and Wardlaw, W.: J. Chem. Soc. 1935, 1475.
(63) Crawford, B. L., and Cross, P. C.: J. Chem. Phys. 6, 525 (1938).
(64) Davies, N. R., and Dwyer, F. P.: J. Proc. Roy. Soc. N. S. Wales 86, 64 (1952).
(65) Davies, N. R., and Dwyer, F. P.: Trans. Faraday Soc. 49,180 (1953).
(66) Davies, N. R., and Dwyer, F. P.: Trans. Faraday Soc. 60, 820 (1954).
(67) Davies, N. R., and Dwyer, F. P.: Unpublished work.
(68) Douglas, B. E., Laitinen, H. A., and Bailar, J. C.: J. Am. Chem. Soc. 72 , 2484
(1950).
(69) Dumazert, C., Levy, M., and Marszak, I.: Compt. rend. soc. biol. 139, 99 (1945).
(70) Dumazert, C., Levy, M., and Marszak, I.: Compt. rend. soc. biol. 139, 219 (1945).
(71) Dwyer, F. P.: J. Proc. Roy. Soc. N. S. Wales 83, 134 (1949).
(72) Dwyer, F. P.: Unpublished work.
(73) Dwyer, F. P. and Ceglowski, M. J.: Unpublished work.
(74) Dwyer, F. P., and Gibson, N. A.: J. Proc. Roy. Soc. N. S. Wales 84, 83 (1950).
(75) Dwyer, F. P., Gibson, N. A., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales
84, 68 (1951).
(76) Dwyer, F. P., Gibson, N. A., and Gyarfas, E. C.: J. Proc. Roy Soc. N. S. Wales
84, 80 (1951).
(77) Dwyer, F. P., Gill, N. S., Gyarfas, E. C., and Lions, F.: J. Am. Chem. Soc. 76,
3834 (1953).
(78) Dwyer, F. P., Gill, N. S., Gyarfas, E. C., and Lions, F.: Unpublished work.
(79) Dwyer, F. P., and Gyarfas, E. C.: Nature 163 918 (1949).
,

(80) Dwyer, F. P., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales 83, 170 (1949).
(81) Dwyer, F. P., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales 83, 174 (1949).
(82) Dwyer, F. P., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales 83, 232 (1949).
(83) Dwyer, F. P., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales 83 263 (1949). ,

(84) Dwyer, F. P., and Gyarfas, E. C.: J. Proc. Roy. Soc. N. S. Wales 84 68 (1950). ,

(85) Dwyer, F. P., and Gyarfas, E. C.: Nature 166, 481 (1950).
(86) Dwyer, F. P., and Gyarfas, E. C.: J. Am. Chem. Soc. 73, 2322 (1951).
(87) Dwyer, F. P., and Gyarfas, E. C.: Nature 168, 29 (1951).
(88) Dwyer, F. P., and Gyarfas, E. C.: J. Am. Chem. Soc. 74, 4699 (1952).
(89) Dwyer, F. P., and Gyarfas, E. C.: J. Am. Chem. Soc. 76, in press (1954).
(90) Dwyer, F. P., and Gyarfas, E. C.: Unpublished work.
(91) Dwyer, F. P., Gyarfas, E. C., Koch, J., and Rogers, W. P.: Nature 170,190 (1952).
(92) Dwyer, F. P., Gyarfas, E. C., Koch, J., Rogers, W. P., and Simmonds, R.: Un-
published work.
(93) Dwyer, F. P., Gyarfas, E. C., and O’Dwyer, M. F.: Nature 166 481 (1950). ,

(94) Dwyer, F. P., Gyarfas, E. C., Rogers, W. P., Shulman, A., and Wright, F.:
Unpublished work.
(95) Dwyer, F. P., Humpoletz, J. E., and Nyholm, R. S.: J. Proc. Roy. Soc. N. S. Wales
80, 212 (1946).
(96) Dwyer, F. P., Lions, F., and Mellor, D. P.: J. Am. Chem. Soc. 72, 5037 (1950).
(97) Dwyer, F. P., and McKenzie, H. A.: J. Proc. Roy. Soc. N. S. Wales 81, 93 (1947).
(98) Dwyer, F. P., and McKenzie, H. A.: J. Proc. Roy. Soc. N. S. Wales 81, 97 (1947).
(99) Dwyer, F. P., and Nyholm, R. S.: J. Proc. Roy. Soc. N. S. Wales 80 28 (1946). ,

(100) Dwyer, F. P., and Wooldridge, H.: J. Proc. Roy. Soc. N. S. Wales 83 235 (1949). ,

(101) Eimer, L., and Medalia, A. I.: J. Am. Chem. Soc. 74, 1592 (1952).
(102) Emmerie, A.: Rec. trav. chim. 66, 489 (1946).
(103) Emmert, B., and Brandl, F.: Ber. 60 , 2211 (1927).
(104) Emmert, B., Diehl, K., and Gollwitzer, F.: Ber. 62, 1733 (1929).
(105) Emmert, B., and Schneider, O.: Ber. 66, 1875 (1933).
(106) Erlenmeyer, H., and Ueberwasser, H.: Helv. Chim. Acta 28,1268 (1940).
(107) Ewens, R. V. G.: Nature 155, 398 (1945).
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1013

(108) Feeney, R. E.: Arch. Biochem. Biophys. 34, 196 (1951).

(109) Feigl, F.: Specific, Sensitive and Selective Reactions. Elsevier Press, New York (1952).
(110) Feigl, F., and Feigl, . E.: Brazil, Ministério agr., Dept. nací, produgáo mineral,
Lab. produgáo mineral, Bol. No. 24, 25 (1946); Chem. Abstracts 43, 8183 (1949).
(111) Feigl, F., and Miranda, L. I.: Ind. Eng. Chem., Anal. Ed. 16, 141 (1944).
(112) Fernelius, W. C.: Record Chem. Progr. (Kresge-Hooker Sci. Lib.) 8, 17 (1950).
(113) Ferrari, G.: Gazz. chim. ital. 67 , 604 (1937).
(114) Ferrari, G.: Ann. chim. appl. 27, 479 (1937). .

(115) Fielding, P. W., and LeFevre, R. J. W.: J. Chem. Soc. 1951, 1811.
(116) Figgis, B. T., and Nyholm, R. S.: Unpublished work.
(117) Foss, . E., and Gibson, C. S.: J. Chem. Soc. 1931, 2415.
(118) Fritz, J. S., Cagle, F. W., and Smith, G. F.: J. Am. Chem. Soc. 71, 2480 (1949).
(119) Gaines, A. Jr., Hammett, L. P., and Walden, G. H.: J. Am. Chem. Soc. 58, 1668
(1936).
(120) Gillis, J., Eeckhout, J., and Poma, K.: Verhandel. Koninkl. Vlaamsche Acad.
Wetenschappen, Belg. 6, No. 10, 5 (1944); Chem. Abstracts 40, 4314 (1946).
(121) Gillis, J., Hoste, J., and Fernandez-Caldas, E.: Anales edafol. y fisiol. vegetal
(Madrid) 9, 585 (1950); Chem. Abstracts 45, 5561 (1951).
(122) Gleu, K.: Z. anal. Chem. 95, 305 (1933).
(123) Goethals, C. H.: Rec. trav. chim. 64, 299 (1935).
(124) Gottlieb, A.: Mikrochemie ver. Mikrochim. Acta 36/37, 370 (1951).
(125) Gould, R. K., and Vosburgh, W. C.: J. Am. Chem. Soc. 64, 1630 (1942).
(126) Guest, R. J.: Can. Dept. Mines and Tech. Surveys, Topical Rept. No. Tr-105/52,
11 pp. (1952); Chem. Abstracts 47, 6304 (1953).
(127) Halcraw, B. E., and Kermack, W. O.: J. Chem. Soc. 1946, 155.
(128) Hale, . N., and Mellon, M. G.: J. Am. Chem. Soc. 72, 3217 (1950).
(128a) Hall, . E., and Smith, A. .: Anal. Chem. 23, 1181 (1951).
(129) Hammett, L. P., Walden, G. H., and Chapman, R. P.: J. Am. Chem. Soc. 53, 3908
(1931).
(130) Hammett, L. P., Walden, G. H., and Edmonds, S. M.: J. Am. Chem. Soc. 66, 1092
(1934).
(131) Hara, R.: J. Pharm. Soc. Japan 71, 1128 (1951).
(132) Harris, C. M.: Private communication.
(133) Harvey, A. ., and Manning, D. L.: J. Am. Chem. Soc. 74, 4744 (1952).
(134) Hein, F.: Chemische Koordinationslehre. S. Hirzel, Zurich (1950).
(135) Hein, F., and Herzog, S.: Z. anorg. Chem. 268, 335 (1951).
(136) Hieber, W., and Muhlbauer, F.: Ber. 61, 2151 (1928).
(137) Hieber, W., and Sonnekalb, F.: Ber. 61B, 555 (1928).
(138) Hoste, F.: Research (London) 1, 713 (1948).
(139) Hoste, F.: Anal. Chim. Acta 4, 23 (1950).
(140) Hoste, J., and Gillis, J.: Mededeel. Koninkl. Vlaamsche Acad. Wetenschappen,
Belg. 13, No. 2, 3 (1951); Chem. Abstracts 46, 7464 (1952).
(141) Hoste, J., Heiremans, A., and Gillis, J.: Mikrochemie ver. Mikrochim. Acta
36/37, 349 (1951).
(142) Hume, D. N., and Kolthoff, I. M.: J. Am. Chem. Soc. 65, 1895 (1943).
(143) Hummel, F. C., and Willard, . H.: Ind. Eng. Chem., Anal. Ed. 10, 13 (1938).
(144) Ignatibff, V.: J. Soc. Chem. Ind. (London) 56, 407 (1937).
(145) Irving, .: A Discussion on Coordination Chemistry, Paper No. 4, Welwyn Herts,
May, 1951.
(146) Irving, H., Butler, E. J., and Ring, M. F.: J. Chem. Soc. 1949, 1489.
(147) Irving, H., Cabell, M. J., and Mellor, D. H.: J. Chem. Soc. 1953, 3417.
(148) Irving, H., and Williams, R. J. P.: Nature 162, 764 (1948).
(149) Irving, H., and Williams, R. J. P.: J. Chem. Soc. 1953, 3192.
(150) Jaeger, F. .: Z. anorg. Chem. 170, 384 (1928).
1014 BRANDT, DWYER, AND GYARFAS

(151) Jab gbr, F. M.: Optical Activity and High Temperature Measurements, pp. 93, 139.
McGraw-Hill Book Company, Inc., New York (1930).
(151a) Jaeger, F. M., and van Dijk, J. A.: Koninkl. Ned. Akad. Wetenschap., Proc. 87,
10 (1934); Chem. Abstracts 28, 2638 (1934).
(152) Jaeger, F. M., and van Dijk, J. A.: Koninkl. Ned. Akad. Wetenschap., Proc. 37,
284 (1934); Chem. Abstracts 29, 1733 (1935).
(153) Jaeger, F. M., and van Dijk, J. A.: Koninkl. Ned. Akad. Wetenschap., Proc. 37,
618 (1934); Chem. Abstracts 29, 1352 (1935).
(154) Jaeger, F. M., and van Dijk, J. A.: Koninkl. Ned. Akad. Wetenschap., Proc. 39,
164 (1936); Chem. Abstracts 30, 3350 (1936).
(155) Jaeger, F. M., and van Dijk, J. A.: Z. anorg. Chem. 227 273 (1936).
,

(156) Jantsch, G., and Zemek, F.: Z. anal. Chem. 139, 249 (1935).
(157) Johnson, J. E., and Hall, N. F.: J. Am. Chem. Soc. 70, 2344 (1948).
(158) Jonassen, . B.: Record Chem. Progr. (Kresge-Hooker Sci. Lib.) 13, 135 (1952).
(159) Kanekar, C. R., and Nyholm, R. S.: Unpublished work.
(160) Kaunitz, H., and Beaver, J. J.: J. Biol. Chem. 156, 661 (1944).
(161) Kimball, G. E.: J. Chem. Phys. 8, 188 (1940).
(162) King, W. R., and Garner, C. S.: J. Am. Chem. Soc. 74, 3709 (1952).
(163) Kleinberg, J.: Unfamiliar Oxidation States and their Stabilization. University of
Kansas Press, Lawrence, Kansas (1950).
(163a) Klemm, W., Jacob, H., and Tilk, W.: Z. anorg. Chem. 201, 1 (1931).
(164) Klockmann, R.: E. Merck’s Jahresber. 53,59 (1939); Chem. Abstracts 34, 7309 (1940).
(165) Knowles, . B., and Lundell, G. E. F.: J. Res. Natl. Bur. Standards 42,405 (1949).
(166) Kolthoff, I. M., Lee, T. S., and Leussing, D. L.: Anal. Chem. 20, 985 (1948).
(167) Kolthoff, I. M., Leussing, D. L., and Lee, T. S.: J. Am. Chem. Soc. 72, 2173 (1950).
(168) Kolthoff, I. M., Leussing, D. L., and Lee, T. S.: J. Am. Chem. Soc. 73, 390 (1951).
(169) Komorovsky, ., and Poluektov, N.: Mikrochim. Acta 1, 264 (1937).
(170) Krumholz, P.: J. Am. Chem. Soc. 71, 3654 (1949).
(171) Krumholz, P.; Nature 163, 724 (1949).
(172) Krumholz, P.: Anais. acad. brasil, cienc. 22 , 263 (1950); Chem. Abstracts 45 , 3209
(1951).
(173) Kruse, J. M., and Brandt, W. W.: Anal. Chem. 24, 1306 (1952).
(174) Lang, R.: Mikrochim. Acta 3, 113 (1938).
(175) Lang, R., and Furstenau, I.: Z. anal. Chem. 133, 72 (1951).
(176) Lascelles, J., and Still, J. L.: Proc. Linnean Soc. N. S. Wales 72, 49 (1947); Chem.
Abstracts 41,7435 (1947).
(177) Lebrun, J., and Delauney, A.: Ann. inst. Pasteur 80, 524 (1951).
(178) Lee, T. S., Kolthoff, I. M., and Leussing, D. L.: J. Am. Chem. Soc. 70, 2348 (1948).
(179) Lee, T. S., Kolthoff, I. M., and Leussing, D. L.: J. Am. Chem. Soc. 70,3596 (1948).
(180) Linstead, R. P.: Ber. 72A, 93 (1939).
(181) Livingstone, S. E.: J. Proc. Roy. Soc. N. S. Wales 85, 151 (1951).
(182) Livingstone, S. E.: J. Proc. Roy. Soc. N. S. Wales 86, 32 (1952).
(183) Longuet-Higgins, H. C., and Coulson, C. A.: J. Chem. Soc. 1949, 971.
(184) MacFadyen, D. A., and Van Slyke, D. D.: J. Biol. Chem. 149, 527 (1943).
(185) Malaguiti, A.: Atti accad. nazi. Lincei 9, 349 (1950).
(186) Malby, L. E., and Mellor, D. P.: Nature 159, 370 (1947).
(187) Malby, L. E., and Mellor, D. P.: Nature 161, 436 (1948).
(188) Malby, L. E., and Mellor, D. P.: Australian J. Sci. 2A, 92 (1949).
(189) Mann, F. G., Purdie, D., and Wells, A. F.: J. Chem. Soc. 1936, 1503.
(190) Margerum, D. W., and Banks, C. V.: Anal. Chem. 26, 200 (1954).
(191) Markunas, P.: Ph.D. Dissertation, University of Illinois, 1940.
(192) Martell, A. E., and Calvin, .: Chemistry of the Metal Chelate Compounds. Pren-
tice-Hall, Inc., New York (1952).
(193) McClure, J. H., and Banks, C. V.: Proc. Iowa Acad. Sci. 65, 263 (1948).
 CHELATE COMPLEXES OF 1,10-PHENANTHROLINE 1015

(194) McCurdy, W. H., and Smith, G. F.: Analyst 77, 846 (1952).
(195) McNaught, M. L., and Owen, E. C.: Intern. Congr. Biochem., Abstr. of Commune.,
1st Congr., Cambridge, Engl., p. 304 (1949).
(196) Mellor, D. P.: Chem. Revs. 33, 137 (1943).
(197) Mellor, D. P., and Gold acre, R. J.: J. Proc. Roy. Soc. N. S. Wales 73 233 (1940).
,

(198) Merritt, L. L., and Walker, J. K.: Ind. Eng. Chem., Anal. Ed. 16, 387 (1944).
(199) Michablis, L., and Granice, S.: J. Am. Chem. Soc. 66 , 481 (1943).
(200) Michablis, L., and Stern, K. G.: Biochem. Z. 240, 192 (1931).
(201) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1930, 2594.
(202) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1931, 2213.
(203) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1932, 20.
(204) Morgan, G. T., and Burstall, F. H.: J. Indian Chem. Soc., Ray Commemoration
Vol. 1933, 1.
(205) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1934, 965.
(206) Morgan, G. T., and Burstall, F. .: Inorganic Chemistry. Heffers, Cambridge
(1936).
(207) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1937, 1649.
(208) Morgan, G. T., and Burstall, F. H.: J. Chem. Soc. 1938, 1672.
(209) Morgan, G. T., and Davies, G. R.: J. Chem. Soc. 1938, 1858.
(210) Morris, B. S.: Unpublished work.
(211) Morris, R. L.: Anal. Chem. 24, 1376 (1952).
(212) Moss, M. L., and Mellon, M. G.: Ind. Eng. Chem., Anal. Ed. 15, 74 (1943).
(213) Moss, M. L., and Mellon, M. G.: Ind. Eng. Chem., Anal. Ed. 15, 116 (1943).
(214) Moss, M. L., Mellon, M. G., and Smith, G. F.: Ind. Eng. Chem., Anal. Ed. 14,
931 (1942).
(215) Murmann, R. K.: Ph.D. Dissertation, Northwestern University, 1953.
(216) Nason, A., and Evans, H. J.: J. Biol. Chem. 202, 655 (1953).
(217) Nicholson, D. G.: J. Am. Chem. Soc. 64 , 2820 (1942).
(218) Nicholson, D. G.: Ind. Eng. Chem. 34, 1175 (1942).
(219) Nyholm, R. S.: J. Chem. Soc. 1951, 3245.
(220) Nyholm, R. S.: Chem. Revs. 53, 263 (1953).
(221) Nyholm, R. S.: Private communication from F. H. Burstall.
(222) Nyholm, R. S., and Short, L. N.: J. Chem. Soc. 1953 2670.
(223) Onstott, E. I. and Laitinen, H. A.: J. Am. Chem., Soc. 72, 4724 (1950).
(224) Panimon, F., Horvitt, . K., and Gerard, R. W.: J. Cellular Comp. Physiol. 17,
17 (1941).
(225) Pauling, L.: The Nature of the Chemical Bond. Cornell University Press, Ithaca,
New York (1945).
(226) Pauling, L.: J. Chem. Soc. 1948, 1461.
(227) Pfeiffer, P., and Christeleit, W.: J. prakt. Chem. 151, 127 (1938).
(228) Pfeiffer, P., and Christeleit, W.: Z. anorg. Chem. 239, 133 (1938).
(229) Pfeiffer, P., Dominik, V., Fritzen, A., and Werdelmann, B.: Z. anorg. Chem.
260, 84 (1949).
(230) Pfeiffer, P., and Nakatsuka, Y.: Ber. 66B, 415 (1933).
(231) Pfeiffer, P., and Quehl, K.: Ber. 64, 2667 (1931).
(232) Pfeiffer, P., and Quehl, K.: Ber. 65, 560 (1932).
(233) Pfeiffer, P., and Quehl, IC.: Ber. 65, 2560 (1932).
(234) Pfeiffer, P., and Tapperman, Fr.: Z. anorg. Chem. 215, 273 (1933).
(235) Pfeiffer, P., and Werdelmann, Br. : Z. anorg. Chem. 261, 197 (1950).
(236) Pfeiffer, P., and Werdelmann, Br.: Z. anorg. Chem. 263, 31 (1950).
(237) Pflaum, R. T.: Ph.D. Dissertation, Purdue University, 1953.
(238) Phillips, J. P., and Fernando, Q.: J. Am. Chem. Soc. 75, 3768 (1953).
(239) Poluektov, N. S., and Nazarenko, V. A.: J. Appl. Chem. U.S.S.R. 10, 2105 (1937).
(240) Pucheault, J., and Lefort, M.: J. chim. phys. 50, 196 (1953).
1016 BRANDT, DWYER, AND GYARFAS

(241) Pullman, A.: Rev. sci. 86, 219 (1948).

(242) Quagliano, J. V., and Schubert, L.: Chem. Revs. 50, 201 (1952).
(243) Rechenberger, J., and Pollack, C.: Z. physiol. Chem. 281, 186 (1944).
(244) Richardson, C. H., Shepherd, . H., and Smith, C. R.: J. Econ. Entomol. 23,
863 (1930).
(245) Richardson, C. H., and Smith, C. R.: J. Agr. Research 33, 597 (1926).
(246) Ruben, S., Kamen, M. D., Allen, . B., and Nahinsky, P.: J. Am. Chem. Soc.
64, 2297 (1942).
(247) Ryan, D. E.: Can. J. Research 27B, 938 (1949).
(248) Ryan, D. E.: Analyst 77, 46 (1952).
(249) Ryan, D. E., and Fainer, P.: Can. J. Research 27B, 67 (1949).
(250) Ryan, J. A., and Botham, G. H.: Anal. Chem. 21, 1521 (1949).
(251) Schmidt, H. G.: Biochem. Z. 305, 104 (1940).
(252) Sidgwick, N. V., and Brewer, F. M.: J. Chem. Soc. 1925, 2379.
(253) Simon, A., and Haufe, W.: Z. anorg. Chem. 230, 160 (1936).
(254) Simon, A., Morganstern, G., and Albrecht, W. .: Z. anorg. Chem. 230, 225 (1936).
(255) Skeline, R. K., and Pitzer, K. S.: J. Am. Chem. Soc. 72, 1107 (1950).
(256) Smirnoff, A.: Helv. Chim. Acta 4, 802 (1921).
(257) Smith, C. R.: J. Am. Chem. Soc. 50, 1936 (1928).
(258) Smith, G. F.: Cerate Oxidimetry. G. F. Smith Chemical Company, Columbus, Ohio
(1942).
(259) Smith, G. F.: Private communication.
(260) Smith, G. F., and Brandt, W. W.: Anal. Chem. 21, 948 (1949).
(261) Smith, G. F., and Cagle, F. W.: Anal. Chem. 20, 574 (1948).
(262) Smith, G. F., and Duke, F. R.: Ind. Eng. Chem., Anal. Ed. 13, 558 (1941).
(263) Smith, G. F., and Duke, F. R.: Ind. Eng. Chem., Anal. Ed. 15, 120 (1943).
(264) Smith, G. F., and Getz, G. A.: Ind. Eng. Chem., Anal. Ed. 10, 191 (1938).
(265) Smith, G. F., and McCurdy, W. H.: Anal. Chem. 24, 371 (1952).
(266) Smith, G. F., McCurdy, W. H., and Diehl, H.: Analyst 77, 418 (1952).
(267) Smith, G. F., and Richter, F. P.: Ind. Eng. Chem., Anal. Ed. 16, 580 (1944).
(268) Smith, G. F., and Richter, F. P.: Phenanthroline and Substituted Phenthroline
Indicators. G. F. Smith Chemical Company, Columbus, Ohio (1944).
(269) Smith, G. F., and Wilkins, D. H.: Anal. Chem. 25, 510 (1953).
(270) Steigman, J., Birnbaum, N., and Edmonds, J. .: Ind. Eng. Chem., Anal. Ed. 14,
30 (1942).
(271) Stenlid, G.: Physiol. Plantarum 2, 61 (1949).
(272) Sutton, G. J.: Roy. Australian Chem. Inst., J. & Proc. 16, 115 (1949).
(273) Sutton, G. J.: Australian J. Sci. 4A, 651 (1951).
(274) Szebelledy, L., and Madis, V.: Z. anal. Chem. 114, 116 (1934).
(275) Tartarí ni, G.: Gazz. chim. ital. 63, 597 (1933).
(276) Taube, H.: Chem. Revs. 60 , 69 (1952).
(277) Taylor, W. H.: Ph.D. Dissertation, University of Illinois, 1942.
(278) Traube, W., and Passarge, W.: Ber. 46, 2505 (1913).
(279) Tschugaeff, L.: Z. anorg. Chem. 46, 160 (1905).
(280) Tschugaeff, L.: Ber. 41,1678 (1908).
(281) Turian, G.: Schweiz. Z. allgem. Pathol, u. Bakteriol. 14,338 (1951); Chem. Abstracts
45, 9124 (1951).
(282) Turner, E. E., and Harris, . M.: Quart. Revs. (London) 1, 299 (1948).
(283) Uri, N.: Chem. Revs. 50, 375 (1952).
(284) Vosburgh, W. C., and Cooper, G. R.: J. Am. Chem. Soc. 63, 437 (1941).
(285) Wadelin, C. W., and Mellon, M. G.: Anal. Chem. 24, 894 (1952).
(286) Walden, G. H., Hammett, L. P., and Chapman, R. P.: J. Am. Chem. Soc. 55, 2649
(1933).
 CHELATE COMPLEXES OF 1,10-PHENANTHEOLINE 1017

(287) Walden, G. H., Hammett, L. P., and Chapman, R. P.: J. Am. Chem. Soc. 63, 3908
(1931).
(288) Walden, G. H., Hammett, L. P., and Edmonds, S. M.: J. Am. Chem. Soc. 66, 350
(1934).
(289) Walden, G. H., Hammett, L. P., and Gaines, A. Jr.: J. Chem. Phys. 3, 364 (1935).
(290) Webster, G. C., and Frenkel, A. W.: Plant Physiol. 28, 63 (1953).
(291) Welcher, F. J.: Organic Analytical Reagents, Vol. I, p. 262. D. Van Nostrand Co.,
New York (1947).
(292) Werner, A.: Z. anorg. Chem. 3, 267 (1893).
(293) Werner, A.: Her. 45, 433 (1912).
(294) Wheeler, G. K.: Ind. Eng. Chem. 39, 1115 (1947).
(295) Wilkins, D. H., McCurdy, W. H., and Smith, G. F.: Anal. Chim. Acta 8, 46 (1953).
(296) Willard, . H., and Young, P.: J. Am. Chem. Soc. 65, 3280 (1953).
(297) Willard, . H., and Young, P.: Ind. Eng. Chem., Anal. Ed. 6, 48 (1934).
(298) Willink, J. P., and Wibaut, H. D.: Rec. trav. chim. 64, 275 (1935).
(299) Worthington, E. A., and Nicholson, D. G.: Paint Oil Chem. Rev. 112, No. 11, 20
(1949).
(300) Yamasaki, K.: Bull. Chem. Soc. Japan 12, 390 (1937).
(301) Yamasaki, K., Haba, T., and Yasuda, M.: Proc. Japan. Acad. 29, 337 (1953).
(302) Zettlemoyer, A. C., and Nace, D. M.: Ind. Eng. Chem. 42, 491 (1950).
(303) Zuckerkandl, F., and Messiner-Klebermass, L.: Biochem. Z. 261, 55 (1933).