View Online / Journal Homepage / Table of Contents for this issue
1033
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
Dent, Linstead, and Lowe : Phthalocyanines. Part V I .
217, Phthalocyanines. Part V I . The Xtructure of the Phthalocyanines.
By C. E. DENT, R. P. LINSTEAD,and A. R. LOWE.
COMPARATIVE
experiments have shown that the reagents which readily yield phthalocyanines
with o-phthalonitrile give no similar products with terephthalonitrile (I),homophthalonitrile (11), o-xylylene dicyanide (111), o-cyanocinnamonitrile (IV), and 2 : 2'-diphenonitrile (V).
CN
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
1034
Dent, Linstead, and Lowe : Phthalocyanines. Part V I .
View Online
These facts indicate that the two nitrile groups participating in phthalocyanine formation must be linked to adjacent carbon atoms of an aromatic nucleus and cannot be separated by an additional saturated atom or by an extended unsaturated or aromatic unit.
The obvious inference, that the production of a phthalocyanine involves the formation of a
ring fused to the aromatic nucleus in the 1 : 2 positions and containing six atoms or less,
is supported by the hydrolysis and oxidation of phthalocyanines to derivatives of o-phthalic
acid.
The evidence of synthesis and of fission and the results of analysis * both indicate the
/\/c-N
11
. The analytical figures prove that the non-benzenoid
presence of the skeleton
\/'C-N
carbon and the nitrogen atom carry no oxygen and little or no hydrogen (the exact content
of hydrogen, which is left uncertain by the figures, is discussed later). The only structural
units which satisfy these conditions are an isoindole ring with an extracyclic nitrogen atom
(VI) and a phthalazine ring (VIZ) :
a)- a;>II
N=
(W
QCC>NH
NH
(IS.)
(VIII.)
(~711.)
The available evidence is strongly in favour of the first of these formulae. The controlled
decomposition of phthalocyanine to phthalimide and ammonium nitrate by nitric acid
(p. 1021) indicates that the molecule contains two dissimilar types of nitrogen atom. This is
not, however, a definite proof, because it is known that certain phthalazines can be converted into derivatives of isoindole by extrusion of one nitrogen atom, as, for example, in
the reduction of chlorophthalazines to dihydroisoindoles (Gabriel and Neumann, Ber., 1893,
26, 521, 705). A comparison has therefore been made of the capacity of the two systems
for phthalocyanine formation. Experiments with iminophthalimidine (VIII) and the
isomeric phthalazone (IX) have shown that the isoindole derivative readily yields phthalocyanines under conditions where phthalazone is unchanged. Indeed we have been unable
to obtain definite indications of phthalocyanine formation by prolonged treatment of (IX)
under the most favourable conditions. Further experiments by Dr. E. G. Noble and one of
us show that methylphthalazone, methylphthalazine, and phthalazine itself (X, XI, and
XII) also give negative results, the last even in the presence of oxidising agents such as
cupric and ferric chlorides.
(X.1
(XI1.)
(XI-)
The isoindole formula is also to be preferred on stereochemical grounds. Inspection of
the complete formulze (such as XIV) given below, which contain four C, units, shows that
it would be impossible to join four phthalazine nuclei in the form of a suitable large complex
* It will be convenient t o summarise here the analytical data for pure phthalocyanine from various
sources :
Method of preparation.
Mg compound (ex cyanobenzamide)
Found
,#
(ex phthalonitrile)
+ H,SO, ..........
+ HC1 ............
............
+ H2S0,
Cyanobenzamide
antimony ..............................
Phthalonitrile + sodium amyloxide .....................
for (C8H,N2), ...................................................
::
for (C,H,N,),H,
................................................
................................................
c , 76.
74.8
74.0
74%
74.3
74.5
75.0
74.7
74-4
H, %.
3-6
3.7
3.6
3-7
3.8
3.1
3.5
3.1)
N,
%.
21-5
21-6
21-5
21-9
21.8
21.7
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
Dent, Limtead, and Lowe : Phthalocyanines. Part V I .
View Online
1035
ring, as the components would have to be linked directly without the intervention of the
connecting atoms, which in the isoindole formula provide the necessary angular divergence.
The intense colour of the phthalocyanines makes it obvious that their molecules cannot
be composed of one isoindole ring. The analyses of the metallic derivatives show that four
such units are combined for every atom of metal-an aggregation which is confirmed by the
molecular weight of magnesium phthalocyanine (see Part V). The close similarity in
physical properties of free phthalocyanine and its metallic derivatives indicates that the
parent compound also has four C, units in the molecule. A four-unit formula is further
supported by the results of the quantitative oxidation of both the free and the metallic
phthalocyanines.
The only reasonable way in which these isoindole units can be joined is by means of
the extracyclic nitrogen atoms : the remaining valencies will then form a conjugated chain,
in keeping with the strong colour. This leads to the " chain formula (XITI) or the ring
formula (XIV).
"
If two imino-hydrogen atoms are left available for the fixation of the metal, (XIII) is
equivalent to (C,H,N,), and (XIV) to (C8H,N2)4H2. The analytical difference is too small
to be detectable. The formation of copper phthalocyanine, (C,H,N,),Cu, from phthalonitrile, and copper can only be simply accounted for on the basis of the ring formula, and
the chain formula is excluded by the results of quantitative oxidation (p. 1038).
A slightly different ring formula is (XV), which contains no imino-hydrogen atoms and
has all the aromatic rings in the benzenoid condition, whereas in (XIV) one ring is
o-quinonoid.
Formulz (XIV) and (XV) differ in their implications respecting the attachment of the
metal on the derivatives. Copper phthalocyanine could be expressed as a derivative of
(XIV) by the formula (XVI), in which the metal has displaced two hydrogen atoms and is
attached by covalencies to two isoindole nitrogen atoms (and may be co-ordinated to the
other two; see p, 1037). In a formula based on (XV),however, the metal would be held
purely by co-ordinate links to the nitrogen atoms (XVII).
CH
CH C
CH
(XVIII .)
These formu12 are analogous to the alternative representations of the complex metallic
derivatives of indigo (Kunz and co-workers, Ber., 1922, 55, 3699; 1923, 56, 2027 ; 1925,
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
1036
Dent, Linstead, and Lowe : Phthalocyanines. Part V I .
View Online
58, 1860; 1927, 60, 367 ; 1930, 63, 2600) for which the formula involving " primary "
valencies (corresponding to XVI) has been proved correct by Kuhn and Machemer (Ber.,
1928, 61, 118).
There is experimental evidence in favour of formulz (XIV) and (XVI) and opposed to
(XV) and (XVII). The stability of the copper compound in the vapour phase a t 580"would
be inconceivable if the metal were held solely by co-ordinate links. If there were no iminohydrogen atoms in phthalocyanine, hydrogen should be liberated during the preparation of
this substance by the action of sulphuric acid on the magnesium compound : none has been
observed. If, on the other hand, the imino-hydrogen atoms are present, the formation of
the metallic compounds from free phthalocyanine should be accompanied by an evolution
of hydrogen. We find, however, that hydrogen is slowly evolved when magnesium alone
is boiled with quinoline ; the reaction, therefore, is not diagnostic.
It was hoped to obtained decisive evidence by the Tschugaeff-Zerewitinoff method for
the estimation of active hydrogen (Ber., 1902, 35, 3912; 1907, 40, 2023, etc.; compare
Hibbert and Sudborough, J., 1904, 85, 933, etc.), which has recently been applied to the
investigation of complex metallic compounds and their precursors by Kuhn and others (Zoc.
cit. ; Ber., 1928, 61, l28), although it has not been so successful in the porphyrin, haemin,
and chlorophyll series (see a summary by Fischer and Rothemund, Ber., 1931, a,201).
Unfortunately the method or known modifications could not be applied to the phthalocyanines owing to their insolubility in the usual media, and the use of quinoline was also
unsatisfactory (p. 1038).
Fission by oxidation provided the desired evidence. It was known that phthalocyanine
and its heavy-metal derivatives were stable in cold sulphuric acid solution for long periods,
but were decomposed by hot acid permanganate so readily that suspensions in dilute
sulphuric acid could be titrated with this reagent. The fission of a molecule of a phthalocyanine of formula (XIV) to phthalimide requires one atom of oxygen (a), whereas a molecule of formula (XV) requires no oxygen ( b ) :
+-
(a) (C8H4N2)4H2 7H20
0 = 4C8H502N 4NH3
(b) (CsH4N2)4 8H20 = 4C8H502N 4NH3
The ease of the oxidative fission is therefore strong evidence in favour of (XIV), and to
make this conclusive, we endeavoured to find conditions under which equation (a) could be
realised quantitatively. This could not be achieved by the use of permanganate, as there
was some autocatalytic decomposition of the reagent , oxygen being partly liberated and
partly dissolved in the solution. Ceric sulphate, which has been used as a stable substitute
for permanganate in inorganic analyses (Mitchell and Ward, " Modern Methods in Quantitative Chemical Analysis," p. 15),proved to be excellent for our purpose. We are indebted
to Dr. A. M. Ward for suggesting its use.
Ceric sulphate reacts very rapidly with a suspension of phthalocyanine in dilute sulphuric acid a t room temperature. Exactly one atom of oxygen is taken up (for each four
c8 units) and about 90% of the theoretical quantity of phthalimide can be isolated from the
product. Under similar conditions copper phthalocyanine is oxidised according to the
equation :
(C8H4N2)4CU SH20 0 = 4C8H502N CUO 4NH3
Oxidation of magnesium phthalocyanine (dihydrate) proceeds similarly :
(C8H,N2)&2H20
+ 6H20 + 0 = 4C,H,O,N + MgO + 4NH,
This shows that the combined oxygen in this compound is not effective for oxidation.
Phthalonitrile takes up no oxygen under these conditions, which justifies experimentally
the contention that the hydrolysis of a substance of formula (XV), which is a polymeride
of phthalonitrile, to phthalimide would not involve oxidation (equation b).
The absence of oxidisable hydrogen in the metallic derivatives and its presence in the
parent compound proves that a bivalent metal takes the place of two atoms of hydrogen
and is therefore bound by covaIencies.
Dent, Linstead, and Lowe : Phthalocyanipzes. Part V I .
If phthalocyanine had the open-chain structure [XIII = (C,H,N,),], each molecule
should either take up two atoms of oxygen and form four molecules of phthalimide or take
up one atom of oxygen and form three molecules of phthalimide and one molecule of
phthalimidine. Both of these possibilities, and hence the open-chain structure, are therefore excluded and the structures represented by (XIV) and (XVI) are proposed for phthalocyanine and its metallic derivatives. The fine structure, i.e., the relative positions of the
imino-hydrogen atoms and the consequent arrangement of the double bonds, cannot be
decided by the chemical evidence. The hydrogen atoms have been placed on opposite
instead of on adjacent rings in the formulz on the basis of X-ray measurements by Dr. J. M.
Robert son.
The stability of copper phthalocyanine indicates that the metal is co-ordinated, presumably with the other two isoindole nitrogen atoms, but it is necessary to consider this
more fully in relation to the stereochemistry of the molecule as a whole. On the basis of
formula (XIV) the cyclic framework of phthalocyanine should exist in one plane and the
only possibility of deviation lies in the ability of ring A to rotate on the two single links
connecting it with the extracyclic nitrogen atoms. If all the rings lie in one plane, the
valencies connecting the imino-nitrogen atoms will be inclined to this, so that the hydrogen
atoms of phthalocyanine will lie out of the great plane.
Both cis- and trans-forms of this compound then become possible, according to the
relative position of the two hydrogen atoms. Models show that a metal exhibiting normal
tetrahedral symmetry can take the place of the two hydrogen atoms without appreciable
strain in a cis-phthalocyanine and with but little strain in a trams-form. On the other hand,
if the other two isoindole nitrogen atoms become co-ordinated, the new links would tend to
lie in the great plane. Hence, for a metal with tetrahedral symmetry to be co-ordinated,
the molecule must accommodate (1) some distortion of the nitrogen valencies and (2) considerable distortion of the metallic valencies. A metal with planar symmetry (e.g., nickel)
could, however, be co-ordinated in the centre of the molecule with distortion of only two
nitrogen valencies. It therefore appears that, if co-ordination of a metal normally exhibiting tetrahedral symmetry occurs, either the molecule must lose its planar configuration
or the normal direction of the metallic valencies must be modified.
Connexion with the Porfihyrins.-There is a close connexion between phthalocyanines of
formulae (XIV) and (XVI) and the porphyrins, which form the basis of many important
natural colouring matters and have been shown by Kuster, Willstatter, Hans Fischer, and
others t o contain the fundamental porphin ring (XVIII). (The imino-hydrogen atoms are
placed on opposite rings by analogy with those of phthalocyanine.) This differs from
phthalocyanine in two respects ; the pyrrole units do not carry benzene nuclei and are connected to one another by methine groups in place of nitrogen atoms. These factors would
not be expected to influence the general spatial arrangement or the stability of the complex
molecule, for it is well known that a nitrogen atom may take the place of a methine group
in rings without greatly affecting their general character. In agreement with this, there
are a number of resemblances between the two types : both are stable to alkalis, less so to
acids ; both are highly coloured, and form complex metallic compounds ; and both can be
degraded by oxidation to the imides of dibasic acids. Moreover, there is a similarity in the
order of stability of the metallic derivatives of the porphyrin and of the phthalocyanine series.
For instance, the metallic phthalocyanines may be compared with the corresponding derivatives of substances of the type of phytochlorin and phytorhodin. Willstatter and Stoll
( Chlorophyll, 1913, Chapter XIX) state that the magnesium compounds of these are
intermediate in stability between the potassium derivatives, which lose the metal in dilute
alcoholic solution, and the copper compounds, which are of unparalleled stability. *
The spatial arrangement of the great ring and the metal may be presumed to be identical
in the phthalocyanines and the porphyrins and it is hoped that further investigation
of the metallic phthalocyanines may throw some light on the general stereochemical
problem.
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
View Online
1037
* Geh. Professor Dr. R. WillstFitter predicted the great stability of copper phthalocyanine to one
of us (R. P. L.) some months before the compound was first isolated.
1038
Dent, Linstead, and Lowe : Phthalocyaizines. Part V I .
View Online
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
EXPERIMENTAL
.
Experiments with Dinitriles other than o-Phthalonitri1e.-Terephthalonitrilc, homophthalonitrile, o-xylylene dicyanide, o-cyanocinnamonitrile, and 2 : 2'-diphenonitrile were each treated
with the following reagents : magnesium, copper, and cuprous chloride a t the boiling point of
the dinitrile, and with sodium amyloxide in boiling amyl alcohol. No coloured products could
be isolated in any of the experiments. o-Xylylene dicyanide and o-cyanocinnamonitrile gave a
green colour with sodium amyloxide, but no pigment could be isolated from the solutions.
2 : 2'-Diphenonitrile gave a similar green colour with molten metallic sodium.
Comparison of Iminophthalimidine and Phthalazone (VIII and IX).-As o-cyanobenzamide
passes apparently completely into iminophthalimidine above its melting point (Braun and Tclierniac, Zoc. cit.), the results of Part I1 serve for one side of this comparison. Pure recrystalliserl
iminophthalimidine (m. p. 203'), made by acidification of a solution of o-cyanobenzamide in alkali
with acetic acid a t O", also readily gave magnesium phthalocyanine'when heated with magnesium
oxide a t 240-260'.
Phthalazone, m. p. 182O, made from phthalide by Gabriel and Neumann's
method (Bey., 1893, 26, 521), yielded no trace of green or blue colour when heated even t o the
boiling point (337") with sodium, sodium amyloxide, magnesium, or cuprous chloride. With
magnesium oxide a weak green colour was formed, but no pigment could be isolated from the
product. Zinc chloride also gave a slight green colour.
A ttempted Determinations of " A dive Hydrogen.-A two-chambered reaction vessel similar
to that recommended by Zerewitinoff (Bey., 1907, 40, 2023) was used. The Grignard compound
was prepared from methyl iodide and magnesium in pure amyl ether. The quinoline was freed
from aromatic bases by the action of nitrous acid and converted into the dichromate, which was
repeatedly crystallised. The base was regenerated with alkali and distilled in steam and under
reduced pressure, b. p. 115'/16 mm. It gave no colour with acid hypochlorite and no methane
with the Grignard reagent.
The procedure followed that previously used (Zoc. cit. ; Sudborough and Hibbert, J., 1904, 85,
933; 1909, 95, 477) except that quinoline was employed as solvent. The results were satisfactory for phthalimide (Found : active H, 1.0, 0.9, 1.0, 0.9, 1.1, 1.0 atom per molecule). Phthalocyanine slowly evolved methane with the reagent, but the results of a large number of experiments varied with the time of reaction t o such an extent that they could not be satisfactorily
corrected by blank experiments.
Quantitative 0xidations.The following results were obtained by dissolution of the phthalocyanine in concentrated sulphuric acid, precipitation with water, and titration at the boiling
point with N/20-potassium permanganate : Found for phthalocyanine : oxidisable H, 0.49,
0.56. (C,H,N,),H, requires 0.39%. Found for copper phthalocyanine : oxidisable Cu, 15.7,
15.7. (C,H,N,),Cu requires 11-05y0. The results for both compounds are high in the same proportion and the method must therefore be at fault. When the precipitated copper phthalocyanine was filtered off, washed, and titrated in dilute sulphuric acid, a rather better result was
obtained (Cu, 12.9%).
Use of ceric sulphate. A known weight (about 0.3 g.) of the finely divided compound was
triturated with a little dilute sulphuric acid, and a known excess of exactly N/lO-ceric sulphate
solution (Mitchell and Ward, op. cit.) added. The bulk of the pigment rapidly disappeared. To
ensure the completion of the reaction, the solution was kept a t 60" until clear, any lumps being
crushed. N/lO-Ferrous ammonium sulphate equivalent to the original amount of ceric sulphate
was then added, and the excess of ferrous salt titrated with N/lO-ceric sulphate, xylene cyanole
FF being used as indicator. The colour changes from greenish-yellow t o brown in the presence
of excess of the ceric salt ; the end-point is sharp and can be approached from either side. The
final ceric sulphate titre is equivalent to the amount originally employed to oxidise the phthalocyanine.
The results are summarised below. The last column gives the molecular weight assumed in
the calculations :
I'
....................................
Mg in magnesium phthalocyanine (dihydrate) ......
Cu in copper phthalocyanine ...........................
yo Oxidisable H in phthalocyanine
,,
,,
Found.
0.39, 0.38
Calc.
0-39
4.2
4-25
11.2, 10.6
11.05
4.3,
M.W.
514
572
676
When the solution obtained in the oxidation of the magnesium compound was kept for 2
days, long needles of phthalimide formed (yield, 87.5%), m. p- and mixed m. p. 228'. The
mother-liquor on continuous extraction with ether for a further 5 days yielded a further 4.5% of
phthalimide. Similarly, phthalocyanine itself yielded 89% of pure phthalimide.
Experiments on the Synthesis of Anthocyanim. Part X X I .
View Online
1039
Our thanks are due to Professor J. F. Thorpe, C.B.E., F.R.S., for his kindness in enabling us
to carry out this work and for his encouragement and advice, We are also indebted t o Imperial
Chemical Industries Limited for grants and gifts of chemicals, to Dr. E. H. Rodd for many valuable suggestions, to Dr. H. F. Harwood for advice on analytical matters, and to t he Salters
Company for a grant to one of us (A. R. L.).
Downloaded by University of Sussex on 07 October 2012
Published on 01 January 1934 on http://pubs.rsc.org | doi:10.1039/JR9340001033
IMPERIAL
COLLEGE,LONDON,
S.W. 7.
[Received, M a y 22nd, 1934.1
