Group Title: preparation and characterization of several hexaaza macrocyclic compounds of cobalt(II), nickel(II), and copper(II) /
Title: The preparation and characterization of several hexaaza macrocyclic compounds of cobalt(II), nickel(II), and copper(II) /
Full Citation
Permanent Link:
 Material Information
Title: The preparation and characterization of several hexaaza macrocyclic compounds of cobalt(II), nickel(II), and copper(II) /
Physical Description: xi, 116 leaves : ill. ; 28cm.
Language: English
Creator: Myers, Frederick Felder, 1948-
Publication Date: 1975
Copyright Date: 1975
Subject: Cobalt compounds   ( lcsh )
Copper compounds   ( lcsh )
Nickel compounds   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 111-115.
Statement of Responsibility: by Frederick Felder Myers, Jr.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098317
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000161374
oclc - 02671253
notis - AAS7714


This item has the following downloads:

preparationchara00myer ( PDF )

Full Text







To my wife and parents

for loving encouragement .


The author wishes to express his sincere appreciation

to Dr. R. C. Stoufer, Chairman of the author's Supervisory

Committee, and to the other members of his Supervisory


The author is indebted to Neil "the Clam" Weinstein for

his tremendous help which permitted the completion of this

work. Thanks Neil! Thanks also go to the various members

of the research group with whom this author was associated

for their invaluable bits of knowledge.



ACKNOWLEDGEMENTS .............. ....

LIST OF TABLES. .............. .....

LIST OF FIGURES. ............. ....


ABSTRACT. .............. .......

INTRODUCTION ........... ..~~~







EXPERIMENTAL. .............

Preparation of Starting Hlaterials..



General .....
Infrared Spectra.....
Electronic Spectra......
Magnetic Susceptibility......
Electron Spin Resonance.....
Ligands... . .

SUMM~ARY ......




.. . 30

.. . 34
.. .. 41
.. . 48
.. . 55
. .. 73
.. . 78

.. . 95

.. . 98


1. Electronic spectra data..... ....... .......... 42

2. Selected molar conductance and limiting conduc-
tance of prepared complexes in H20 at 25.000......... 49

3. Known electrolyte types and their molar conductivity. 50

4. Magnetic data................. .... . 56

5. Temperature-dependent magnetic susceptibility of
Ni(ketomacr)C124H20.. .............. ........_ 59

6. Temperature-dependent magnetic susceptibility of
Ni(ketomacr)(NO3 2-6H20...... ......... 60

7. Temperature-dependent magnetic susceptibility of
Co(ketomacr)C12-4H20. ........ _. ._ ........... 61

8. Temperature-dependent magentic susceptibility of
Co (aldomacr)C12-4H20 ........... ......... 62

9. Temperature-dependent magnetic susceptibility of
Cu(ketomacr)(NO3 2..... .... 63

10. ESR data. ........ .............. ...... 76

A-1 Selected infrared absorption bands (cm-1) for Ni(II)
macracyclic complexes.. ............. ......... 99

A-2 Selected infrared absorption bands (cm-1) for Co(II)
macrocyclic complexes. ............ .. ..........101

A-3 Selected infrared absorption bands (cm-1) for Cu(II)
macrocyclic complexes........... ... ..........103

A-4 Selected infrared absorption bands (cm~-1) for
"ketomacr" and its precursors. ........ .. ........ ...1OS

A-5 Selected infrared absorption bands (cm-1) for
"aldomacr" and its precursors. ....... ....... ...106i

List of Tables (Continued)

A-6 Mass spectral cracking pattern for the
condensate of DAP and DAPH... .......... .. ..... 107

A-7 M~ass spectral cracking pattern for the
condensate of PDC and PDCH ............. ........ 109



1. Porphyrin skeleton. ............. ........ 2

2. Corrin skeleton.. .......... ............ 2

3. Structural formula of DAP and PDC. ......... ..... 7

4. Structural formula of DAPH and PDCH............... 7

5. Structural formula of ketomacr, aldomacr, and
aldoketomacr. ............_ .. ......... 8

6. 2,6-Bis(propynone)pyridine... ........ ..... .... 10

7. Infrared spectra of Ni(aldomacr)C12 20O,
Ni(aldoketomacr)C12-4H20, and Ni(ketomacr)C124H20 35

8. Infrared spectra of Co(aldomacr)C124H20 and
Co(ketomacr)C124 20. ............. . ........_ 36

9. Infrared spectra of Cu(aldoketomacr)Cl2-2H20 and
Cu(aldomacr)C12-4H20. 37

10. Infrared spectra of Co(ketomacr)(NO3)24H20
Ni(ketomacr)(NO3)2-6H20, and Cu(ketomacr)(N3).. 3

11. Diffuse reflectance spectra of Ni(ketomacr)Cl2'
4H20 and Ni(ketomacr)(NO3)2-6H20......... ....:... 45

12. Diffuse reflectance spectra of Co(ketomacr)Cl2'
4H20 and Co(ketomacr)(NO332 20. 465
13. Diffuse reflectance spectra of Cu(aldoketomacr)-
C12-2H20 and Cu(aldomacr)C12 H20 ......... ....... 47
14. Molar conductance vs. C" for BaC12. ............... 51

15. Malar conductance vs. C for Ni(ketomacr)C12-4H20,
a Ni(ketomacr)(a -r6H20 Ni(aldoketomacr)C12-4H20,
andNialomcrC 2'4H20......... ....... 52

16 olrconductance vs. C frCkeoa)C24H20,
Co(aldomacr)C124H20, and Co(ketomacr)(NO3)2 420. 53

List of Figures (continued)


17. olarconductance vs. C for Cu(ketomacr)(NO3)2,
Cu(aldoketomacr)Cl2-2H20,and Cu(aldomacr)C12-4H20..5
18. Temperature dependence of inverse magnetic
susceptibility for Ni(ketomacr)C12-4H20...... ....... 64

19. Temperature dependence of inverse magnetic
susceptibility for Ni(ketamacr)(NO3 2-6H20.......... 65
20. Temperature dependence of inverse magnetic
susceptibility for Co(ketomacr)C12-4H20.......6
21. Temperature dependence of inverse magnetic
susceptibility for Co(aldoketomacr)C12-4H20. ........ 67

22. Temperature dependence of inverse magnetic
susceptibility for Cu(ketomacr)(NO3 2....... 68

23a. Splitting diagram for a single electron. ........... 75

b. Zero-field splitting diagram. ......... .. ......... 76

24. Infrared spectra of "ketomacr" and its precursors... 83

25. Infrared spectra of "aldomacr" and its precursors... 93












A o






See Fig. Sa

See Fig. 5b

See Fig. Sc


benz.o [b,m][1,4,12,15] tetraazacyclodocosine
-1 -1
molar conductance, micromohs M cm

limiting conductance

corrected molar magnetic susceptibility

Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy



Frederick Felder Myers, Jr.

December, 1975

Chairman: R. Carl Stoufer
Major Department: Chemistry

Ten new macrocyclic complexes containing the ligands,

2,5,11,14- cecramethyl-3,4,12, 13,19, 20-hexaazatricyclo ll3.3 1 16,10t

eicosa-1(19),2,4,6,8,10(20),11,13,15,17-dean (ketomacr),


6,8,10(20),11,13,15,17-decaene (aldomacr), and 2,14-dimethyl-

3,4,12,13,19, 20-hexaazatricyclo [,1]eicosa-1(19),2,6,-

6,8,10(20),11,13,15,17-decaene (aldoketomacr), have been prepar-

ed via the template method. The complexes were produced with

chloride and nitrate salts of cobalt(II), nickel(II), and

copper(II) by a Schiff base condensation of 2,6-diacetylpyri-

dine and 2,6-pyridinediacetyldihydrazone for ketomacr,

2,6-pyridi~nedicarboxaldehyde and 2,6-pyridinedialdihydrazone

for aldomacr, and 2,6-pyridinedicarboxaldehyde and 2,6-pyridine-

diacet~yldihydrazone for aldoketomacr. Isolation of the free

macrocyclic base could neither be accomplished by direct

synthesis nor by removal of the complexes' metal ion.

The complexes were characterized by elemental analysis,

infrared, ultraviolet, visible, and electron spin resonance

spectra, conductance, and magnetic susceptibility determinations.

The results of these studies support the formulation of each of

the complexes as a metal ion surrounded by a planar, quadra-

dentate ligand with water either very loosely held in the axial

positions or present as part of the crystal lattice.

The nickel(II) chloride and nitrate complexes of ketomacr

have above-normal room temperature magnetic moments which are

reduced by 0.5 BM upon dehydration of the samples. The temper-

ature dependent magnetic susceptibility data, obtained for

Ni(ketomacr)C12-4H20, Ni(ketomacr)(NO3 2-6H20, Co(ketomacr)-

C12-4H20, Co(aldoketamacr)C12 6 20, and Cu(ketomacr)(NO3 2'
followed the normal Curie-Weiss law.

Metal template reactions have been defined as ligandd

reactions which are dependent on, or can be significantly

enhanced by, a particular geometrical orientation imposed by

metal coordination" (1). A macrocyclic complex is one in

which the metal ion is circumscribed by a ligand which is,

itself, a closed ring. Review articles concerning various

aspects of the coordination chemistry of macrocyclic ligands

reflect the increased attention this class of compounds has

received in the past fifteen years (1-12). The entire

Volume 100 of Advances ~in Chemistry Series is dedicated to the

biological involvement of metal complexes, many of which are

macrocyclic complexes. Such macrocyclic complexes--in particu-

lar, those which have a planar arrangement of four nitrogen

donor atoms--are related in varying degree to a number of

biologically important molecules such as porphyrins (3) and

corrins (4) (Figures 1 and 2). Metal complexes are involved

in material transfer such as oxygen transport by hemoglobin,

in material storage such as that of iron by ferritin and in

energy transfer as performed by chlorophyll systems. Some

workers have proposed that a macrocyclic complex is involved

in natural nitrogen fixation (5).

In an attempt to more completely elucidate the behavior




Fig. 1 Porphyrin skeleton

Fig. 2 Corrin skeleton

of metal ion-macrocyclic ligand systems, a study of model

systems less complicated than the macrocyclic systems and

complexes found in nature is reasonable; that is, knowledge

of the properties and behavior of macrocyclic complexes is

desirable for the understanding it may impart to biological

behavior. Synthetic macracyclic complexes prepared during

the investigations reported herein are potential analogs of

the natural systems and consequently are of use in model

experiments (6).

One factor enhancing the stability of macrocyclic

complexes is the chelate effect. This term means that a

metal chelate complex is more stable than a related complex

containing only monodentate ligands (13). For transition

metal ions the chelate effect consists of an enthalpy and an

entropy contribution.

Further stability may be gained by the formation of

multiple metal-ligand bonds. Multiple bonding of this type

may occur provided (14) : (i) the metal contains electrons in

the dxz and dyz orbitals (where the axis system is defined such

that the organic macrocycle coordinated to the metal lies in the

xy plane) and (ii) the ligand contains empty Egg-antibonding

molecular orbitals of proper symmetry to which the contribution

of the metal (pi-donor) atom is finite. The satisfaction of

these two requirements allows back donation of the d-electrons

from the metal to the Ei-antibonding orbitals of the ligand

imparting double bond character to the metal-ligand bond. The

C=N linkage, frequently occurring in natural systems, meets

the criteria listed above; experiments have demonstrated that

there appears to be considerable stabilization of the metal-

ligand bond through partial double bond formation (15). It

is likely that a combination of all the above effects results

in the exceptional stability of macrocyclic complexes.

Many unsaturated macrocyclic complexes have been observed

to be inert (16). This inertness has been attributed to

very large dissociational activation energies. Because of thle

closed ring structure of the macracycle, no simple dissociative

step involving the metal ion or donor atom can occur. It is

not possible to extend the metal-donor distance sufficiently

to constitute bond breaking without either bond breaking with~-

in the ligand or extensive rearrangement within the coordina-

tion sphere.

Synthetic routes to the preparation of macrocyclic

complexes can be divided into two broad categories. One is

derived from a Schiff base condensation of carbonyl compounds

with bis(diamine) complexes (11,16,17). Another is derived

from the reaction of coordinated mercaptides with alkyl and

aryl dihalides. Examples of the latter type of complexes have

been prepared by Busch and co-workers (18,19).

Metal template reactions are almost mandatory for the

synthesis of macracyclic metal complexes (20). The metal ion

coordinates to and orients the reactant species in such a waa

as to render the cyclization more probable than were

the metal ion absent. Reaction of the organic moieties in the

absence of the metal ion either leads to oligomers and/or

very low yields of the desired macrocycle. For a more complete

discussion of this "template effect" the reader is referred to

review articles by Busch et al. (2,10,21).

The design of a ligand must take into account the factors

affecting the stability of metal complexes including the

geometry of the molecular construction. The geometrical

consideration can be grouped into four categories: (i) size of

metal ion, (ii) type of donor atoms, (iii) "hole size" of the

mnacrocycle, and (iv) other steric considerations relating to

the periphery of the macrocyclic base.

Black and Hartshorn have compiled a review (22) of these

features of ligand design and synthesis, with emphasis placed

upon the nature of the ligands. Busch and co-workers have

demonstrated (23) the relationship of metal ion radius and

macrocyclic ring size to values of the ligand field parameter

22xIy. They conclude that the metal-donor distance can have a

profound effect on the strength of the metal-donor interaction.

Up until the last decade (10), few macrocyclic complexes

with extended conjugation have been prepared. Certainly more

stringent steric considerations have to be observed in the

design and, preparation of complexes of this type. Since many

biologically important systems contain this extended conjugation

within the macrocyclic ring, it was decided to attempt the pre-

paration and characterization of complexes of this type via

Schiff base condensations and metal template techniques.

The organic reactants chosen for the Schiff base con-

densations (Figs. 3, 4, and 5) were: DAP and DAPH, PDC and

PDCH, and PDC and PDCH (see page ix for a list of symbols).

Metal ions of the first row transition series were chosen for

their known (24) tendency to promote metal template reactions

involving systems similar to those listed above. Also metal

ions of this type are catalysts in the sense that such Schiff

base condensations are known to proceed by way of a nucle-

ophilic attack by the amine nitrogen on the carbon of the

carbonyl group to yield a carbinol-amine intermediate.

Coordination of the carbonyl oxygen on DAP or PDC to a positive

center--the metal ion--would favor the reaction by making the

carbonyl carbon atom more susceptible to nucleophilic attack.

The presence of the pyridine group between the carbonyl

functions in DAP or PDC promotes initial tridentate chelation

of the reactant (21). This chelation then activates the

coordinated carbonyls toward reaction with the amine groups onr

DAPH or PDCH leading to the macrocyclic complex.

Work on this study was almost complete when, to this

writer's consternation, a recent article by Goedken et al.

(25) was discovered in the literature which reported the svnthe-

sis and structural characterization of iron(II) complexes of

the ketomacrocyclic ligand. This discovery does take some of

the novelty away from this work; but little overlap exists.

Different solvent systems and a different approach to the

template reaction were employed. Goedken did prepare the Co(HI)

derivative, but this is the only part where both works are

R ;



DAP RL=R2=CH3; 2, 6-diacetylpyridine
PDC RL=R2=H; 2, 6-pyridinedicarboxaldehyde

Fig. 3 Structural formula of DAP and PDC

R ,- R2

N N\
H N ~N H,


R =R2=CH3; 2, 6-pyridinediacetyldihydrazone
R =R2=H; 2, 6-pyridinedialdihydrazone

Fig. 4 Structural formula of DAPH and PDCH


a) R1=R2=R3=R4=CH3; 2,5,11,14 Tetramethyl-3,4,12,13,19,20-

Abbreviated as ketomacr.

b) RI=R2=Ry=R 3,4,12,13,19,20-Hexaazatricyclo-
[13..1.1 ]eicosa-l(19),2,4,6,8,10(20),11,13,15,17-

Abbreviated as aldomacr.

c) RI=R2=H, R =R4=CH3; 2,14-Dimethyl-3,4,12,13,19,20-hexa-
azatricyclo[,101eicosa-1(19),2, 4680(),

Abbreviated as aldoketomacr.

*Fig. 5 Structural formulas of ketomacr, aldomacr, and

*The symbol M placed within the structure of the macrocycle
may represent Ni(II), Cu(II) or Co(II). Its presence here
is to clarify sites of bonding.

identical. Further comparison to Goedken's work will follow

in the discussion.

A different but novel type of system considered to be

extremely interesting is represented in Fig. 6. This tri-

dentate ligand would be expected to form complexes of mixed

type, i.e., it would contain both M-N and M-C bonds. Indeed,

preparation of the original ligand and complexes could be

followed by cyclization to form novel macrocyclic complexes.

Such an investigation would include the study of the

interactions of the metal-ligand bonds. Two configurations

could exist (26) in this type of metal-ligand arrangement.

The alkyne type would give rise to a typical acetylenic

stretching vibration between 2000 and 2200 em-1 (27, 28). The

other type, the alkene type, would have associated with it

an acetylenic stretching vibration around 1800 cm-1 (29).

The alkyne type would result from the donation of the

electron density in an acetylene Ei-bonding orbital into an

empty E- or d-orbital on the metal and back donation of the

electron density from a filled d-orbital to an acetylene pi-

antibonding orbital.

The lowering of the energy of the acetylenic stretching

vibration in the alkene type would be due to the reorganization

of the electron density of the acetylene bond to give an 01efin-

like arrangement. As a result of this rearrangement two

sigma bonds would be formed between the metal ion and the





Fig. 6 2,6-Bis(propynone)pyridine

3F.~T~:ys~ws~;\TT",'c~ir~:r*~~,r,,:i ;r- "2-~r"7-`~-~"~ r~y



Unless otherwise specified all chemicals were

commercially available as reagent grade and were used

without further purification.

2,6-Pyridinedimethanol. This compound was purchased

from Aldrich Chemical Co. and used without further purifi-

cation: mp 113-1140

2, 6-Diacetylpyridine. This compound was purchased from

Aldrich Chemical Co. and used without further purification:

mp 78-790

Preparation of Starting Mlaterials

2,6-Pyridinediacetyldihydrazone. A modification of the

reported procedure of Curry et al. (30) was followed. Fi ce

grams (0.03 mole) of 2,6-diacetylpylridine were dissolved in~

150 ml of absolute ethanol and this solution was added dron-

wise to a stirred solution of 7 ml (0.15 mole) of 99-100%

hydrazine hydrate (Matheson, Coleman and Bell) kept at 50C.

Stirring was continued for six hours as the temperature

slowly rose to room temperature. The crude white product was

collected on a sintered glass funnel and recrystallized from

hot absolute ethanol giving 5.2 g (85%) of the dihydrazone:
mp 185-1900 [1it. (31) mp 1810l

Repeated attempts to purify the product by recrystalli-

zation were unable to lower the melting point to the liter-

ature values of 1810. However~ only one component was shown

by thin layer chromatography.

Anal. Called for C H3N: C, 56.53; H, 6.85; N, 36.62.

Found: C, 56.50; H, 6.85; N, 36.65.

2,6-Pyridinedicarboxaldehyde. A modification of the

method of Papadopoulos et al. (32) was employed. Seventy

grams of freshly prepared manganese dioxide were suspended

in 500 ml of chloroform containing 5.7 g of 2,6-pyridinedi

methanol. The mixture was stirred at reflux for five hours,

filtered with suction and the oxide washed with five 100 al1

portions of ether. The filtrates were combined and evapor-

ated under a stream of N2. The off-yellow residue was taken

up in a minimum amount of solvent containing 80% benzene and

20% ethyl acetate. This was placed on a 3 cm X 55 cm silica

gel (60-200 mesh, Matheson, Coleman and Bell) column and

eluted with the same solvent mixture collecting the middle

portion, 250 ml, in 50 ml fractions after discarding the

first 150-200 ml. Flow rate of the column was 5 ml per minute.

Each of the desired fractions was evaporated with a stream of

N2 and the melting points of the white crystalline residues
were checked. The product fractions were combined giving

3.0 g (53%) of the dialdehyde: mp 120-1220 [1it. (33) mp 1240!

2,6-Pyridinedialdihydrazone. A modification of the
method rep-orted by Stoufer and Busch (34) was followed. Two

grams of 2,6-pyridinedicarboxaldehyde (0.015 mole) in 35 ml of

warm absolute ethanol were added dropwise to a solution of

5 g (0.1 mole) of 99-100% hydrazine hydrate (Matheson,

Coleman and Bell) and 10 ml of absolute ethanol. Stirring

was continued at room temperature for four hours. After 30

minutes a fine white crystalline solid formed. The product

was stored in a freezer overnight, filtered on a glass

sintered funnel and recrystallized from hot ethanol giving

1.7 g (70%) of the dihydrazone: mp 144-1460 [1it. (35) mp

130-1350). Further recrystallization did not lower the

melting point and thin layer chromatography showed only one


Anal. Called for C7HgN5: C, 51.47; H, 5.56; N, 42.89.
Found: C, 51.44; H, 4.81; N, 43.64.

Freshly prepared Mn02. A modification of the method

reported by Sondheimer et al. (36) was followed. A solution

of 70 g (0.45 mole) of K~lnO4 and 700 ml water in a 2-liter

beaker was made acidic with 25 ml of cone. H2SO4. To this

hot, stirred solution of Kf~nOq was added slowly a solution of

100 g (0.60 mole) MnS04-H20 in 400 ml of water. After adding

the MlnS04 solution, excess K~InO4 was added until the intense

purple color was obvious. Stirring was continued for six

The brown MnO2 product was filtered with suction and

washed as many times as necessary to remove any excess per-

manganate ion. The product was then dried 24 hours at 1300,

ground to a fine powder giving 90 g (93%) of oxide and stored

over P4010-


Ni(ketomacr)C12-4H20. A solution of 0.951 g (0.004 mole)

of nickel(II)chloride hexabydrate and 0.653 g (0.004 mole) of

2,6-diacetylpyridine in 300 ml absolute ethanol was brought to

reflux at which point two drops of concentrated hydrochloric

acid were added. A solution of 0.765 g (0.004 mole) 2,6-

pyridinediacetyldihydrazone in 200 ml warm absolute ethanol

was added dropwise with stirring over a period of one hour to

the above solution. The mixture was refluxed for an additional

18 hours. The initial yellow-green solution turned to a

yellow-brown and finally to a green-brown heterogeneous mixture.

Two hundred milliliters of solvent were stripped off the hetero-

geneous mixture and 300 ml of diethyl ether were added to preci-

pitate additional product. The green product, filtered with

suction through a medium frit glass sintered funnel, was washed

via a Soxhlet extractor with refluxing methanol. The remaining~

grey-green solid was dried in~ vacuo over P 010 giving 1.913 g

(92%) of product.

Upon drying in an oven at 120 C for 24 hours a water loss

of four moles of water per mole of complex was observed.

After allowing the dried product to remain in contact with

atmospheric moisture, the four moles of water per mole of

complex were reabsorbed.

Anal. Called for Ni(C18 18 6)C12'4H20: C, 41.57,

H, 5.04; N, 16.16; Ni, 11.29; C1, 13.63. Found: C, 41.23;

H, 5.21; N, 16.03; Ni, 11.45, C1, 13.23.

Ni(aldomacr)C1 -AH O. A procedure similar to that re-

ported for the Ni(ketomacr!Cl24H20 preparation was followedl
using a solution of 0.540 g (0.004 mole) 2,6-pyridinedi-
carboxaldehyde and 0.951 g (0.004 mole) nickel(II) chloride

hexahydrate in 300 ml of absolute ethanol. This solution was

brought to reflux and two drops of concentrated hydrochloric
acid were added to the above solution. To this was added

dropwise a solution of 0.653 g (0.004 mole) 2,6-pyridinedi-
aldihydrazone in 200 ml absolute ethanol. The resulting
brown-green solid was worked up as in the preceding prepara-

tions yielding 0.919 g (49%) of product.

Upon drying in an oven at 1200C for 24 hours a water
loss of four moles of water per mole of complex was observed.

After allowing the dried product to remain in contact with

atmospheric moisture, the four moles of water per mole of

complex were reabsorbed.

Anal. Called for Ni(C14 1 6g)C12-4H20: C, 36.21; H, 3.91;
N, 18.12; Ni, 12.65; C1, 15.28. Found: C, 36.50; H, 3.43;
N, 18.68; Ni, 13.0; C1, 15.12.

Ni(ketomacr)(NO -26H 0. A procedure similar to that

reported for the Ni(ketomacr)C12-4H20 preparation was followed
using a solution of 1.163 g (0.004 mole) nickel(II) nitrate
hexahydrate and 0.653 g (0.004 mole) 2,6-diacetylpyridine in

300 ml of absolute ethanol. This solution was brought to

reflux at which point two drops of concentrated nitric acid
were added. A solution of 0.765 g (0.004 mole) 2,6-pyridine-

diacetyldihydrazone in 200 ml absolute ethanol was added drop-
wise. The resulting light green solid was worked up as in

the preceding preparations yielding 0.651 g (32.5%) of product.

Upon drying in an oven at 1200C for 24 hours a water loss

of six moles of water per male of complex was observed.

After allowing the dried product to remain in contact with

atmospheric moisture, the six moles of water per mole of

complex were reabsorbed.

Anal. Called for Ni(C18H1 )NO N0 ) 2 620: C, 35.48;

H, 4.96; N, 18.39; Ni, 9.64. Found: C, 35.31; H, 4.92;

N, 18.38; Ni, 9.48.

Ni (aldoketomacr)C1? 4H90. A procedure similar to that re-

ported for the preparation of Ni(ketomacr)C12-4H20 was followed

using a solution of 0.951 g (0.004 mole) nickel(II) chloride

hexahydrate and 0.540 g (0.004 mole) 2,6-pyridinedicarbox-

aldehyde in 300 ml 95% ethanol. When the solution was

brought to reflux, five drops of concentrated hydrochloric

acid were added. A solution of 0.765 g (0.004 mole)

2,6-pyridinediacetylydihydrazone was added dropwise over a

period of one hour. The light yellow-green solution turned

dark green as reflux was continued for 20 hours. A volume of

200 ml of solvent was stripped off and petroleum ether was

added to precipitate the product. The product was worked up

as in the preceding preparations giving 0.901 g (46%) of a

brown-green solid.

Anal. Called for Ni(C )Cl -6)C2'4H 0: C, 39.06; H, 2.87:

N, 17.08; Ni, 11.9. Found: C, 39.53; H, 3.01, N, 17.43;

Ni, 12.3.

Co(ke tomacr) C12 -4H20. A solution of 0.653 g (0.004 nol~e)

2,6-diacetylpyridine and 0.952 g (0.004 mole) cobalt(II)

chloride hexahydrate in 300 ml 95% ethanol was brought to

reflux. Two to three drops of concentrated hydrochloric

acid were then added. A solution of 0.765 g (0.004 mole)

2,6-pyridinediacetyldibydrazone in 200 ml warm absolute

ethanol was added dropwise over a period of one hour with

stirring. The mixture was then refluxed for an additional

20 hours during which time the royal blue solution turned

black. Approximately 150 ml of solvent were removed by

distillation and 300 ml of petroleum ether were added to

precipitate the product. The dark black-green precipitate

was filtered with suction, washed with refluxing methanol

in a Soxhlet extractor and dried in vacuo over P4010 giving

1.569 g (75%) of the desired product.

Anal. Called for Co(C 1 N)C12-4H20: C, 41.55; H, 5.06;

N, 16.15; Co, 11.33. Found: C, 41.13; H, 5.38; N, 16.40;

Co, 11.1.

Co(aldom~acr)C12.4H20. A prodecure similar to that

reported for the Co(ketomacr)C12-4H20 preparation was
followed using a solution of 0.952 g (0.004 mole) cobalt(II)

chloride hexahydrate and 0.540 g (0.004 mole) 2,6-pyridine-

dicarboxaldehyde in 300 ml 95% ethanol at reflux to which

had been added two drops of concentrated hydrochloric acid.

A solution of 0.653 g (0.004 mole) 2,6-pyridinedialdihydrazone

and 200 ml absolute ethanol was added dropwise over a period

of one hour. The black-green product was precipitated with

ether and worked up as in the preceding preparations giving

1.657 g (89%) of a black pow~der.

Anal. Called for Co(C14H10 6)C12-4H 20: C 62:H

3.91: N, 18.12; Co, 12.70. Found: C, 36.45: H, 3.61; N,

17.91; Co, 12.40.

Co (k etoma cr) (NO )2-2-'4H ?. A procedure similar to that

reported for the preparation of Co(ketomacr)C12*4H20 was
followed using a solution of 2.911 g (0.01 mole) cobalt(HI)

nitrate hexahydrate and 1.633 g (0.01 mole) 2,6-diacetylpyr-

idine in 500 ml absolute ethanol with two drops of concentrated

nitric acid. A warm solution of 250 ml absolute ethanol and

1.91 g (0.01 mole) 2,6-pyridinediacetyldihydrazone was added

dropwise to the above refluxing solution. After the reaction

was complete, 400 ml of solvent were stripped off and ether

was added to precipitate the product. The product was worked

up as in the preceding preparations giving 1.337 g (23%) of
a brown solid.

nl.Caled for Co(C18 18 6) (NO3)2-4H20: C, 37.71;

H, 4.57; N, 19.54; Co, 10.28. Found: C? 38.04; H, 4.34;

N, 19.82; Co, 10.4.

Cu(aldomacr)C12 H 0. A solution of 0.682 g (0.004 mnole)

copper(II) chloride dibydrate and 0.540 g (0.004 mole)

2,6-pyridinedicarboxaldehyde in 300 ml 95% ethanol was brought

to reflux. To this bright green solution were added two drops

of concentrated hydrochloric acid. A solution of 0.653 g

(0.004 mole) 2,6-pyridinedialdihydrazone in 200 ml absolute

ethanol was then added dropwise over a period of two hours

The reaction mixture turned dark green-brown and was refluxed
an additional 18 hours.

Two hundred and fifty mnilliliters of solvent were

removed by distillation. Ether was added to precipitate the

product. The black solid was filtered with suction, washed
with refluxing methanol in an extractor and dried in vacuo

giving 1.277 g (68%) of a fine brown-black powder.

Anal. Called for Cu(CL4H10"6)C12.4H20: C, 35.87;

H, 3.87; N, 17.93; Cu, 13.55. Found: C, 35.45; H, 3.41;

N, 17.41; Cu, 13.9.

Cu(ketomacr)(NO3)2 A procedure similar to that reported

for the preparation of Cu(aldomacr)C12 4H20 was followed using

a solution of 0.653 g (0.004 mole) 2,6-diacetylpyridine and

0.966 g (0.004 mole) copper(II) nitrate hexahydrate in 300 ml

95% ethanol. This solution was brought to reflux and two

drops of concentrated nitric acid were added. A solution of

0.765 g (0.004 mole) 2,6-pyridinediacetyldihydrazone in 200 ml

warm absolute ethanol was added dropwise over a period of

90 minutes. The mixture was refluxed for 20 hours. The prod-

uct was worked up as in the preceding preparations giving

1.416 g (70%/) of a black powder.

Anal. Called for Cu(C18H1186) (NO3)2:C, 42.73; H, 3.59;

N, 22.15; Cu, 12.60. Found: C, 43.19; H, 3.40; N, 21.83;

Cu, 12.7.

Cu(aldoketomacr)Clp2-220. A procedure similar to that

reported for the preparation of Cu(aldomacr)C12 4~20 was
followed using a solution of 0.540 g (0.004 mole) 2,6-pyridline-

dicarboxaldehyde and 0.682 g (0.004 mole) copper(II)

chloride dihydrate in 300 r1l absolute ethanol. Two drops of

concentrated hydrochloric acid were added to this refluxing

solution. A solution of 0.765 g (0.004 mole) 2,6-pyridine-

diacetyldihydrazone in 200 ml warm absolute ethanol was then

added dropwise over a period of one hour. The reaction

mixture was refluxed for 18 hours. Approximately 250 ml of

solvent were removed and 300 ml ether were added to pre-

cipitate the product. The green-brown solid was worked up

as in the preceding preparations giving 1.216 g (66%) of


Anal. Called for Cu(C16H 4N6))Cl2H2H20 C, 41.71; H, 3 93;

N, 18.24; Cu, 13.79. Found: C,41.62; H. 3.73; N, 18.33;

Cu, 14.2.

Attempted preparation of the ketomacr ligand. In a one-

liter round bottom flask 0.653 g (0.004 mole) 2,6-diacetyl-

pyridine was dissolved in 350 ml of absolute ethanol. This

solution was brought to reflux and one drop of concentrated

hydrochloric acid was added. To this clear colorless

solution a solution of 0.765 g (0.004 mole) 2,6-pyridinedi-

acetyldihydrazone in 175 ml absolute ethanol was added drop-

wise over a period of one hour. The reaction mixture

immediately turned lemon-yellow. Reflux was continued for

18 hours during which time a yellow solid formed.

The yellow product was filtered with suction, washed

with hot ethanol and dried over P 010 for 2 days. The prod-
uct did.not melt up to a temperature of 3000C

Anal. Called for C18H ,N : C, 67.91; H, 5.70; N, 26.39.
Found: C, 66.56; H, 5.91; N, 25.52.

Attempted preparation of the aldomacr ligand. A

solution of 0.135 g (0.001 mole) 2,6-pyridinedicarboxaldehyde

in 250 ml absolute ethanol was brought to reflux and one drop

of concentrated hydrochloric acid was added. To this

reaction mixture a solution of 0.163 g (0.001 mole) 2,6-pyridine-

dialdihydrazone in 250 ml absolute ethanol was added dropw!ise

over a period of 45 minutes. The colorless solution turned

yellow with the first drop of dihydrazone. Reflux was

continued for 20 hours. A yellow precipitate had formed by

this time and was filtered with suction, washed with ethanol

and dried over P4010 ~-in vacuo. The product did not melt up

to a temperature of 3000C.

Anal. Called for C14 10 6: C, 64.11; H, 3.84; N, 32.04.

Found: C, 62.63; H, 4.22; Nj, 29.02.

Attempted preparation of Cu(ketomacr)C12. A solution of

1.306 g (0.008 mole) 2,6-diacetylpyridine and 1.802 g (0.008

mole) copper(II) chloride dihydrate in 250 ml methanol was

brought to reflux and two drops of concentrated hydrochloric

acid were added. To this solution was added dropwise over a

period of one hour a solution of 1.530 g (0.008 mole)

2,6-pyridinediacetyldihydrazone in 350 ml methanol. Reflux

was continued for 24 hours and the lime-green reaction mixture

turned dark green.

The reaction mixture was concentrated to a volume of

100 ml and a volume of 500 ml of ether was added to precipitate

the dark green product. The product was filtered, washed with

ether and dried over P41 in vacuo giving 2.575 g of green


Anal. Called for Cu(Cl8H18N )C12; C, 47.74; H, 4.01;

N, 18.56; Cu, 14.03; C1, 15.66. Found: C, 43.39; H, 4.10;

N, 15.91; Cu, 15.20; C1, 13.92.

Attempted preparation of Zniketomacr)C12. A solution

of 1.306 g (0.008 mole) 2,6-diaccetylpyridin and 1.090 g

(0.008 mole) zinc(II) chloride in 250 ml methanol w~as

brought to reflux and one drop of concentrated hydrochloric

acid was added. To this was added dropwise a solution of

1.530 g (0.008 mole) 2,6-pyridinediacetyldihydrazone in

350 ml methanol over a period of one hour. Reflux was

continued for 24 hours during which time the reaction mix-

ture had turned lemon yellow.

A volume of 200 ml of methanol was distilled off and

the remainder of the solution was further concentrated using

a flow of Ni2 gas to a final volume of 100 ml. The yellow

solid was filtered with suction, washed with ether and

dried over P4010 in vacuo giving 1.859 g of product.

Anal. Called for Zn(C18H18N6)C12: C, 47.55; H, 3.99;

N, 18.48; Zn, 14.38; C1, 15.60. Found: C, 57.71; H, 5.15;

N, 21.02; Zn, 3.26; C1, 7.25.

Attempted preparation of Mln(ketomacr) (C104}3. A

lavender solution of 2.855 g (0.004 mole) hexaureamanganese(IrI)

perchlorate in 250 ml absolute ethanol, to which had been

added two drops of concentrated perchloric acid, was added to

0.653 g (0.004 mole) 2,6-diacetylpyridine. The now red-violet

solution was brought to reflux and a solution of 0.765 g

(0.004 mole) 2,6-pyridinediacetyldihydrazone in 200 ml

absolute ethanol was added dropwise. The red solution turned

yellow immediately. Reflux was continued overnight and a

yellow-orange solid w~as filtered, washed with ethanol and

dried in vacuo over P4010 giving 1.230 g of product.

Anal. Called for Mn(C 31N)C0)(C0 C, 32.19; H, 2.70;

N, 12.51; C1, 15.84. Found: C, 58.41; H, 5.26; N, 22.76;

C1, 3.97.

Attempted preparation of Mn(ketomacr)C19. A solution of

1.306 g (0.008 mole) 2,6-diacetylpyridine and 1.583 g (0.008

mole) manganese(II) chloride tetrahydrate in 250 ml methanol

was brought to reflux and two drops of concentrated hydro-

chloric acid were added. A solution of 1.530 g (0.008 mole)

2,6-pyridinediacetyldihydrazone in 350 ml methanol was

added dropwise over a period of one hour. Reflux was continued

for 24 hours with the reaction mixture turning a dirty yellow

color during this time. The yellow-orange solid that had

formed was filtered with suction, washed with ether and dried

over P4010 in vacuo giving 2.421 g of product.

Anal. Called for Mn(C S~N)Cl : C, 48.67; H,'4.08:

N, 18.92; MIn, 12.37; C1, 15.96. Found: C, 65.56; H, 5.42;

N, 23.51; M~n, 1.78; C1, 2.14.

Attempted preparation of Hg(ketomacr) (NO3)2. A solution

of 0.653 g (0.004 mole) 2,6-diacetylpyridine and 1.370 g

(0.004 mole) mercury(II) nitrate monohydrate in 200 ml

absolute ethanol was brought to reflux and a pinch of NH4N03

solid was added. To this solution was added dropwise a

solution of 0.765 g (0.004 mole) 2,6-pyridinediacetyldihydra-

zone in 250 ml absolute ethanol. The white cloudy reaction

mixture turned yellow as the hydrazone was added. Reflux

was continued for 24 hours. Free mercury was observed in

the bottom of the reaction flask at the conclusion of the


The reaction mixture was filtered with suction and a

yellow-oirange solid was recovered. The product was washed

three times with ethanol and twice with ether and then dried

~in vacuo over P .1

Anal. Called for Hg(C 81N6(0)(NO C, 33.61; H, 2.8i0;

N, 17.42; Hg, 31.21. Found: C, 38.57, H, 3.44; N, 16.95;

Hg, 30.65.

Sodium acetylide. A method similar to that of Rutledge (37)

was followed using a suspension of sodium dispersion in hexane.

Acetylene gas was purified by passing it through a dry ice-

acetone trap and concentrated sulfuric acid to remove any


The sodium dispersion was prepared by dissolving six

grams of metallic sodium in 150 ml of liquid NH3. Precooled

hexane was then added to the blue solution of Na and NH3. The

NH3 was allowed to evaporate at room temperature with stirring

resulting in a fine dispersion of sodium metal in hexane.

The purified acetylene was bubbled into the sodium dis-

persion at the boiling point of hexane (690C) for 3-5 hours.

The greyish sodium dispersion became beige in color at the

completion of the conversion of the sodium metal to the


The hexane was removed under reduced pressure. A

pale yellow solid was recovered. Themonosodium acetylide

product was identified based upon the results of acid-base


2,6-pyridinedicarboxylic acid chloride. A sample of

16.8 g (0.1 mole) 2,6-pyridinedicarboxylic acid was placed in

a 250 ml three-neck round bottom flask fitted with a water

cooled reflux condenser and magnetic stirring bar. A volume

of 100 ml thionyl chloride was added and the mixture was

refluxed for 24 hours. The solid acid dissolved and a yellow!-

orange solution resulted. The excess thionyl chloride was

removed with reduced pressure and a pink solid remained.

The product was purified by sublimation giving 16.9 g

(81%) of the pure white crystalline acid chloride; mp 58-600

[1it. (38) mp 56-580]

Attempted reaction of 2,6-pyridinedicarboxylic acid

chloride with sodium acetylide. In trying to prepare the

acetylene derivative of 2,6-pyridinedicarboxylic acid chloride,

(Fig. 6). Attempts were made to carry out this reaction in

various solvent systems: diethyl ether, 1,4-dioxane, tetra-

hydrofuran, liquid SO2 and benzene. Care was taken to dry each
solvent. The organic solvents were observed to become warm or

to boil upon the addition of solid sodium acetylide to a


solution of 2,6-pyridinedicarboxylic acid chloride.

A violent reaction was observed at the mouth of the

flask when sodium acetylide was added to a solution of the

acid chloride in liquid SO2. A yellow-orange flame resulted
each time.

The two solids, sodium acetylide and 2,6-pyridinedicar-

boxylic acid chloride, were mixed together under a N2 atmo-

sphere resulting in a violent reaction evolving heat and

smoke. Charring of the solids resulted. In each case prod-

uct isolation was attempted with no success. Either the acid

chloride or the parent acid was recovered.


Spectrometers. Solution spectra in the visible and

ultraviolet regions were obtained using a Cary Model 15

recording spectrophotometer. The solid state diffuse re-

flectance spectra were obtained using a Cary Model 1411

Diffuse Reflectance Accessary in conjunction with a Cary

Model 14 recording spectrophotometer. Magnesium carbonate

was employed as reference material.

Infrared spectra were obtained using a Perkin-Elmer

Model 137 NaC1 prism and 237B grating spectrophotometers;

also employed was a Beckman Model IR-10 grating spectro-

photometer. All spectra were calibrated with polystyrene.

The pressed KBr pellet technique was used.

Electron spin resonance spectra were obtained on finely

ground samples using the Varian Associates Mlodel E-3 recording


Mass spectra were obtained on AEl Scientific Apparatus

Model MS-30 double-beam, double focusing mass spectrometer

equipped with a DS-30 data system. Each solid sample was

run by direct introduction probe. Probe temperature ranged

from 2000 to 3400C.

H nmr spectra were measured on Varian Associates Model

A60-A and XL-100 nuclear magnetic resonance spectrometers.

Conductance apparatus. Conductances were measured using

an Industrial Instruments, Inc., Model RC-18 Conductivity

Bridge and a cell of constant 1.431+0.006 cm- at constant

temperature of 25.00C+0.010, maintained by the use of a water


Melting point apparatus. A Thomas Hoover "Uni-melt"

capillary melting point apparatus was employed; the tempera-

tures are uncorrected.

Analyses. Carbon, hydrogen, nitrogen, and halogen anal-

yses were performed by Galbraith Laboratories, Inc., Pennin-

sular Chem. Research, Inc., and Atlantic Microlab, Inc. Al-

though the metal analyses reported herein were performed by

Galbraith Laboratories, Inc., they were initially obtained

by using a Perkin-Elmer Model 290B atomic absorption spectro-

meter. All samples were analyzed in aqueous solutions after

digesting the complex almost to dryness in 20 ml of a 1:1

mixture of concentrated nitric and perchloric acids.

Gouy apparatus. Magnetic susceptibilities were deter-

mined by the Gouy method using equipment described previously

in greater detail (39). The maximum field strength attained

was 66200oersteds. The magnetic field was calibrated using

mercury(II) tetrathiocyanatocobaltate(II) (40). The cryostat

and temperature control apparatus used were of the basic

design of Figgis and Nyholm (41). Temperatures between 1150

and 4000K could be maintained within +0.10 as determined by a

platinum resistance thermometer. The sample tube was made of

a cylindrical piece of quartz, approximately 3.0 mm inside

diameter and 16 cm in length which was sealed at one end.

Approximately 15 cm was used for containing the sample volume.

It was suspended in the cryostat from a semi-micro balance by

a gold chain attached to a tapered Teflon plug. The diamagnetic

correction for the tube was measured as a function of temper-

ature. A Mettler Model B-6 semi-micro balance of 0.01 mg

sensitivity was used to measure the force exerted by the magnet-

ic field upon the sample.

Polarograph. A Sargent Model XVI Polarograph was used in

conjunction with a dropping mercury electrode for obtaining

the polarogram of the Ni(ketomacr)C12 complex. A drop rate of
one drop per two seconds was employed. Water, used as solvent,

was deoxygenated by passing a stream of N2 gas through the water

for 15 minutes. The supporting electrolytes used were 0.1 M KC1

and 0.1 M tetraethylammronium bromide. A lx10-3% solution of

Triton X-100 was used to suppress maxima. A N2 blanket was kept

over the solution under study to prevent redissolution of oxygen.

Electrolysis Cell. A mercury cathode cell was prepared

using the design of A.D. Meloven described by Willard, Merritt

and Dean (42). A platinum spiral was used as the anode. The

cell, 7 cm in diameter by 14 cm high, was fitted with a three-

way stopcock. One arm of the stopcock was connected to the

leveling bulb of mercury. The other arm permitted removal of

the electrolyte. A d.c. power supply was used to supply the




The macrocyclic complexes containing Ni(II) ion proved to

be more easily prepared than those of Co(II) or Cu(II). The

preparation of Ni(ketomacr)C12 was attempted first because of

the availability of DAP and DAPH, the precursors of the macro-

cyclic ligand. The preparation of H~i(ketamacr)C12 had beeni
claimed previously (43) but its characterization was incomplete.

Although the infrared spectrum and analytical data are identical

(within experimental error) to those obtained for the product

isolated during the course of these investigations, the room~

temperature magnetic susceptibility obtained by this writer

differs markedly from that reported previously. A more extensive

comparison will be made later.

It was anticipated that a successful synthetic technique

developed for this complex could be applied to the preparation

of other complexes employing similar ligand types, i.e., PDC and


The "template reaction" method was employed first because

a vast majority of the successful syntheses of complexes which

incorporate a macrocyclic ring circumscribing a central metal

ion have only been possible using the template technique videe

suEra). It was observed that the dropwise addition of a solution

DAPH in ethanol to a refluxing solution of DAP and NiC12 in

ethanol (containing a few drops of concentrated hydrochloric

acid as catalyst) produced the desired macrocyclic complex

after about 18 to 24 hours of reflux.

Attempts were made to prepare Ni(ketomacr)C12 by

changing the sequence of the addition of reactants. An etha-

nolic solution of DAP and DAPH were added simultaneously to

NiC12. Neither of these two variations proved to produce the
desired complex as evidenced by poor elemental analysis; thus,

these two methods were not used. Goedken (25) failed in

attempts to prepare any nickel macrocyclic complex. But the

order of addition of hydrazine and DAP he employed is different

from that reported in this work; i.e., he added hydrazine

directly into an acetonitrile solution of a Ni(II) salt and DAP

The green, hydrated Ni(ketomacr)C122 was precipitated,
after distilling off approximately half of the solvent, by

adding diethyl ether. The light green product, after being

filtered and dried at room temperature over P 010 in vacuo,
corresponded to a tetrahydrate. The nitrate salt, prepared in

a similar manner, was isolated as a hexahydrate. The complexes

Ni(aldomacr)C124 20 and Ni(aldoketomacr)C12-4H20 were prepared
in a similar manner; both complexes analyzed as tetrahydrates.

The Ni(aldoketomacr)Cl2-4H O was a brown-green powder; the

Ni(aldomacr)C12 420 was a darker brown-green powder. Inspection
of the complexes under a microscope revealed the complexes to

be homogeneous, amorphous solids.

The macrocyclic complexes of cobalt and copper were

prepared using similar techniques and are described in the

Experimental section. The three cobalt complexes

Co(ketomacr)(NO3 2-4H20, Co(aldoketomacr)C124H20,

Co (aldomracr)C12 4H20 all analyzed as tetrahydrates. No color

change was observed for the cobalt complexes as the water was

driven off with heat. The three copper complexes prepared and

characterized are Cu (aldomacr) C12 4H20, Cu(aldoke tomacr) C12 -2H20,

and Cu(ketomacr) (NO3 2. Again no color change was observed wJith

water loss. The reason that Cu(ketomacr)(NO )2 should be

isolated as the only anhydrous complex is not apparent. It is

also noted that even though several attempts were made to pre-

pare Cu(ketomacr)C12, no pure product was isolated.
It was observed that the waters of hydration could be re-

moved by drying the samples overnight in an oven set at 1200C.

The waters of hydration were subsequently readsorbed when the

complexes were allowed to remain in the atmosphere a few days.

The gain and loss of water was accompanied by a barely per-

ceptible, reversible color change for the nickel complexes only,

i.e., from a yellow-green characteristic of the anhydrous nickel

complex to the light green color of the hydrate.

The character of this water is uncertain; but three

possibilities exist: lattice water, coordinated water, and

solvolytic water, i.e., water contained in the form of a carbinol-

amine (44). A combination of these three can not be excluded a

priori, either. Because there is, effectively, no color change

when the complexes are dried, it is concluded that the water pres-

ent is not in the coordination sphere of the metal ions. The

removal of a coordinated species from a metal ion changes the

ligand field about that metal ion and accordingly, the relative

energies of the spectroscopic terms. Familar examples of this

phenomenon include the conversion of the red, hexaaquocobalt(II)

chloride to the blue, anhydrous cobalt(II) chloride.

For the water to be present as a carbinol-amine, the for-

mation of the complex must stop at an intermediate stage in the

Schiff base condensation reaction on one end and reverse itself

one step on the other. Since two of the C=N bonds have already

been formed in the DAPH, it would seem unlikely that such a

reversal would occur if the ring was not puckered. However, if

the ring was puckered, the C=N would be very susceptible to

nucleophilic attack to form the carbinol-amine (45). The

presence of four C=N bonds in each macrocycle would correlate

with four H20 molecules included as carbinal-amines, but would

not account for the presence of six H20 groups in Ni(ketomacr)-

(NO3 )26H20, the two H20 groups in Cu(andoketomacr)C12-2H20,o

the anhydrous nature of Cu(ketomacr)(NO3 2. The problem could
not be resolved by infrared spectroscopy because of the extreme

broadness of the characteristic absorptions.

Goedken et al. (25) have reported the preparation and

crystal structure of iron(II) and cobalt(II) complexes of

ketomacrocycle. The crystal structure of Fe(ketomacr)-

(CH CN)2 (C10 )2 showed the nitrogen atoms in the macrocyclic

ring to be coplanar with the iron(II) ion. The axial ligands,

acetonitrile, were above and below this plane. Therefore,

because there is no structural evidence for puckering of the

ring, it is concluded that the water, contained in the macrocyclic

complexes prepared in this study, does not result from solvol-

ysis of the C=N linkages.

It was found that if a small excess of DAP was used ini the

synthesis, the product gave a consistently better analysis.

The method whereby the hydrazone was added dropwise to a

solution of the metal ion and DAP or PDC was used exclusively

When the products did not give the expected elemental analysis,

because of apparent excess organic material, it was found that

the solid product could be extracted with absolute methanol

in a Soxhlet extractor to produce a product which gave the ex-

pected analysis. This Soxhlet extraction technique was used

on all of the complexes except for Ni(ketomacr)C124H20 and

Ni(ketomacr) (N03 2'6H20 for which it was unnecessary.

Recrystallization to improve the Durity of a particular

complex usually proved to be ineffective because of its limited

solubility in either water or ethanol. Although other solvents

were tried, the complexes were found to be more soluble in

either water or ethanol. Saturated solutions of the complexes

in water ranged from 10 2M to 10 M; thus recrystallization on

a large scale was not feasible.

Infrared Spectra

The infrared spectra of the nickel, cobalt, and copper

complexes, Figures 7, 8, 9, and 10, contain a recurrent absorption

near 3350cm-1, which is customarily assigned to OH or NH stretch-

ing vibrations. Incorporation of adsorbed water was assigned as

the source of this band as a result of the previous consider-

ations. Also the intensity of this band could be reduced, but

Ni (aldomacr) C12 -4H20

~I Ni(aldoketomacr)Cl2*4H20

3000 2000. 1500 /oo 00 8b 00- 700
Frequency (cm-1)
Fig. 7 Infrared spectra of Ni(aldomacr)C1 41120,
Ni (aldoketomacr) C12 H20, and Ni (k omc)C2 2

Co (aldomacr) C12 -4H20

Co(ketomacr)C12 420
3000 1000 1500 1100 1000 900 800 700
Frequency (cm-1)
Fig. 8 Infrared spectra of C1(aldomacr)C12 4H20 and

C)adkeoar)12 2

Cu(aldokomacr)C1 2H20O

3000 2000 1500 12.00 1000 900 800 700
Frequency (cm-1)

Fig. 9 Infrared spectra of Cu(aldoketomacr)C12-2H20
and Cu (aldomacr) C12 -4H20 .

I I ~Cu(ketomaler) (NO3 2
3000 2000 \$00 1200 1o000 900 800 700
Frequency (cm-1)
Fig. 10 Infrared spectra of Co(ketomacr) (NO )24H20
Ni(ketomacr) (NO3 2 6H20 an uktoarN3)2

not entirely removed, by carefully drying the complex and Dre-

naring the KBr pellet under a dry nitrogen atmosphere. These

spectra were presented because of their convenient size. The

spectra of the same compounds were obtained using a Beck~man

Model IR-10 grating spectrophotometer, however, these IR-10

spectra contained no absorptions which were not identifiable in

the soectra presented herein.

The infrared spectra of the metal complexes tend to be

very broad and poorly resolved as is characteristic of mrany

highly conjug~ated macrocyclic complexes (46). Following the

approach of Nakamoto (46), the infrared spectra of the complexes

might be expected to be a composite of the soectra of the

corresponding precursors: DAP and DAPPH, PDC and PDCH, PDC and

DAPH. Similarities are found except for the absence of the rela-

tively sharp primary amine and carbonyl stretching absorptions.

The C=N absorption in conjugated systems interact to a

large extent with other double bonds in the compound and thus it

is frequently difficult to assign a definite position to this

group. Busch (17), however, has reported the imine absorption

of a related macrocyclic system to occur at 1570 em- Other

studies (47) have reported completed imnine absorptions between

1610 and 1615 cm- Bellamy (48) has reported C=N stretching

vibrations as high as 1690 em-1~ The profusion of bands in the

1600 cm- region arising from the C=N grouping and the Dyridine

ring absorptions made absolute assignment of bands in this

region difficult. To complicate matters even more, the OH

deformation vibrations are also found in the 1600 em1 region.

The C=N vibrations could be strongly coupled wJith the ring

vibrations in which case, it would be impossible to make an

unequivocal assignment. However, it is still very temating to

try to make some assignment of the C=N absorption to the 1600 cm-1

region. Tables A-1, Al-2, and A-3 of the Appendix list selected
absorptions and their corresponding assignments.

All four nickel complexes show a broad, intense band ca

1585 cm- which contains several weak shoulders. This band is

attributed to a composite of ring vibrations (Bands I and II)

and C=N stretching vibrations. The C=N band is better resolved

in the soectra of Ni(ketomacr)+2 salts. The very strong, broad

absorption in the spectrum of Ni(ketomacr)(NO3 2-6H20 lying in

the 1400-1350 cm- region is assigned to an ionic NO)3 stretching

vibration (46). Corrdinated nitrate complexes have NO2 anti-

symmetric and symmetric stretching vibrations in the 1531-
141c-1 ad19-23m-1 regions respectively (49).

The infrared spectra of the cobalt complexes also contain

the broad undefined absorptions in the 1650-1575 em-1 region

which can be assigned to a combination of acyclic C=N and ring

vibrations. The spectrum of Co(ketomacr)(NO3 2-4H20 is better
resolved in this region, exhibiting three strong peaks at

1645, 1600, and 1550 em- The band at 1600 em- is attributed

to the C=N stretching vibration. The broad band centered on

1380 em-1 is assigned to an ionic NO3 stretching vibration (4j).
The infrared spectra of the copper complexes are very

similar to those of the nickel and cobalt complexes. Broad

unresolved bands are again observed. The broad band in the

1640-1560 em- region is present and is attributed to a combin-

ation of ring and acyclic C=N stretching vibrations. The

NO3 stretching vibration is observed at 1380 em-1 n h
spectrum of Cu(ketomacr)(NO3 2.
The infrared spectra of the complexes prepared support the

assertion that a Schiff base condensation has occurred, i.e.

there are neither characteristic C=0 nor NH2 stretching vibra-

tions observed; but it is not possible to make an unequivocal

assignment of the C=NJ stretching vibration. Nor is it possible

to conclude on the basis of these data that the ligand exists

in an imine form.

Electronic Spectra

A study of the ultraviolet-visible absorption spectra of

the complexes was undertaken to ascertain the coordination

geometry present in the macrocyclic compounds. As mentioned

previously, the solubilities of the complexes limited the choice

of solvents; but, event then, the solute concentrations of all
but Ni(ketomlacr)C124 20adN~eomc)N326H20 were, less

than 103M However, the large molar absorntivity coefficients

of the complexes permitted a characterization under the very

dilute conditions dictated. The Ama and molar absorptivity

coefficients are listed in Table 1 together with the absorations

observed in the diffuse reflectance spectra.

The spectra of aqueous solutions of both Ni(ketomacr)C12 4H20

and Ni(ketomacr) (NO3 2-6H20 contained absorpt~ions at 267 andl 340
4 --1 -1
nm with molar absorptivity coefficients of magnitude 10 M em

The complexes Ni(aldomacr)C12 420 and INi(aldoketomacr) C12 4r20
showed three absorptions at 265, 320, and 450 nm with molar

absorptivity coefficients of similar magnitude. There wias no

Table 1. Electronic spectral data

-1 -3
cm x10













i 07








Ni(ketomacr) (NO )2-H0

Ni~adoktomar)C2 4 20

Ni(aldoktmacr)C12 H20


Co(ketomacr)(NO3 ~20















Table 1. Continued

-1 -3 --l 3 -1 --1l
Compound A (nm) cm x10 el Mc

Cu(ketomacr)(NO3 2 255 39.2 4.16
320 31.2 2.53
450 22.2 0.42
375*' 26.7--

*Diffuse reflectance absorption
fDiffuse reflectance snectrum~ contains no absorption maxima

change in the solution soectra of the complexes upon standing

for several days. The diffuse reflectance spectra of

Ni(ketomacr)C12-4H20 and Ni(ketomacr)(NO )2-6H20 (Fig. 11)
contain absorations at 360 nm which can be considered to be

the same one arising at 340 nm in the solution spectra.

The solution spectra of Ni(ketomacr)C12 H20 and

Bi(ketomacr) (NO3 2.6H20 contained no well-defined absorptions
in the visible region. The observation is not surprising if

one expects to see d-d transitions at the concentrations

required. The spectra of Ni(aldomacr)C12 H20 and

Ni(aldoketomacr)C12-4H20 do show a band in the visible region
of the solution snectra at 450 nm and the diffuse reflectance

spectrum of Ni(aldonacr)C12- 4H20 shows a weaker shoulder at 455
nm. The diffuse reflectance spectrum, obtained by spreading

the powdered samale on 3M masking ta~c, of Ni(aldoketomacr)-

C12-4H20 is a straight line. Only the spectra of five of the
complexes are illustrated because 04 the lack of characteristic

features in the spectra of the other complexes.

The solution electronic soectra of the cobalt and conner

complexes also contain the twso characteristic bands in the

ultraviolet region, viz., at 260 and 330 nm. The diffuse

reflectance spectra (Figures 12 and 13) contain broad bands

around 350-450 nm, except for the spectrum of Co(aldomacr)

C12-4H20 which is a straight line. However, the molar absorn-
tivity coefficients of the complexes in this region are too

large to result from d-d transitions (14); rather, they are




o j

o C)


O r

LD o


o U



\ -

asueqiosqtr an~~~Ta~






o 0r



E0 o




to u

o o


0 )



c I

a3ueqlosq~ anr~~~a~

asueqiosqtr ahT~eTa~



O r






o E

d n

to '--

O bD



O n


attributed to charge transfer bands. The charge transfer

bands may vary in position and intensity from one system to

another. The band observed near 320 nm in the solution

spectra may be present in the reflectance spectra; but it is

so poorly resolved as to escape detection. Comparison of the

solution electronic spectra with the diffuse reflectance

spectra failed to reveal additional similarities.

The transitions in the ultraviolet region of the spect;ra

are assigned to charge transfer absorptions localized upon the

(aromatic) ligands. The lower energy (visible) transitions

are attributed to charge transfer bands to the complex. Thus,

the coordination geometry cannot be inferred from either the

solution electronic spectra or the diffuse reflectance spectra

of the complexes.


The conductance of an aqueous solution of each of the

marcocyclic complexes was obtained in order to infer whether

or not anion coordination is present. The molar conductances

for the complexes at selected concentrations are reported in

Table 2. The slight solubility of several of the complexes

prevents a straightforward comparison of the molar conductance

at the same concentration, however, trends are apparent.

Several electrolyte types, i.e., 1:1, 1:2, 2:1, 1:3, 3:1, and

1:4, and their molar conductance in water for lx10-3M solutions

are listed in Table 3. The values obtained for the complexes

lie in the range of 2:1 electrolytes. Therefore, it is con-

cluded that coordination with the anion is unimportant. A

standard curve for a 2:1 electrolyte (Fig. 14) was prepared using

Table 2. Selected molar coniductance and limiting conductance
of prepared complexes in H20 at 25.00C


2.2 x10
4.48x10 4
4.31x10 4
3.06x10 4
1.66x10 4
2.52x10 4

5.0 x10
5.0 x10


Ni(ketomacr)(NO3 2.6H20
Ni(aldomacr)C12 H20
Co(ketomacr)C12 H20
Co(ketomacr)(NO3 2-4H20
Cu(ketomacr)(NO3 2




Table 3. Known electrolyte types and their molar
Electrolyte Molar Conductivity*

(PtNH 3C13C 96.8
K[Pt(NH )C151 106.8
NaC1 123.7

1:2 & 2:1.

CaC12 260.8
[Co(NHI3 501]C12 261.3
[Co(NH3 5BrjBr2 257.6
[Cr(NH3 5C1]Cl2 260.2
[Cr(NH3)5BrjBr2 280.1
IPt(NH3 4C12]C12 228.9
K2[PtC16] 256.8
1:3 &r 3:1

LaC13 393.5
[Co(NH3)6]C13 431.6
[Co(NH ) ]Br3 426.9
[Cr(NH3 6]C13 441.7
[Pt(NH )5C1]C13 404.0


[Pt(NH3) JCl4 522.9

*l x 10 31 solutions in H20 at 2.0 5)



en C
a, Lc


b Di








a oo


< o

023 N 01
000 0


1Ju *


8 1

:t \








0 O

r 1CI a\
-- u



c7 au


0 '

O 0

N N-

u 0

O 00
6 5 5d

0 CO

P o
o i

o~ ;

,g I

je s
I d












i 00
n C

to t

E on






c o a c







*I e;




standard solutions of BaC12. The plots of molar conductance

versus the square root of concentrations of all the complexes

(Figs. 15, 16, and 17) and the standard curve for BaC12 are
similar. At concentrations larger than 10 M there is a

linear portion of the curve which then breaks upward into a

steeper slope at the lower concentrations. The limiting con-

ductance of each of the complexes is comparable to that observed

for BaC12. The limiting conductance was obtained by linear

extrapolation of the straight portions of the molar conductance

versus square root of concentration curve (52). These values

are reported in Table 2. The limiting conductance for a

5.0x10 71 solution of BaC12 a on ob 4(h m-

Those of the complexes are listed in Table 2.

Magnetic Susceptibility

The magnetic susceptibilities of the complexes were deter-

mined using the Couy method, except for Cu(aldomacr)Cl2-4H20

in which case the nmr technique was used. The magnetic data

for the macrocyclic complexes, discussed herein, are listed in

Table 4.

The room temperature magnetic moments for Ni(ketomacr)-

C1-H n N~etmc)(O '6H20 were found to be 3.89 and
3.77 BM respectively. These moments are considered to be high

for high-spin octahedral nickel(II) ion (53); but they fall

within the range of values reported for either high-snin tetra-

hedral or some five-coordinate nickel(II) ions. Effective

magnetic moments from 3.69 to 3.85 BM have been reported (5:)

for five-coordinate nickel(II) comolexes. Complexes of

Table 4. Magnetic data



Ni(ketomacr)(NO )2'6H20


Ni(aldomacr)C12 M20


Co (ket omacr) (N03 2 t20

Co(aldomacr)C124E 0



Cu(ketomacr)(NO3 2

(cg-s units)

4 .610









ToK) Go Ieff !I)

3001 -6 3.89
308 3.40

302 -10 3.77

301 -- 2.08

300 -- 2.62

300 -20 2.86

300 -- 4.73

300 -30 2.51

308 -- 1.19(nmr)

301 -- 1.81

299 -14 1.50

*TI = Temoierature Indecondent Moment

nickel(II) with N-methyldabconium ion (L ), H20 or NH3, and

halide ion were prepared (55), NiL(H20)C13, and found to be

nonionic. The magnetic data for these five-coordinate

complexes are consistent with either a trigonal bipyrimidal

structure or a square pyramidal structure. Based upon the

electronic spectra obtained, the authors concluded that the

structures were trigonal bipyrimidal.

A sample of Ni(ketomacr)C12-4H20 was dried to determine

what effect the presence of the water has on the magnetic

susceptibility. The room temperature moment of the anhydrous

product observed was 3.40 BrI, about 0.5 BM less than that

found for the hydrated complex. This change is considered

to be significant because the moment of the anhydrous complex

lies in the range of high-spin octahedral nickel(II) ions

(14, 53). There is some definite interaction of the water

molecules with the paramagnetic center of the complex as

evidenced by the change in the magnetic moment. The tyoe of

interaction involved is not fully understood. Weller (43)

reported the preparation of Ni(ketomacr)C12; but his product
is not the same as that prepared during the course of this

study. For example, he reported a magnetic moment of 1.03 BHl:

but, using the separation techniques reported herein, an

anhydrous product having a room temperature moment of 3.40 B11

was isolated. The differences in separation techniques are

considered to be the major factor governing the lack of corres-


Most transition metal compounds are magnetically dilute,

i.e., their paramaginetic centers are isolated from each other

by inert ligand molecules. In such compounds the paramagnetic
ions act independently of each other. In these cases the

idealized behavior for the variation of magnetic susceptibility

with temperature is the Curie lawj, Xm = C/T, where Xm is the

susceptibility per mole of paramagnetic material corrected for

the diamagnetism of the constituent atoms, C is the Curie

constant and I the absolute temperature. However, the majority

of the paramagnetic substances do not obey this law but a

modified version called the Curie-W~eiss law, X_ = C/(T+9), in

which 9 is an empirical quantity and is a measure of the

deviation from the idealized Curie law description. This para-

meter is determined by plotting 1/X, versus T and determi~ninF

the intercent at 1/Xm = 0. For magnetically dilute paramaginetics,

9 is usually a small quantity.

The data obtained from temperature-decendent magnetic

susceptibility studies of Ni(ketomacr)C12-4H20, Ni(ketomacr)

(NO3 26H20, Co(ketomacr)C12-4H20, Co(aldomacr)C12-4H20, and
Cu(ketomacr)(NO03 2 are given in Tables 5-9 and represented

graphically in Figs. 18-22. All the comolexes measured followed
the Curie-Weiss law with small 9 values listed with the temo-

erature-dependent data.

Because the effective moments for the Ni(ketomacr)C12 420

and Ni(ketomacr)(NO3 2-6H20 were considered to be large for a

high-spin d8 ion in an octahedral or tetrahedral field, the

moments were checked using an nmr technique (56). This nmr

technique allows on~e to determine the paramagnetic moment of a


Table 5. Temperature-dependent magnetic
of Ni(ketomacr)C12-4H20.





8 .47








Temperature, OK











6 = -60, Curie

X = 0.261 x

X~ = corrected


molar magnetic susceptibility

Table 6. Temperature-dependent magnetic susceptibility
of Ni (ketomacr)(NO3 2 -6H20 .

Temperature. oK x103

260 6.74

240 7.31

220 7.93

200 8.77

180 9.69

160 10, 88

150 11.46

140 12.25

130 12.98

120 13.89

e = -100

Xda= 0.275 x 10-3

Table 7 Temperature-dependent magnetic
of Co (ketomacr) 012 4H20 .












Temperature, oK










6 = -200

Xdia = -0.261 x

Table 8. Temperature-dependent magnetic susceptibility
of Co (aldomacr) C12 -4H20 .

X ix103










Temperature, oK










S= -300

Xdia = -0 .214 x

Table 9. Temperature-dependent magnetic susceptibility
of Cu(ketomacr) (NO3)2.

Temperature, oK x xl3

260 1.40

240 1.52

220 1.62

200 1.79

180 1.94

160 2.19

150 2.37

140 2.52

129 2.68

B = -140

Xdi = -0.196 x 103



Lo h

o C)
N co

o o

O 60
HO i
00 m

rlo E

-j CL
E o

N O*

IO 04
O 0-
I aL


x ~ ~~ ~ o 0 o a c

\ 4 N O co~ aD N

o ,D

o n



"N U




ar o

a r


O -H
N *

O C)


oo ar



00 at




I ~ t I t i l i r



O rl



co or


o u

o OO

\ r-4 H 4

O 'ir

a ao

\ Ho~ ~





O -H

"N o


oc ~

-3 rd

0 E


\~- 0 .3c-



.0 ma
o n4

\ Hiu

xo a 0 0
100 0 O O O
H > \LO LA 4CO N

solution of the complex under study by measuring the shift

difference of the nuclei (1H) of an indicator (t-butanol) in

two compartments of the nmr tube. A capillary melting point

tube was half filled with a 2.99x10-6 mole/ml solution of

Ni~ktomcr)C2-420 in 2% aqueous t-butanal and sealed. Ti
capillary tube was placed in a standard nmr tube which was

filled just to the top of the capillary tube with a 2%

solution of t-butanol. An nmr spectrum was run on a 60 M~Hz

instrument and two resonances due to the shift difference in

the methyl protons on the t-butanol were observed, one from the

unshielded solution of t-butano1 (outer compartment) and the

other (less intense) from the shielded protons (inner compart-

ment). The methyl protons in the inner compartment, shielded

by the paramagnetic nickel ion of the complex, were shifted

upfield from the unshielded methyl protons of the reference.

The shift difference of the absorption signals of the t-

butano1 protons in the two compartments is related to the

volume susceptibility of the two solutions and, therefore, to

the molar susceptibility and magnetic moment of the paramag-

netic substance in the inner tube. The magnetic moment is

related to the shift difference by the equation = a(Tav/C*)~
where a = 2.522x10- ( oe K-1 m-1 H-1 %, T = absolute

temperature, C*' = concentration of complex in mole/ml, and aJ
is the frequency difference in Hz of the two signals of the


This technique was used with both Ni(ketomacr)C12 4120 and
Ni(ketomacr)(NO3 2-6H20 complexes and the magnetic moments were
determined to be 3.88 and 3.74 BM respectively. These values

agree, to within experimental error, with those values obtained

by the Gouy method. It is concluded that the oaramagnetic

center which gives rise to the observed magnetic moment is the

same in the solid state and in aaueous solution. Based on the

information gathered from the conductance studies, one must

discount the Dresence of a five-coordinate species in solution

in which the fifth ligand is a coordinating anion. However.

interaction of solvent molecules cannot be discounted. This

interaction is considered to be finite because the moments which

are the same for both the hydrated solid solution states differ

from that of the anhydrous solid.

The room temperature magnetic moments for Ni(aldomacr)-

C12`4H20 and Ni(aldoketomacr)C12-4H20 were found to be 2.62 and
2.08 BM1 respectively. Both are considered to be too low for

simple high-sain nickel(II) (14,53). Nor are the data presented

in keeping with the expected sauare-Dlanar arrangement about the

nickel ion. The majority of square complexes of nickel involving

unsaturated ligands containing N-donors are diam~agentic (53). It

is difficult to rationalize these intermediate moments. Th~e

presence of inequivalent nickel ions, a folded ligand, or a

conformational equilibrium of the complex are possible contribut-

ors to the unusual magnetic moments. Further investigation is

needed; but it was considered to lie beyond the SCOoe Of this


The room temperature magnetic moments for Co(ketomacr)

212 4H20, Co(aldomnacr)C12'4H20, and Co(ketomracr)(NO3 2 H20
were observed to be 2.86, 2.51, and 4.73 BM1 respectively. Souare

planar cobalt(II) complexes are not numerous but have been found

to invariably give magnetic moments in the range of 2.1-2.8

BM (53). Tetrahedral complexes of cobalt(II) fall in the range

of 4.4-4.8 BM. The room temperature moments for Co(ketomacr)-

Cl2-4H20 and Co~aldomacr)C12-4H20 both lie in the range observed
for square planar cobalt(II). These moments are too large for

a low-spin d7 ion in an octahedral environment for which the

spin-free moment would be 1.73 BM. Thus, these two complexes

are considered to be of square planar geometry.

The square planar arrangement about the metal ion has been

demonstrated with the ketomacrocyclic complex of iron(II) and

inferred for cobalt(II) which were prepared by Goedken et al.

(25). The crystal structure presented for the iron(II) complex

illustrates that the macrocyclic ligand lies in a planar arrange-

ment around iron(II) ion as depicted in Fig. 5. However, their

complexes are approximately of octahedral geometry because of

the two axial ligands present and would not necessarily exhibit

the same properties as those complexes prepared in this studio.

Two types of geometry that could be considered are octahedral

and square planar, in which a tetragonal distortion is consider-

ed to be an extension of the square planar geometry.

Goedken etal. (25) report the preparation of a low-spin

cobalt(II) complex with the ketomacrocyclic ligand. This

octahedral complex has a reported magnetic moment of 1.83 BM.

Unfortunately, Goedken has not reported any additional charac-

terization of the cobalt complex. The magnetic data of the

cobalt complexes discussed herein are listed in Table 4 and

the Curie-Weiss plots for Co(ketomacr)C12^ H20 and Co(aldomacr)-

C12-4H20 are illustrated in Figures 20 and 21.
Spin-free octahedral cobalt(II) complexes customarily have

magnetic moments between 4.50 and 5.20 BM~ and values of 9

between -90 and 300 (41, 53). In octahedral complexes one does

not observe magnetic moments as low as the "spin-only" moment

of 3.89 BMI. High-spin octahedral cobalt(II) compounds reflect

an unusually large orbital contributions to the magnetic

moments. Considerable orbital contribution is seen in

Co(ketomacr)(NO3) 24H20 as evidenced by a magnetic moment of
4.73 BM at room temperature. Although this moment is on the

low end of the range expected for octahedral or distorted

octahedral cobalt(II) complexes (53),the complex is considered

to have actahedral geometry.

The magnetic properties of copper(II) complexes fall into

two broad classes (57). First, there are those having

essentially temperature-independent magnetic moments in the

range 1.75-2.20 BM. The complexes exhibiting such moments are

mononuclear complexes, i.e., having no major interaction

between unpaired electrons in different copper ions. The

second class (58), in which the moments are substantially below

the spin-only value and markedly temiperature-dependent, is

composed of complexes in which pairs of copper(II) ions are

held close together, usually by carboxylate anions.

The magnetic moments of the macracyclic copper(II) complexes

are listed in Table 4. The temperature-dependence of the mag-

netic susceptibility of Cu(ketomacr)(NO3 2 was found to follow

Curie-Weiss law and is illustrated in Fig. 22. The room temper-

ature moments of all the copper complexes tend to be on the

low side of the expected value. The moment reported for

Cu(aldomacr)C12-4H20 was obtained using the nmr techniques.
The use of this method was necessitated by the small quantity
of sample available which also precluded obtaining tempera-
ture-dependent data for the complex.

Because the moments for the copper complexes are less

than expected for spin-free copper(II) ion, it is tempting to
incorporate some sort of Cu-Cu interaction which is known:

(59,60) to reduce the observed moment. This Cu-Cu interaction

could be a direct interaction of the copper atoms within the

crystal or it could be indirect, i.e., through a bridging
atom of the complex such as the N atom. However, there are

no data which would lead to the conclusion that the complex
is binuclear. Although the Cu-Cu interaction is expected
to be more probable in the solid state than in solution, it

is the solution moment of Cu(aldomacr)Cl2-4H20 that is
smallest (peff = 1.19 BM). A clear-cut explanation is not
possible with the available data.

Electron Spin Resonance

For an electron of spin s = 4, the spin angular momentum

quantum number can have values of ms = +k, which, in the absence
of a magnetic field, leads to a doubly degenerate spin state.
When a magnetic field is applied, the degeneracy is resolved as
represented in Fig. 23. In an electron spin resonance experiment

a transition from the M~ = -t; to the MS = +k state occurs upon
absorption of a quantum of radiation. The spectroscopic

splitting factor, g, is inversely proportional to the field
strength at which the resonance is observed, i.e., E = g6HI,
where H is the field strength and B is the Bohr magneton. For

S'='' MS


Fig. 23a Splitting diagram for a single electron.

M ='+1

S'=1 S' S'



Fig. 23b Zero-field splitting diagram.

a free electron g has the value of 2.0023. In general, the

magnitude of g depends upon the orientation of the molecule

containing the unpaired electron with respect to the magnetic

field (61) and the spin and orbital angular mnomenta.

The electron spin resonance spectra were obtained for the

nickel complexes at liquid nitrogen temperatures (770K), The

esr data are listed in Table 10. The room temperature spectra

were attempted but no resonances were observed. The spectra of

Ni(ketomacr)C124H20 and Ni(ketomacr)(NO )2-6H20 both showed
a single, very broad and unsymmetric resonance anproximately

5000 gauss wide. Indeed, it is so broad that no p value can

be determined. The spectra of Ni(aldoketomacr)C12-4H20 n

Ni(aldomacr)C12 4H20 also gave extremely broad spectra but w~ith
a relatively narrow resonance superimposed upon the broad

absorption at g values of 2.00 and 1.99, respectively. In

hexaaquonickel(II), it is found experimentally that g = 2,25

(61). The difference from a g value of 2.00 is attributed to

spin-orbit coupling. The broad resonances are typical of most

nickel(II) complexes (62, 63).

The esr spectra of the cobalt complexes were obtained at

both room temperature and liquid nitrogen temperature. The g

values, listed in Table 10), fall in the range of 2.0 to 2.2

The spectrum of Co(ketomacr)(NO3 24H20 (the only high-spin

complex prepared) at liquid nitrogen temperatures contains an

eight-line hyperfine structure with coupling constants of about

78 gauss. Such hyperfine structure can be ascribed to the

interaction of unpaired electrons with the cobalt ion nucleus

(62, 64) which possesses a nuclear spin of 7/2.

Table 10. ESR data

Compound g(RT) LWI(gauss) g(770K() LWi(raus s)

Ni(aldomacr)C12-4H20 --- --- 1.996 30

Ni(aldoketomacr)C12 -4H20 --- --- 2 000 11.00

Ni (ke toma cr) C12 -4H -- ---

Ni(ketomacr)(NO3 2 6H20 -- ---

Co(aldomacr)C12-4H20 2.000 35 2.001 25

Co(ket omacr)C124H20 2.003 25 2.206 275
2.186 300

Co(ketom~acr)(NO3 24H20 2.204 375 2.221* 550

Cu(aldomacr)C12'4H20 2.101 260 2.099 160

Cu(aldoketomacr)C12-2H20 2.126 195 2.199 150

Cu(ketomacr)(NO )2 2.078 170 2.078 140

LW = line width
*center of gravity of 8 line hyperfine split.

Low-spin cobalt(II) complexes with planar, four N or tw~o N

and two O,coordinating ligands show esr spectra with gl values

of 2.0 and gl values of 2.2-2.9 (65). The spectrum of cobalt(II)
ion in the cubic field of a MgO host lattice has been analyz7ed

by Low (66) and by Bleaney and Hayes (67). The results showed

a g value of 4.2. No gl value near 4.0 was observed for any of
the macrocyclic complexes. The spectra, similar to those of

the nickel complexes, had narrow resonances superimposed upon

a very broad resonance; therefore, it was not possible to

ascertain the g value for the broader absorptions. Presumably

this broadening arose from a decrease in the spin-lattice re-

laxation time which accompanies either an increase in tempera-

ture or an increase in orbital contribution to the magnetic

susceptibility (68).

Although,based upon the esr spectra data, little insight

can be gained concerning the structure of the nickel and cobalt

complexes, the esr spectra of the copper complexes were more

ideal. Single resonances were observed at g values of approx-

imately 2.1 at room and liquid nitrogen temperatures. The

spectra exhibited level base lines. The g values listed in

Table 10 are typical of both octahedral and planar copper(II)

complexes (57, 69, 70).

It is noted that for the nickel complexes which exhibit

the large magnetic moments no esr signal was observed. For

those having the smaller magnetic moments an esr signal was

observed. Thus, there must be two different types of nickel in

these complexes.


Standard procedures were followed (2, 10, 19, 27) in

attempting to synthesize the uncoordinated ketomnacrocycle and

aldomacrocycle by direct reaction of Drescursors. The pro-

cedure presented in the experimental section is to be consider-

ed representative of a number of attempted syntheses. The

attempted syntheses of the ketomacrocycle differed according

to changes in the variables related to solvent, catalyst,

reactant ratios, and concentrations of the reactants.

The solvents used were: 95% ethanol, methanol, n-butanal,

acetone, and 2,2-dimethoxypropane was chosen to effect the

removal of H,O produced by the Schiff base condensation thus

enhancing the condensation.

Schiff base condensations are frequently acid catalyzed;

thus, it was expected that a few drops of a strong acid would

be needed to promote the condensation reaction. None the less,

one attempt was made to prepare the macrocyclic ligiand in the

absence of added acid to decrease the number of system components.

For each solvent system employed, the solutions of reactants

were colorless as, subsequently, were the combined solutions.

And no solid was produced. Because acyclic aldazines are known

to be yellow in color (71) and the expected oligomeric or

macrocyclic products sparingly soluble (59), it was concluded

that no desired condensation had occurred. But, if a few drops

of either concentrated sulfuric acid or concentrated hydrochloric

acid were added as catalyst to the reaction mixture of DAP, the

reaction mixture became vellow as the addition of the solution

of DAPH proceeded- upon refluxing for several hours, a yellow

solid began to form. The annearance of the yellow oroduct wans

the same whether sulfuric or hydrochloric acid was used as

evidenced by similar infrared spectra.

Other attempts to alter the condensation reaction included

changing of both the mole ratios of DAP to DAPH, by adding a

solution of DAPH to an excess of DAP, and the concentrations

of reactant solutions within the range of 10- to 10-6(t

prevent or, at least, to minimize the formation of 01igomers).

However, these variations proved to be ineffective as were a~l

of the previous attempts, i.e., each orecaration resulted in

the formation of the same lemon-yellow solid as evidenced by

their insolubility in numerous organic solvents, their rela-

tively high melting point (>3000C), and the elemental analysis

which are inconsistent with those expected for the macrocyclic


Attempts were made to separate ketomacrocycle from the

yellow~ solids described above. These attempts, motivated by

mass spectral data videe infra), included sublimation and an

in situ reaction of the oligomeric solid with nickel(II)

chloride. It was impossible to recrystallize any of the yellow~

solids formed since none was soluble in any of the solvents

used even though differences in solubility have been reported

for at least one macrocycle (TMCD) and an oligomeric side

product (59). Attempted sublimation of the various condensates

was unsuccessful, i.e., no solid was collected on the cold

finger of the sublimator although the temperature was slowly

raisd t 2S0C nd he pressure was reduced to 105 torr.

It would be reasonable to expect Ni(II) ion to preferentially

complex with the oligomeric condensate rather than the macrocycle.

As discussed previously in the Introduction, a step-wise reclace-

ment of coordinated solvent molecules on the N~i(II) ion is

Possible with non-cyclized ligands similar to the oligomeric

condensate; whereas, reaction of Ni(II) ion with the macrocvcle

would require that all of the solvent molecules of the Ni(TI)

ion be removed prior to coordination so that the Ni(II) ion

would "fit" into the central hole of the macrocycle. The sten-

wise process has a smaller entropy requirement associated with

it than does the latter process, thus making: energy requirements

smaller for the step-wise reaction. If the condensate were more

soluble than the macrocycle, the step-wise complexation

reaction could occur without any competition from the macro-

cycle. On the other hand, if the macrocycle were the more

soluble, the activational energy requirements might be too

large for the Ni(II) ion to react directly with the macrocycle

regardless of solubility. Thus, this technique might also be

ineffective in removing or separating the free-ketomacrocycle.

The attempt to remove any free ketomacrocycle from the

yellow condensate by reacting the yellow product inf situ with

a refluxing ethanolic solution of nickel(II) chloride also

proved to be ineffective, as evidenced by the lack of any color

change of the mixture. Experience has shownm that formation of

Ni(ketomacr)C12 H20 is associated with a color change of

yellow-green to a dark green or brown-green color. The yellow

condensate was still suspended in the yellow-green solution

of nickel(II) chloride and ethanol. It is concluded that

neither the oligomreric condensate nor the Letomacroevele

completed with the Ni(II) ion,

To provide direction in the preparation and isolation of

the free macrocycle the yellow Droducts produced were character

ized by mass spectra, infrared soectra, and elemental analysis.

Mass spectral data for the product isolated in the

attempted synthesis of the ketomacrocycle, obtained under

relatively extreme conditions, are contained in Table A-6 of

the Appendix. It was difficult to obtain a good mass snectrum;

the probe temperature had to be set at its maximum of 3500C to

achieve reasonable intensity, accesting to the stability and

nonvolatile nature of this product.

The peak corresponding to the greatest mass was weak

(m/e equal 624; relative intensity of 1.5%); its mass was ca.

twice the mass of the desired macrocycle. This peak could

correspond to a condensate consisting of four units of DAP to

three units of hydrazine. A weak peak corresponding to the

parent ion was observed at m/e 318, Although the presence of

this peak would normally lead one to exnect that some of the

free ketomacrocycle was present in the sample, little credence

was placed in this peak because the mass spectrum had the

general appearance of a ''picket fence and about any neak m/e

< 479 could be found. Peaks of lesser m/e could correspondl

to fragments of the particle that gives rise to the peak

m/e 624. A systematic cracking pattern can be followed by

beginning at the larger values of m/e and proceeding to the

lesser values of m/e. This cracking pattern can be achieved

by subtracting appropriate atoms and/or groups to form the

next fragment.

The infrared spectra of the condensate, DAP, and DAPH are

presented in Fig. 24. In general the spectrum of the conden-

sate had the appearance of a composite of the soectra of DAP

and DAPH with the major exceptions of the C=0 and N-H vibra-

tions. The spectrum contained characteristic absorptions at

1700 and 1610 cm-1 which were assigned to a carbonyl stretching

vibration and an acyclic C=N stretching vibration, respectively

(44, 47). The presence of the carbonyl band was not unexpected

because other evidence indicated that not all of the product

was macrocycle. The relative intensity of the C=0 absorption

is less than that observed in the spectrum of DAP which is

consistent with a terminal carbonyl. The high energy band

associated with the ring vibrations (Band I-ring) is absent in

the spectra DAP and DAPH but can be found in the spectra of PDC

and PDCH at 1580 em- Because of the numerous bands between

1580 and 1200 em-1 there was no attempt to make these assign-


The soectrum of the condensate lacked the absorptions at

3300 and 3150 cm- observed in the spectrum of DAPH and assian-

ed to the antisymmetric and symmetric stretching vibrations

of the NH2 group. Nor was the NH2 deformation mode, observed

at 1640 cm1 in the spectrum of DAPH present in the spectrum

Frequency (cm l)

Fig. 24 Infrared spectra of"ketomacr"and its precursors .

of the condensate. The absorotions present at 1560, 1450, and

1425 cm-1 generally can be assigned to ring vibrations. The

assignment of absorptions are presented in Table A-4 of the


It is concluded on the basis of the infrared spectral data

that the yellow solid is an acyclic product, terminated on at

least one end with a carbonyl function. The presence of any

unreacted DAP in the product is precluded by the good solubility

of DAP in the solvents that were used to wash the product.

Therefore, any similarities in the spectrum of the condensate to

that of DAP must be assigned to functional groups bound to the

product, a conclusion consistent with the mass spectra data.

The sum of the weight percentages of C, H, and N from the

elemental analysis is 98%. The 2% difference is attributed to

the presence of oxygen arising either from the oxygen in water

or the oxygen of a carbonyl group. Based upon the infrared

spectrum, it is proposed that the oxygen is a carbonyl oxygen.

Also, since the product was dried in vacuo over P 010 for
24 hours at 1000C Drior to analysis and no gain in weight was

observed when the solid was allowed to stand overnight in

atmospheric moisture, it is unlikely that any water was present

prior to the analysis. The CHN ratio approximates a moiety

containing a linear arrangement of two units of macrocycle

terminated on one end by a carbonyl, i.e., C36H36 ~120, (Cale:d:

C, 66.24; H, 5.56; N, 25.75. Found: C, 66.56; H, 5.91;

N, 25.52). Thus it is concluded, on the basis of mass spectral

data, infrared data, and elemental analysis, that oligomeriza-

tion occurred instead of the desired cyclization.

Because all of the attempts to prepare the ketomacrocycle

were unsuccessful, alternate methods were considered. The

most logical alternative was to separate the free base fromn a

complex of the ketomacrocyclic ligand videe infra). The complex

Ni(ketomacr)C124H20 was chosen as the reactant species because
the complex is the most easily prepared of all the other

complexes and is reasonably soluble in polar solvents. Methods

which might be expected to remove the Ni(II) ion from the

macrocycle would include: electrochemical reduction of the

metal, chemical reduction of the metal, precipitation of Ni(II)

ion with sulfide ion or complexation of Ni(II) ion with cyanide

ion (16).

Before alternate methods of preparation of the free

ketomacrocycle were considered, it w~as necessary to settle the

question: "Why bother with preparing free ligand when the

complexes are easier to prepare?"' The evidence for the prepara-

tion of the macrocyclic complexes is good videe supra), but more

information might be gained about metal-donor interactions.

for example, if the free ligand were available for comparison

studies. It would also be informative to be able to prepare

the complexes from two different synthetic routes and compare

the products. The ligand would, of course, be fully character-

ized and possibly reveal properties suggesting further study.

To assess the optimum conditions for electrochemical

reduction, a polarogram of Ni(ketomacr)C12-4H20 a ae.I
was anticipated that electrochemical reduction would oreferen-

tially reduce the metal within the complex allowing the metal

to "pop" out of the central h-ole of marcocycle. Indeed, it

was also anticipated that the polarogram might give some in-

formation about the coordinated macrocycle itself. The

Sargent Model XVI Polarograph employed had the capability to

mearure current changes within the voltage range of + 3 volts

using a saturated calomel electrode as the reference electrode.

After obtaining a half-wave potential, E of -12.5 volts of

a standard 0.01 M solution of Ni(II) ion [1it. (72) E =-1.1

volts] in 1.0 M KCl solution, the polarogram of the complex

was attempted. Various voltage settings, concentrations of

complex (102 to 105 M), and type of supporting electrolyte

were tried. However, no polarographic wave of any kind, except

that of the supporting electrolyte, could be observed for the

several solutions of the complex. It was concluded that the

complex was too stable, i.e., the voltage range of the instru-

ment was not sufficiently large. It is known (73) that the

half-wave potential for the reduction of a metal complex is

generally more negative than that for the corresponding simple

metal ion. Thus, it would be anticipated that a more negative

potential would be required to reduce the metal from the com-

plex than could be achieved by the polarograph.

A mercury cathode cell described in the Experimental

section was used in the subsequent attempt to reduce the compl exed

Ni(II) ion under more stringent conditions using electrolysis.

This particular cell was chosen because of two factors (42).

Many metals depositing on mercury can form an alloy (analgami)

with the mercury. The deposition potentials of these metals

on the mercury are now displaced from their value in the

positive direction with respect to reduction potentials. The

deposition is also aided by the fact that the hydrogen over-

voltage on mercury is particularly large.

A saturated solution, 10- M, containing suspended solid of

Ni(ketomacr)C12-4H20 together with a supporting electrolyte to
0.1 M sodium perchlorate was placed in the cell. A potential

was applied to the cell but no reaction was observed until the

potential was increased to 3.0 volts at which point gas

evolution was observed at both the platinum and mercury elect-

rodes. Increasing the potential to 20.0 volts only served to

increase, linearly, the current from 0.00 to 1.00 amp. Most

of the suspended solid rose to the top of the cell because of

the stirring action of the bubbles.

Even though it was concluded that no reduction had taken

place, a control experiment was performed to provide additional

basis for this conclusion. Accordingly, a solution of 0.1 M1

sodium perchlorate, the supporting electrolyte, was electrolyzed

in an identical manner. No current flow was observed until the

potential had been increased to 3.0 volts; at 5.0 volts both

solutions, the control and complex, had a corresponding

current of 0.15 amp passing through them. A solution contain-

ing 0.02 M Ni(II) ion in 0.1 M NaC104 deposited metallic nickel
on the mercury electrode at 3.0 volts (current 0.05 amp) as

evidenced by a blackish coloration of the mercury pool. Gas

evolution was noted at 3.0 volts in all three cases. Had there

been any reductive processes occurring at the mercury cathode

before or after the reduction of water began, a non-linear

increase in current would have been observed. Since no such

increase was observed in comparing the electrolysis of the

supporting electrolyte to that of the electrolysis of the

complex, it is concluded that no reduction of Ni(II) ion out

of the complex was achieved and that only water was electro-

lyzed to H2 and 02.
Chemical reduction of the metal from the macrocycle was

attempted using sodium-potassium alloy. The sodium-potassium

alloy has been used as a rich source of electrons for reactions

requiring a powerful reducing agent (74), e.g., reduction of

bis(trimethylamine)-boronium iodide in 1,2-dimethoxyethane to

produce 1,1,3,3-tetramethyl-1,3-diazonia-2,4-diborlccoet

ane (B2 2C5H18). The utility of this alloy is enhanced because

it is a liquid at room temperature when prepared as a mixture

of 25% sodium and 75% potassium by weight. The use of the alloy

as a source of electrons is known (75); but, the alloy is not

widely used because of its high reactivity toward any easily

reduced material. It was expected that solvated electrons

might reduce the complexed Ni(II) ion of the macrocycle and,

thereby, promote its removal from the complex. Alternatively,

the ligand could be reduced preferentially allowing the reduced

form of the macrocycle to be more easily separated from the

metal ion. The application of the alloy for such a purpose as

this was considered to be both novel and reasonable.

The sodium-potassium alloy, when added to and stirred

vigorously with monoglyme under N2, produces a system of
solvated electrons as evidenced by the blue color of the

solution; however, upon cessation of stirring, the blue color

is lost within a few seconds. Thus, while a solution is

stirred there is good contact of "free" electrons with any

material that would be present in the solvent.

In the application of this technique, a one-half gram

sample of the sodium-potassium alloy (equivalent to 15 mmoles

of electrons) was added under N'2 to a 25 ml yellow-green

monoglyme suspension containing 0.5 g (1.0 mmole)

Ni(ketamacr).4H20. Vigorous stirring with a glass-encased

magnetic stir-bar caused the suspension to become blue-green

in color. No change in color was noted after seven hours of

stirring; the mercury-like alloy was still present. The

reaction was decanted off the alloy leaving a yellow-green

suspension of the macrocyclic complex in monoglyme. The yellow-

green suspension was filtered under N2. The clear, colorless

filtrate was evaporated under a stream of N2; but, no residue

formed. The green solid collected on the filter was washed

several times with diethyl ether and the combined washings

evaporated; no non-volatile product remained. Because neither

nickel(II) chloride (anhydrous or hydrated) nor the oligomeric

condensate are soluble in monoglyme or ether, the separation

of the free ligand was dependent upon its solubility in those

solvents. The ligand is expected to be yellow in color and

since no increase in yellow color was noted in the suspension

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs