Preparation and deprotonation of certain pi-arene-pi-cyclopentadienyliron(11) salts /

preparationdepro00hend ( .pdf )

PREPARATION AND DEPROTONATION OF CERTAIN
PI-ARENE-PI-CYCLOPENTADIENYLIRON(II) SALTS













By

WILLIAM A. HENDRICKSON


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY







UNIVERSITY OF FLORIDA















ACKNOWLEDGMENTS


I would like to thank Dr. John F. Helling, my research chairman,

whose guidance has made this dissertation possible. I would also like

to recognize the great support of Dr. Jackie Dugan who provided me with

excellent mass spectra of my compounds.

Finally, I wish to thank Lee, my wife, for typing this dissertation.








































i i















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS . . . . . . . . ... ....... i

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

ABSTRACT . . . . . . . . . . . . . . . vi

INTRODUCTION . . . . . . . . ... . . . . . 1

PREPARATION OF n6-ARENE-rn-CYCLOPENTADIENYLIRON(II) SALTS ... 15

DEPROTONATION OF n6-ARENE-n5-CYCLOPENTADIENYLIRON(II) SALTS . . 28

EXPERIMENTAL . . . . . . . . ... ....... . 59

r6-Anisole-n5-cyclopentadienyliron(ll) Hexafluorophosphate .. 60

n6-Chlorobenzene-n-cyclopentadentadienyron(I) Hexafluoro-
phosphate . . . . . . . . . . . . . 61

T6-Diphenylmethane-5-cyclopentadentadienyliron() Hexa-
fluorophosphate . . . . . . . . . . . . 62

n6-Fluorene-n5-cyclopentadennyliron(ll) Hexafluorophosphate 63

n6-Hexamethylbenzene-n5-cyclopentadienyliron( I) Hexa-
fluorophosphate . . . . . . . . . . . . 64

n6-Toluene-ns-cyclopentadienyliron(Il) Hexafluorophosphate 64

n6-Aniline-nS-cyclopentadienyliron(Il) Hexafluorophosphate .. 65

n6-Carbazole-n5-cyclopentadienyliron(lI) Hexafluorophosphate .66

r6-Diphenylamine-n5-cyclopentadienyliron(II) Hexafluoro-
phosphate . . . . . . . . . . . . . 67

n6-1,3,5-Triisoopropylbenzee-n-cyclopentadienyliron(II)
Hexafluorophosphate . . . . . . . . . . .. 69

n6-Thiophenol-n5-cyclopentadienyliron(li) Hexafluorophosphate 70

n6-Phenol-qn-cyclopentadienyliron(ll) Hexafluorophosphate . 71


_~~








TABLE OF CONTENTS (continued)


Page

r6-N,N-dimethylaniline-n5-cyclopentadienyliron(lI) Hexa-
fluorophosphate . . . . . . . . . . . 72

n6-Thioanisole-n5-cyclopentadienyliron(ll) Iodide ...... 74

n6-Anisole-n5-cyclopentadienyliron(II) Iodide . . . ... 75

r6-Carbazolyl-n5-cyclopentadieny]iron(ll) . . . .... 75

n6-Fluorenyl-Tn-cyclopentadienyliron(lI) . . . . .. 76

(1-5-r5-6-Diphenylmethylenecyclohexadienyl)-ns-cyclo-
pentadienyliron(ll) ... . . . . . . . . 77

(1-5-ns-6-Phenyliminocyclohexadienyl)-n5-cyclopenta-
dienyliron(ll) . . . . . . . . . . . 78

(1-5-ns-Cyclohexadienyl-6-one)-n5-cyclopentadienyliron(ll) . 78

(1-5-n5-Cyclohexadienyl-6-thione)-ns-cyclopentadienyliron(l) 79

(1-5-n5-6-Iminocyclohexadienyl)-n5-cyclopentadienyliron(lI) . 81

a. Reaction in Dichloromethane . . . . . .... 81
b. Reaction in Ammonia . . . . . . .... .. .82
c. Reaction in Tetrahydrofuran . . . . . . .. 82
d. Reaction in Acetonitrile . . . . . . . . 82

Reaction of n6-Diphenylmethane-n5-cyclopentadieny1iron(ll)
Hexafluorophosphate with Sodium Amide . . . . . . 83

Regeneration of n6-Diphenylmethane-n5-cyclopentadienyliron(II)
Cation . . . . . . . . . . . . . 83

Attempted Reaction of Methyl Iodide with Certain r-Arene-
i-cyclopentadienyliron(II) Salts . . . . . . . 84

a. r6-Aniline-n5-cyclopentadienyliron(II) Hexafluoro-
phosphate . . . . . . . . . . . 84
b. r6-Phenol-n5-cyclopentadienyliron(II) Hexafluoro-
phosphate . . . . . . . . . . . 84
c. n6-Thiophenol-n5-cyclopenaendienyliron(I) Hexa-
fluorophosphate . . . . . . . . . 84

Preparation of n6-Anilinium-n5-cyclopentadienyliron(II)
Cation . . . . . . . . . . . . . 84

a. From n6-Aniline-ns-cyclopentadienyliron(l I) Hexa-


fluorophosphate . . . . . . . . . . 84









TABLE OF CONTENTS (continued)


Page

b. From (1-5-Tn-6-iminocyclohexadienyl)-n5-cyclopenta-
dienyliron(II) . ... .. . . . . . . . 85

Preparation of n6-Benzene-n5-cyclopentadienyliron(ll)
Hexafluorophosphate in Ethylpyridinium Bromide-Aluminum
Chloride Eutectic . . . . . . . .. . .85

Reaction of Hydroiodic Acid with n6-Anisole-n5-cyclo-
pentadienyliron(ll) Hexafluorophosphate . . . .... 86

Attempted Synthesis of n6-Phenol-rn-cyclopentadienyliron( I)
Hexafluorophosphate . . . . . . . . ... . 86

Attempted Synthesis of rn-Aniline-n5-cyclopentadienyliron( I)
Hexafluorophosphate Using Anilinium Sulfate . . . ... 87

Sodium Hexamethyldisilylamide . . . . . .... .. . 88

Sodium Hydrogen Sulfide . . . . . . . . . 88

REFERENCES . . . ... . . . . . . . . . 89

BIOGRAPHICAL SKETCH . . . . . . . .... ... . 92















LIST OF TABLES


Table Page

I. n6-Arene-n5-cyclopentadienyliron(ll) Salts . . . ... 17

II. 'H NMR Spectral Data for n6-Arene-n5-cyclo-
pentadienyliron(ll) Salts . . . . . . . . .19

III. Deprotonated 16-Arene-ns-cyclopentadienyliron(l I)
Complexes . . . . . . . . . . . . 3

IV. 'H NMR Spectral Data for Deprotonated n6-Arene-
s5-cyclopentadienyliron(II) Complexes ..... .. . . 40









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


PREPARATION AND DEPROTONATION OF CERTAIN
PI-ARENE-PI-CYCLOPENTADIENYLIRON(II) SALTS

By

William A. Hendrickson

August 1978

Chairman: Dr. John F. Helling
Major Department: Chemistry

A number of known i-arene-i-cyclopentadienyliron(I) salts have

been prepared including the t-complexes of benzene, toluene, mesitylene,

diphenylmethane, fluorene, hexamethylbenzene, anisole, aniline, and

chlorobenzene by the direct reaction of the appropriate arene, ferrocene,

aluminum chloride, and aluminum at elevated temperatures. Also, new w-

complexes of carbazole, diphenylamine, triphenylmethane, and 1,3,5-tri-

isopropylbenzene were prepared directly from ferrocene and the respec-

tive arene as above. Finally, the n-phenol and i-thiophenol complexes

were prepared by the reaction of n6-chlorobenzene-n5-cyclopentadienyl-

iron(ll) hexafluorophosphate with potassium hydroxide or sodium hydrogen

sulfide respectively.

The i-complexes of fluorene, triphenylmethane, diphenylamine, car-

bazole, phenol, thiophenol, and aniline were deprotonated with appro-

priate bases including sodium amide, sodium hexamethyldisilylamide,

potassium hydroxide, alumina, or the sodium salt of acetonitrile. The

T-fluorenyl and i-carbazolyl complexes seem to possess a greater degree

of dipolar character than the others. The remaining complexes re-

sulting from deprotonation have properties consistent with a structure

containing an exocyclic double-bonded cyclohexadienyl unit.



v1l















INTRODUCTION


With the increased interest in organometallic chemistry following

the discovery of ferrocene slightly more than twenty-five years ago, the
1 ,2
field has developed into a major research area.2 This rapid develop-

ment was, in part, promoted by the unique aspects of the field. The

interface of inorganic and organic chemistry has led to a number of ex-

citing reactions and compounds of unique structure and bonding. In an

organometallic complex the interaction of the metal and the organic lig-

and changes the properties of both the ligand and the metal. The quint-

essence, then, of organometallic chemistry is the way in which this in-

teraction manifests itself in the reactions and structure of a particu-

lar compound.

The field of catalytic chemistry is, to a large extent, dependent

on the favorable interaction between metal atoms or ions and organic

molecules. The formation of either a transient or a stable organo-

metallic species allows the organic ligand to undergo reactions that

are not normally possible under similar conditions in the absence of a

metal atom. Synthetic organic chemistry is also heavily dependent upon
4
metal-organic compounds. Whether it is the formation of alcohols from

ketones via a Grignard reagent or the preparation of aldehydes via the

Wacker process, metal atoms have become deeply entwined with the field

of synthetic organic chemistry. Finally, formation of metal to carbon

bonds has allowed the isolation and characterization of organic species

that are normally considered unstable or transient intermediates. A


1









classical example of this is the formation of stable metal complexes

with cyclobutadiene such as n4-tetramethylcyclobutadienenickel(II)

chloride dimer. Cyclobutadiene and its derivatives are so reactive

that their isolation is not normally possible except by using matrix
6
isolation techniques. A more recent example of this is the isolation
7
of a metal-coordinated allyl system containing a silicon atom. Com-

pounds with silicon-carbon double bonds or even partial double bonds

are too reactive to isolate and this represents the first case of this

structure being successfully isolated.

While the field of organometallic chemistry may be subdivided ac-

cording to reaction or structural type, it has also been classified ac-

cording to ligand type such as complexes of carbon monoxide, alkenes,

dienes, alkyls, and arenes. Metal complexes of aromatic ligands or lig-

ands that are considered aromatic upon complexation are known where the

ligand is bonded to the metal via a ligand-to-metal sigma bond or

through interaction of i orbitals of the ligand with metal orbitals.

The ligands that have been used in these T complexes include cyclobuta-
5 1 8 9
diene, cyclopentadienide, arenes, cycloheptatrienyl cation, and cy-
10
clooctatetraene (Figure I). A particularly interesting group of 7-

coordinated aromatic metal complexes are those classified as sandwich

compounds. In general, the type of compound that is considered a sand-

.wich compound is one in which a metal atom is "sandwiched" between two

parallel aromatic ligands. The first sandwich compound, bis(ri5-cyclo-

pentadienyl)iron(II) (Figure lI),commonly called ferrocene, was prepared

in 1951 by Kealy and Pauson who described it as a sigma bonded complex,

analogous to a Grignard reagent. In 1953 the structure was correctly
11
elucidated by Wilkinson et al. as the sandwich structure that is now

accepted.











Ni

/\
Ni


Fe


.Mo+ BF4
Oc" \c
C0
0 0
0


FIGURE I


FeCl2 + 2 NaC5H5


CrCI3


- PhMgBr


L-200C, THF

2. 20-400C
3. H20, H3P02


FIGURE II


C C
0 0


U


Fe





Cr









Following the discovery and characterization of ferrocene,Zeiss and
12
Herwig showed that the previously reported reaction of phenylmagnesium

bromide with chromium(Ill) chloride did not give sigma complexes, but

rather bis(n6-arene)chromium(O) compounds (Figure II). From this mod-

est beginning, the number of compounds and the scope of the field have

expanded tremendously. The sheer number of compounds runs into the

thousands and includes most transition metals.

In 1957 Coffield, Sandel, and Closson3 reported the first prepara-

tion of a transition metal sandwich compound containing both a cyclo-

pentadienyl and an arene ligand (Figure III). By reacting n5-cyclo-

pentadienyliron dicarbonyl chloride or bis(ns-cyclopentadienyl)diiron

tetracarbonyl with aluminum chloride in refluxing mesitylene, followed

by hydrolysis in the workup, they were able to isolate Tb-mesitylene-

q5-cyclopentadienyliron(II) iodide. Following this initial preparation,
13 14
Coffield et al. and others prepared approximately 15 compounds with

similar structures. In 1963 Nesmeyanov et al. reported that it was

possible to prepare the n6-benzene-ns-cyclopentadienyliron(lI) tetra-

phenylborate by the reaction of ferrocene and aluminum chloride in the

presence of aluminum powder and refluxing benzene. A mechanism for

this reaction reported in 1976 by Astruc and Dabard proposed the initial
16
step to be the complexation of aluminum chloride and ferrocene. The

aluminum chloride can complex at the cyclopentadienyl ring which leads

to product or the aluminum chloride can complex at the iron atom to

give an insoluble complex which does not react further (Figure IV).

Since the first paper by Nesmeyanov et al., reporting this reaction in

1963, this procedure has been the preferred synthetic route with well

over 100 new r-arene-r-cyclopentadienyliron(II) salts reported.











Cpe(CO)2C +4- AICl3 --


CpgFe2(CO)4 + AICI3 +


2. H0
3. Nal






2. H20
3. Nal


FIGURE III


Fe -AICI
43


O_ Hc
S2.H
Fe + AICI PhH A ->
43. NH4BF4


AIC13'


AICI3

13


Fe SBF4


2. H2p, NH4BF4


FIGURE IV




6




While the reaction of ferrocene and aluminum chloride has great utility

in the preparation of these mixed sandwich compounds, the use of alumin-

um chloride does put some restraints on the type of arene that can be

reacted successfully. Since aluminum chloride is a very strong Lewis

acid, the coordination of aluminum chloride with basic functional groups

on substituted arenes has interfered with the desired reaction.
18,19
Nesmeyanov et al. reported that no w-arene-a-cyclopentadienyliron(II)

complexes were obtained when aniline, phenol, thiophene, or benzophenone

were reacted with ferrocene and aluminum chloride (Figure V). Subse-
19
quently, the aniline complex was prepared by an interesting route

that will be discussed later while the others are yet to be reported in

the literature. Chlorobenzene and bromobenzene also have the ability

to act as Lewis bases toward aluminum chloride as they both have the re-

quisite unshared pair of electrons. However, their reaction with alum-

inum chloride is less a concern than their reaction with the aluminum
20
powder in the reaction mixture to form benzene (Figure V). Over 50%

of the product in some reactions of bromobenzene and ferrocene is the

n6-benzene-q5-cyclopentadienyliron(II) salt. This is especially unfor-

tunate, not only with respect to the yield, but because the separation

of the resulting products is almost impossible.

Even a cursory inspection of the structure of these r-arene-t-

..cyclopentadienyliron(ll) complexes should provide some insight into the

types of reactions that should be possible. An analogy might be drawn

to reactions of 2,4-dinitrochlorobenzene and anisole. Since anisole

is an electron-rich system,electrophilic substitution on the ring should

be facilitated. Conversely, 2,4-dinitrochlorobenzene is an electron-

poor system and the same reaction would be unlikely; nucleophilic sub-











ANILINE



PHENOL


THIOPHENE

Fe + AICI1 + Al

BENZOPHENONE


CHLOROBENZENE


BROMOBENZENE



L


-CI
Fe





Fe +


NO REACTION



NO REACTION



NO REACTION







NO REACTION


+ Fe+


+ Fe+


FIGURE V










stitution should be favored. Likewise, the r-arene-i-cyclopentadienyl-

iron(ll) cation should favor reactions with electron-rich species.

Coupling this with a transition metal atom's known ability to stabilize

normally unstable species leads one to propose the following four ligand

reactions as viable possibilities: nucleophilic addition, nucleophilic

substitution, ring metalation, and removal of an acidic alpha proton

using a strong base (Figure VI). Although this discussion is specifi-

cally about iron sandwich complexes, such reactions have also been de-

scribed for complexes of other metals and ligands.

The first of these reactions, nucleophilic addition, was discovered

shortly after the initial preparation of these compounds. Green, Pratt,
14
and Wilkinson reported that it was possible to react sodium borohydride

with various r-arene-T-cyclopentadienyliron(II) salts. The resulting

product is viewed as a Meisenheimer intermediate stabilized by coordina-

tion to the iron atom (Figure Via). While the use of hydride as a nu-

cleophile does not allow distinction between exo or endo attack, the
14 21 22
use of sodium borodeuteride, phenyl lithium, and sodium cyanide

showed that attack is always exo. It should be noted that oxygen or

nitrogen nucleophiles did not react with these *-arene-r-cyclopentadi-
22
enyliron(ll) cations as was also the case with the uncoordinated

arenes.

The second reaction, nucleophilic substitution, is quite analogous

to nucleophilic addition with respect to the nature of the first step,

addition of the nucleophile, and with respect to the structure of the

resulting intermediate. However, for nucleophilic substitution to oc-

cur, a good leaving group such as bromide or chloride should be present

which allows the formation of a coordinated arene by expulsion of the












+ Nu-


A. NUCLEOPHILIC ADDITION


-Cl
Fe + + Nu- )



B. NUCLEOPHILIC SUBSTITUTION


FeT


+ MB >


C. PROPOSED RING






F CH3
Fe* + B


Fe+


Nu= RO-,RS-
RHN-,
NH-


S +
Fe+ + HB


METALATION






SHB
S Fe+ + HB+


D. DEPROTONATION AT .THE ALPHA POSITION


FIGURE VI


Nu



) Fe


Nu= H-, D-,
R", CN-


Fe








19
chloride or bromide ion. Nesmeyanov et al. first reported the pre-

paration of the n6-chlorobenzene-n5-cyclopentadienyliron(ll) cation and

its reactions with nucleophiles. Under conditions that were remarkably

mild Nesmeyanov et al. were able to react n6-chlorobenzene-n5-cyclo-

pentadienyliron(ll) salts with methoxide, phenoxide, thiophenoxide,

amide, alkyl amide and phthalimide anions to obtain the substituted

products in 50-90% yield (Figure VIb). The reactions were typically

performed at 500 C for 1-2 hours in ethanol. The free chlorobenzene

was usually unreactive under these conditions. Interestingly, it was

reported that it was not possible to prepare the r-phenol complex via
19
this route.

It should be noted that the removal of an arene proton has not

been reported in sandwich complexes where iron is the metal. However,
23
it has been shown that metalation of bis(n6-benzene)chromium(O)2 is

possible.

Only in the past five years has experimental evidence of increased
22
acidity for alpha protons of the iron-coordinated arenes been available.2

Yet, previous to 1973, there was strong evidence to suggest that the

idea was not without merit. During the course of Nesmeyanov's initial

studies of the T-arene-T-cyclopentadienyliron(lI) salts it was discov-

ered that these mixed ligand sandwich compounds are extremely stable
24
toward oxidation and reduction at the iron atom. This is rather sur-

prising as the analogous ferrocene compounds are easily oxidized at the

metal. Furthermore, if the n-arene-i-cyclopentadienyliron(Il) salt was

substituted with alkyl substituents on either ring, attempted oxidations

oxidized the alkyl substituents to the corresponding carboxylic acids

(Figure VII). This reaction is interesting in light of the differences








~CH3
Fe




Fe+
4=- CH3


- KMn04


+ KMnO4 -


<-COOH
Fe*




ZFe+
4=-COOH


- KMn04 ---


FIGURE VII


LiNH2 or LiOMe -


I + 11 -


FIGURE VIII


Fe


Fe+









observed with respect to ferrocene, but more pertinent to this discus-

sion was the increase in acidity of the coordinated acids with respect
24
to the uncoordinated acids.

The first evidence that arene alpha protons in iron systems might

have enhanced acidity with respect to the uncoordinated arene alpha
22
protons was presented in 1973 by Helling and Cash. In an attempt to

extend the variety of nucleophiles that could react with the coordinated

arenes,they reacted methoxide and amide anions with bis(n6-mesitylene)-

iron(ll) hexafluorophosphate. The expected reaction was the formation

of a bis(n5-cyclohexadienyl)iron(ll) complex obtained by nucleophilic

addition of either methoxide or amide ion. Instead, regardless of the

nucleophile, the structure shown in Figure VIII was obtained. To ex-

plain the formation of this product it was necessary to consider the

possibility that the nucleophiles were acting as bases to initially re-

move an alpha proton from one coordinated mesitylene. This produces a

new nucleophile which adds readily to another unreacted bis(n6-mesityl-

ene)iron(ll) cation. The compound then partially decomposes to give

the product obtained. The key step in this reaction is the removal of

an alpha proton by a strong base. As a check on the possibility that

the iron was substantially affecting the acidity of the alpha protons,

a reaction was carried out in which bis(n6-hexamethylbenzene)iron(Il)

chloride was dissolved in D 0 with a catalytic amount of trisethylene-

diamine. After 24 hours at room temperature 94% of the alpha protons

had exchanged with deuterium. A similar experiment with uncomplexed

hexamethylbenzene showed no deuterium exchange (Figure IX). The

greatly increased acidity of the alpha protons could be attributable

to two effects promoted by the iron atom. The first is a simple induc-












,= Me) 6
Fe4 +

'RMe) 6


D20, C6H12N2
24 HOURS


CD3)6
Fe ++

< (CD3)6


020, C-H,2N2
D2, C6H 112 NO REACTION
24 HOURS


FIGURE IX


< 13BASE
Fe


Fe* --- Fe
I I


FIGURE X


Me)6









tive effect as seen with the carboxylic acids. Also, and perhaps more

importantly, there exists the possibility of resonance interaction of

the anion with the coordinated ring to form an exocyclic double bond

and cyclohexadienyl structure similar to the type seen with nucleo-

philic addition (Figure X). A number of attempts were made to isolate

these exocyclic double-bonded compounds. However, their reactivity and

instability precluded this possibility.

Due to the greater stability of r-arene-r-cyclopentadienyliron(II)

salts with respect to the bis(n6-arene)iron(ll) salts, the possibility

of isolating a stable, exocyclic double-bonded, cyclohexadienyl complex

was pursued using these as reagents. This dissertation will deal with

the preparation and characterization of these complexes.















PREPARATION OF n6-ARENE-n5-CYCLOPENTADIENYLIRON( I) SALTS


Since the initial preparation of n6-mesitylene-n5-cyclopentadienyl-
13
iron(ll) iodide by Coffield, Sandel, and Closson and the subsequent
15
preparation of similar salts by Nesmeyanov, Vol'kenau, and Bolesova

two general procedures have been used. The most widely used is the
25
Fischer-Hafner synthesis with ferrocene as the iron-containing start-

ing material. The second method involves the conversion of an existing

n6-arene-n5-cyclopentadienyliron(II) salt to a new i-arene salt by one

of the methods described in the introduction, such as nucleophilic sub-

stitution or oxidation of alkyl groups on either of the rings. Two

methods were used in this study to obtain good yields of these i-arene-

i-cyclopentadienyliron(lI) salts. A number of other methods were at-

tempted and these will be briefly described.
15
The Fischer-Hafner synthesis, as modified by Nesmeyanov et al.

was the method used to prepare a wide variety of alkylated and substi-
26 27 27
tuted i-coordinated arenes. Toluene, diphenylmethane, fluorene,
28 19
hexamethylbenzene, anisole, and 1,3,5-triisopropylbenzene were re-

acted with ferrocene and aluminum chloride to obtain the respective ir-

arene-i-cyclopentadienyliron(ll) cations (Figure XI). Of the preceding

cations only the last, the r-1,3,5-triisopropylbenzene complex, is a

new cation. Excluding yields, the methods of preparation and charac-

terization data obtained are generally consistent with the reported

values (Table I). The sole important exception to this is the melting

point of n6-toluene-n5-cyclopentadienyliron(lI) hexafluorophosphate


15












H3 H
Fe+ PFg






-OMMe
Fe+ PFl





(IMe)g6

F*+ PFg


K< HMPh
Fe* PFg







Fe* PF-
6







Fe PF-
6


FIGURE XI


+ PhH + Al


I. AlCl3- EtPyr Br-

2.H20, NH4PF6


Fe PFg-


FIGURE XII


Fe









TABLE I

q6-Arene-rS5-cyclopentadienyliron( I) Salts


YIELD (%) DECOMPOSITION POINT
In this work Reported in lit.


a
ARENE


Aniline

Anisole
b
Anisole

Carbazole

Chlorobenzene

Diphenylamine

Diphenylmethane

Fluorene

Hexamethylbenzene

N,N-dimethylaniline

Phenol

b
Thioanisole

Thiophenol


Toluene

1,3,5-Triisopropyl-
benzene


36.9

18.6



42.8

28.3

5.8

38.6

34.1

59.8

29.7

36.3


70.8

50.4


36.0

45.8


242 C

135-136 C

143-1460C

204-2060C

224-228 C

180-181C

136-138 C

163-165C

280-283 C

153-156.5 C

204-206 C

0
119-122 C

236-238 C


286-289

264-265 C


250 C

137.5-138.5 C


1650C


Fischer-Hafner

Fischer-Hafner

Methylation

Fischer-Hafner

Fischer-Hafner

Fischer-Hafner

Fischer-Hafner

Fischer-Hafner

Fischer-Hafner

Methylation

Nucleophilic Sub-
stitutionc

Methylation

Nucleophicic Sub-
stitution

Fischer-Hafner

Fischer-Hafner


a.
With two exceptions all salts are hexafluorophosphate salts.

bComplexes obtained as iodide salts.

Complexes prepared using the T-chlorobenzene salt as the starting compound.


METHOD OF
PREPARATION








26
which was reported to be 165C by Astruc and Dabard, but which was

found here to be 286-2890C (decomposition). This discrepancy is ex-

plainable by assuming that Astruc and Dabard used an open melting

point capillary instead of a sealed evacuated capillary tube. The

'H NMR spectra of the preceding complexes show absorption in the 4.62-

5.356 region, corresponding to the cyclopentadienyl protons. The co-

ordinated arene protons show absorption in the range of 5.80-7.386. The

chemical shift data for each compound are tabulated in Table II.

As referred to in the introduction, reactions of arenes substituted

with heteroatoms in the Fischer-Hafner synthesis are complicated by the

possibility of substituent reactions with either the aluminum chloride

or aluminum present in the reaction mixture. Nesmeyanov et al.,in their

preparation of n6-chlorobenzene-n5-cyclopentadienyliron(II) tetrafluoro-
18
borate, reported that dehalogenation of the chlorobenzene starting

material occurred under the conditions used and subsequently, a mixture
a
of i-biphenyl and r-chlorobenzene cations was obtained. Nesmeyanov

and co-workers reported further that it is possible to separate these

two products on alumina. In this work the Russian synthesis was modi-

fied to avoid the dehalogenation problem. The preparation of pure n6-



aIt should be noted that Nesmeyanov determined the presence of r-
biphenyl by the oxidative cleavage of his product and analysis of
the arenes obtained. In a similar reaction with bromobenzene
Nesmeyanov reports the reaction proceeds to give exclusively n-
benzene and not r-blphenyl. 'H NMR analysis of the product mixture
obtained here when n6-chlorobenzene-n5-cyclopentadienyliron(II)
cation was prepared in the presence of aluminum powder shows that
the product obtained by dehalogenation is n6-benzene-n5-cyclopenta-
dienyliron(ll) cation and not n0-biphenyl-nS-cyclopentadienyliron(lI)
cation. However, Khand et al. reported that under conditions si-
milar to that reported by Nesmeyanov dehalogenation was not a pro-
blem. This is in contrast to the work by Nesmeyanov and the work
presented here. There does not appear to be a rational explanation
for the difference.









TABLE II
a
'H NMR Spectral Data for
n6-Arene-n5-cyclopentadienyliron( I) Salts


CYCLOPENTA- COORDINATED UNCOORDINATED
DIENYL ARENE ARENE


Aniline 4.98
(s,5)

Anisole 5.15
(s,5)


c
Anisole


5.25
(s,5)


Carbazole 4.62
(s,5)



Chloro- 5.35
benzene (s,5)

Diphenyl- 4.95
amine (s,5)

Diphenyl- 5.22
methane (s,5)

Fluorene 4.85
(s,5)


Hexamethyl- 4.74
benzene (s,5)


'Chemical shift is
tetramethylsilane.
(splitting pattern,
hertz).


5.80-6.38
(m,5)

6.34
(s,5)


6.50
(s,5)

6.07-6.42 8.28-8.50
(m,2) (m,1),
7.07-7.78
(m,5)d

6.40-6.90
(m,5)

5.85-6.35 7.02-7.62
(s,5) (m,5)


6.47
(s,5)

6.35-6.63
(m,2)
6.93-7.38
(m,2)


7.37
(s,5)

7.38-7.82
(m,3)
8.00-8.33
(m,l)


NH 2 5.57-5.80 (m,2)


CH3, 4.04 (s,3)


CH 4.08 (s,3)


NH, 10.38 (s,l)


NH, 7.87 (s,l)


CH2, 4.25 (s,2)


exo CH, 4.23 (d,l, J=23)
endo CH, 4.50
(d,l, J=23)


CH3, 2.58 (s,18)


expressed in delta(6) values relative to internal
'H NMR results are expressed as chemical shift
relative number of protons, coupling constant in


b.
With two exceptions all salts were hexafluorophosphate salts and
spectra were obtained using acetone-d6.
C'Salts were obtained as iodides and spectra were obtained using chloro-
form-d as the solvent.
'Overlapping multiplets of 2 coordinated and 3 uncoordinated arene protons.


ARENEbc


OTHER











a,b
ARENE


N,N-dimethyl-
aniline


CYCLOPENTA-
DIENYL

5.11
(s,5)


Phenol 5.12
(s,5)
b
Thioanisole 5.28
(s,5)

Thiophenol 5.20
(s,5)

Toluene 5.18
(s,5)

1,3,5-triiso- 5.07
propylbenzene (s,5)


a.With two exceptions
spectra were obtained
b
Salts were obtained
form-d as the solvent.


TABLE II (continued)


COORDINATED UNCOORDINATED
ARENE ARENE

6.08-6.37
(m,3)
5.83-6.08
(m,2)

6.28
(s,5)

6.65
(s,5)

6.28-6.78
(m,5)

6.40
(s,5)

6.27
(s,3)


OTHER


CH3, 3.21 (s,6)




OH, 9.05 (s,1)


CH3, 2.81 (s,3)


SH, 3.60-3.92 (s,l)


CH3, 2.58 (s,3)


CH, 3.22 (m,3, J=6.5)
CH3, 1.42 (d,18, J=6.5)


all salts were hexafluorophosphate salts and
using acetone-d6.

as iodides and spectra were obtained using chloro-









chlorobenzene-n5-cyclopentadienyliron(ll) hexafluorophosphate was ac-

complished by doing the reaction in the absence of aluminum powder.

Consequently, it was necessary to reduce the ferricenium ion formed

during the reaction prior to the precipitation of the product with am-

monium hexafluorophosphate. Significantly, the yield of product was

not adversely affected by the absence of aluminum as seen on Table I.

The 'H NMR data for this compound are given in Table II and are consistent
20
with the data provided by Khand et al.

While these r-arene-i-cyclopentadienyliron(ll) cations were being
29
prepared,there was a report by Koch, Miller, and Osteryoung about elec-

troinitiated Friedel-Crafts reactions in the room temperature eutectic,

ethylpyridinium bromide-aluminum chloride. The paper was pertinent to

this research as both benzene and ferrocene are soluble in the eutectic.
16
In the mechanism proposed by Astruc and Dabard for the ligand exchange

reaction they attributed the low yields to the formation of an insoluble

complex of aluminum chloride and ferrocene. It was hoped that the use

of this eutectic would obviate this solubility problem and hence greater

yields would be possible. When the ethylpyridinium bromide-aluminum

chloride eutectic was prepared and benzene, ferrocene, and aluminum

were added and then heated at 80C for 3 hours, n6-benzene-n5-cyclo-

pentadienyliron(ll) hexafluorophosphate was obtained (Figure XII).

-However, through 'H NMR spectral analysis the product was shown to be

a mixture of the desired hexafluorophosphate salt and a salt of ethyl-

pyridinium bromide. The mixture was recrystallized and chromatographed.

However, no separation was observed. Further reactions using this sol-

vent were attempted, but no product at all was obtained when either di-

phenylamine or anisole was used as the arene. This reaction potentially









has great promise. However, the purification and general procedure

needs substantially more work.

The preparation of nitrogen, oxygen, and sulfur substituted r-
18, 19
coordinated arenes has been explored primarily by Nesmeyanov et al.

They reported that significant limitations were encountered when arenes

substituted with oxygen, sulfur, or nitrogen were used in the Fischer-

Hafner synthesis. It was possible to prepare n-complexes of anisole and

other aromatic ethers by this method, but it was not possible to pre-
19
pare a r-complex of phenol by this method. Also, Nesmeyanov et al.

reported that while the r-acetanilide complex could be prepared directly,
18
that of aniline could not. The low reactivity of phenol and aniline

reflects the ability of their substituent groups to coordinate with al-

uminum chloride. Our interest in the preparation of these complexes

was stimulated by observations made in the early stages of this work

which suggested that a series of coordinated cyclohexadienyl complexes

bearing exocyclic C=N, C=0, or C=S double bonds might be synthesized

by the reaction of a suitable base with the n-complexes of anilines,

N-alkyl anilines, phenols, or thiophenols. The reaction of rn-fluorene-

n5-cyclopentadienyliron(ll) hexafluorophosphate with sodium hexamethyl-

disilylamide was found to give the relatively stable, neutral depro-
30
tonated complex. The formation of a stable deprotonated t-complex

from the n1-fluorene-n5-cyclopentadienyliron(II) cation provided the

impetus to attempt the preparation of the analogous cationic complex
18
derived from carbazole. On the basis of the work by Nesmeyanov et al.

there was considerable doubt as to the chance of success and consequently,

the high yield, 42.8%, obtained when carbazole, ferrocene, and aluminum

chloride were reacted was unexpected. Subsequent reactions of diphenyl-









amine and aniline, however, were only partially successful. The re-

action of diphenylamine with ferrocene and aluminum chloride provided

the cyclopentadienyliron complex of diphenylamine in only a 6% yield.

Initially, the reaction of aniline to prepare the corresponding w-complex

was unsuccessful. A reaction of anilinium sulfate, ferrocene, and alum-
a
inum chloride also failed.

For aniline this problem was overcome by increasing the reaction

temperature. Excess aluminum chloride was necessary to compensate for

the coordination of aniline and aluminum chloride, while the high temp-

erature offset the lowered reactivity of the aniline due to its complex-

ation with the aluminum chloride. When these changes were made, n6-

aniline-n5-cyclopentadienyliron(II) hexafluorophosphate was obtained

in 36.9% yield.

The 'H NMR spectra obtained for these compounds is consistent with

the proposed structures (Table II). The 'H NMR spectrum of the T-

aniline salt is the same as that reported by Nesmeyanov et al. The IR

spectra of the above compounds show in all cases an absorption at ap-
-l
proximately 840 cm corresponding to the hexafluorophosphate anion.

An N-H or NH2 stretch is seen at 3389 cm 3417 cm or 3499 and

3401 cm-l for the diphenylamine, carbazole, and aniline complexes, re-

spectively.

When the conditions successfully used to prepare the i-aniline com-

plex were applied to a reaction of phenol, ferrocene, and aluminum chlor-

ide, only a black viscous oil was obtained. The reaction of sodium hy-



a'The reaction of anilinium sulfate, ferrocene, and aluminum chloride
was attempted as a possible way to avoid coordination of the aniline
and aluminum chloride. The failure of this reaction may be attributed
to the low solubility of the anilinium sulfate in the decahydronaph-
thalene and the positive charge of the arene salt.









droxide and a n6-chlorobenzene-n5-cyclopentadienyliron(Il) salt would

seem to be a reasonable method for preparing the coordinated phenol.

However, when Nesmeyanov and co-workers attempted this reaction in water
19
at 50'C, only extensive decomposition occurred. Due to this report

another route was sought for the preparation of the r-phenol complex.

The preparation of phenol by the reaction of anisole and hydrogen iodide

is a well-known reaction. However, the reaction of n6-anisole-n5-

cyclopentadienyliron(II) hexafluorophosphate with hydrogen iodide did

not give the desired product; only decomposition was noted. Since it

was not possible to prepare the 7-phenol complex either directly or by

using hydrogen iodide and the i-coordinated anisole complex, attention

was refocused on the reaction of the i-chlorobenzene complex and hydrox-

ide. In retrospect, the use of elevated temperatures was the probable

reason that decomposition occurred when the reaction was attempted by

Nesmeyanov et al. In a strongly basic media any iron-coordinated phenol

formed would immediately react with another equivalent of base to form

the deprotonated r-phenol complex. In this work the deprotonated w-

complexes were found to be unstable in solution at elevated temperatures.

Therefore, when n6-chlorobenzene-n5-cyclopentadienyliron(II) hexafluoro-

phosphate was reacted in 50% aqueous acetone at room temperature for 24

hours, followed by acidification and recrystallization, n6-phenol-n5-

Scyclopentadienyliron(ll) hexafluorophosphate was obtained in 36.3% yield

(Figure Xlla). The 'H NMR spectrum of the proposed coordinated i-

phenol complex shows the expected absorptions for the cyclopentadienyl

group and coordinated arene protons. The phenolic proton is observed

as a broad peak at 9.056. Also, there is a strong, broad absorption at
-1
3495 cm corresponding to an 0-H stretch.










FePF
Fe* PFr'


+ 2 NaOH


L HCI

2. NH4PF6


Fe PF6
6


+ 2 NoHS


AQUEOUS ACETONE

24 HOURS


<=>OH
Fe4 PFE
4


ACETONITRILE

20 MINUTES


I. HCI ----- Fe* PF
2. NH4PF6 6


FIGURE XIII


Fe


Fe









One additional iron salt was prepared by an analogous route. n6-

Thiophenol-n5-cyclopentadienyliron(II) hexafluorophosphate was prepared

by the reaction of sodium hydrogen sulfide with q6-chlorobenzene-n5-

cyclopentadienyliron(II) hexafluorophosphate in acetonitrile. In this

reaction the deprotonated i-thiophenol complex was also initially ob-

tained. The hexafluorophosphate salt was obtained by acidifying the re-

sulting solution with concentrated hydrochloric acid (Figure Xllib).

The 'H NMR spectrum shows an absorption at 3.6-3.926 of relative area

corresponding to one proton. The IR spectrum shows the characteristic
-1
absorption for an S-H stretch at 2584 cm

Mass spectral data have been obtained for all the new compounds.
32
The spectra are consistent with those reported by Games et al. The

parent cation is observed without the hexafluorophosphate anion and char-

acteristically the base peak is ferrocene or the arene ligand. For the

aniline, thiophenol and phenol complexes a substantial number of large

peaks of greater mass than the parent peak are observed. Two high mass

peaks were observed in the mass spectrum of the i-aniline complex that

deserve special note. A peak at m/e 372 corresponding to C H 0Fe2 and

a peak at m/e 307 corresponding to C H Fe were seen. These peaks

were also seen in the mass spectrum of the deprotonated T-aniline com-

plex and will be discussed in the next section.

The observation of peaks at a greater mass than the parent peak

seems to correlate with the observed method of decomposition. In meas-

uring the decomposition temperatures of a number of the w-arene-i-cyclo-

pentadienyliron(II) salts two distinct decomposition modes were observed.

The more common mode is a darkening of the product as the temperature is

raised followed by the formation of a black oil at the decomposition









temperature. An alternative method of decomposition does not form a

black oil at the decomposition temperature, but instead turns white with

a concurrent formation of a green or red oil at the top of the capillary

tube. For those compounds which decomposed in the second mode it was

not only difficult to determine a decomposition temperature, but the

mass spectra showed the peaks above the parent peak. Perhaps in decom-

posing these compounds form dimers, trimers, etc., which are observed

in the mass spectrum as high mass peaks.

The results presented here on the preparation of T-arene-i-cyclo-

pentadienyliron(lI) salts indicate that the limitations imposed on their

preparation by the presence of aluminum chloride or aluminum may be over-

come by a judicious choice of reaction conditions or by choosing an al-

ternate route to their preparation. The work done by Nesmeyanov et al.18' 19

needs to be reviewed further in light of the results presented here. In

the cases where a low yield or no product is reported there is a reason-

able chance that those results could be easily changed with slightly dif-

ferent conditions.















DEPROTONATION OF n6-ARENE-n5-CYCLOPENTADIENYLIRON(II) SALTS


22
The paper by Helling and Cash provided the first evidence of in-

creased acidity of alpha protons in r-arene iron(ll) salts. However, a

substantial amount of work has been reported in which increased acidity

of alpha protons of other r-arene metal systems is observed.

In 1971 Kang and Maitlis33 showed that the protons of 5S-pentamethyl-

cyclopentadienylrhodium(III) chloride dimer could be completely exchanged

using sodium deuteroxide in deuterium oxide (Figure XIV). The exchange

was effected in 72 hours at 700C. To explain this facile exchange of

the alpha protons they suggested that an inductive and a mesomeric stab-

ilization of the intermediate carbanion must occur. In 1971 two papers

were presented that dealt with the r-arenechromium(0) tricarbonyl sys-
34
tem. Wu, Biehl, and Reeves reported that when pKa values for a series

of arene substituted r-phenolchromium(0) tricarbonyl complexes were ob-

tained, there was a large increase in acid strength for all the r-phenol

complexes with respect to the uncoordinated phenols. When the pKa values

were plotted against o- values for the corresponding substituent, a good

correlation was obtained. The fact that o- and not a values gave the

test correlation indicates that mesomeric interaction is of particular
35
importance in this system. Trahanovsky and Card reported that when

n6-1,4-diphenylbutanechromium(O) tricarbonyl is reacted with potassium

t-butoxide in DMSO-d6, followed by cleavage of the ring from the metal
-- 6
using ceric ammonium nitrate, a 70% yield of 1,4-diphenyl-l,l-dideuterio-

butane was obtained. Or, if n6-indanechromium(O) tricarbonyl is treated









,75_ (CH3)5C5Rh /
2


D20

NoOD


FIGURE XIV


K< CH2)4Ph
0 Cr

oC


L KO-t-bu, DMSO-d6

2. Ce'


CD3 5


DO '
D

Rh

-(CD3)5





< )CD 2(CH )3Ph


D


I. KO-t-bu, DMSO-d6

2. Ce'v


FIGURE XV


.- >CH 2CH3
C.CrO +

C
0


KO-t-bu
Mel ---
DMF


.Cr


0 R= i-pr or
t-bu


FIGURE XVI


OC/ CO
C









in the same fashion, cis-l,3-dideuterioindane is obtained (Figure XV).

The deuterium exchange experiments by Trahanovsky and Card and the pKa

measurements by Wu et al. show that increased acidity can be expected

with i-arenechromium(O) tricarbonyls. This was reiterated in 1975 by
36
Jaouen, Meyer, and Simmonneaux. They reported the reaction of potas-

sium t-butoxide, methyl iodide, and a variety of r-arenechromium(O) tri-

carbonyls in DMF. For instance, the reaction of n6-ethylbenzenechromium(O)

tricarbonyl with the above reagents yields n6-isopropylbenzenechromium(o)

tricarbonyl and n6-t-butylbenzenechromium(O) tricarbonyl (Figure XVI).

Under similar conditions the uncoordinated ethylbenzene did not react

with methyl iodide.

By the addition of hydrochloric acid to bis(r5-indenyl)iron(II)

Johnson and Treichel were the first to prepare n6-indene-n5-indenyliron(ll)

hexafluorophosphate. Furthermore, it was possible to regenerate the

starting material by addition of n-butyllithium to the cation. Conceiv-

ably, the regeneration of bis(n5-indenyl)iron(lI) was brought about by the

formation of a T-cyclohexadienyl intermediate with an exocyclic double

bond similar to the intermediate proposed by White, Thompson, and Maitlis
38
in 1976 (Figure XVII). White et al. reported that when trisacetone-

n5-pentamethylcyclopentadienylrhodium(III) hexafluorophosphate or the

iridium analog is reacted with indene or indole, a complex is formed in

which the ligand and metal are bonded via the six-membered ring of the

arene. However, upon standing or upon the addition of a suitable base

the complex isomerizes to a complex in which the metal is coordinated to

a five-membered ring. White and co-workers proposed that the rearrange-

ment occurred via deprotonation to produce an intermediate coordinated

cyclohexadienyl complex with an exocyclic double bond. This then iso-

















[5(CH35C5 FM(ACETONE





[15-(CH3)5C-MTACETONE)j++


HCI
C-------->
n-buLi





H



X = N ,CH


H


M++

(CH35


M = Rh, Ir


-H


FIGURE XVII


<.~ NHPh
n+
o Co
0


NaOH

EtOH


< N/Ph

Mn


0


FIGURE XVIII


M+

s(CH3A5


mH35


0 A









merized to the observed product. The preceding 6 papers and the paper
22
by Helling and Cash provide a variety of examples of increased acidity

of coordinated arene alpha protons which, in some cases, point to the

formation of a r-cyclohexadienyl system as an intermediate.
39
In 1974 Pauson and Segal reported the isolation and characteriza-

tion of (l-5-ns-6-phenyliminocyclohexadienyl)manganese(l) tricarbonyl.

They reacted n6-diphenylaminemanganese(l) tricarbonyl cation with ethan-

olic sodium hydroxide to obtain the neutral complex (Figure XVIII). From

'H NMR and IR spectral analysis they assigned the structure as a i-cyclo-

hexadienyl system bearing an exocyclic phenylimine group. Sheats, Miller,
40
and Kirsh in 1975 also reported a structure that they believed con-

tained an exocyclic double bond (Figure XIX). The deprotonation of n5-

[(diphenylmethyl)cyclopentadienyl]-ns-cyclopentadienylcobalt(l I) cation

with ethanolic sodium hydroxide resulted in a complex with substantial

solubility differences compared to the starting cation that suggested

the exocyclic double bond structure. A second paper on the acidity of
41
alpha protons in cobalticenium cations was presented by Sheats et al.

in which the reaction of n5-cyclopentadienylcobalt dicarbonyl or n5-cyclo-

pentadienylrhodium dicarbonyl with dimethyl-, diphenyl-, or diperfluoro-

phenyl acetylene provided products that were best represented as n4-cyclo-

pentadienone-n5-cyclopentadienylcobalt( II) complexes or the analogous

rhodium(lll) complexes rather than zwitterionic structures like that pro-
42
posed by Markby, Sternberg, and Wender.
43
In 1976 Johnson and Treichel reported the reaction of n6-fluorene-

n5-cyclopentadienyliron(Il) hexafluorophosphate with potassium t-butoxide

in toluene to give a new product that is best represented as a zwitter-

ionic structure. The analogously prepared,deprotonated T-fluorenemanganese(1)










^'CHPh2
Co+


PC
0 0
M:Co, Rh


NaOH

EtOH


/ Ph
Co


R4


+ RC=CR M


R= Me, Ph, PhF


FIGURE XIX


M


M=175-C5HSFe+, Mn(CO)3+


Mn
oC\ '-'Co
C
0






Fe*


FIGURE XX









tricarbonyl complex was proposed to be a complex in which the exocyclic

double bond is substantially more important. In 1977 further elabora-
44, 45
tion of this work was published in two separate papers' in which
13
both compounds are shown by C NMR to have substantial negative charge

residing at the 9-position. However, the neutral manganese complex ap-

pears to have less charge density at the 9-position and hence is less

dipolar in nature than the neutral iron complex. The x-ray structure

for the neutral iron complex was also obtained and it showed only a

slight difference in the bond lengths of the two carbon-carbon bonds

to C(9) and only a 110 deviation of C(10) from the plane of the coor-

dinated arene. Finally, the neutral iron complex was much more reactive

toward electrophiles than the neutral manganese compound. Therefore, the
13
x-ray structure, the C NMR, and the reactivity all argue for substan-

tially more dipolar character in the neutral iron complex than in the

neutral manganese complex (Figure XX).
46
In 1976 Cole-Hamilton, Young, and Wilkinson reported the prepara-

tion of (l-5-n5-cyclohexadienyl-6-one)bis(triphenylphosphine)ruthenium(II)

hydride. Based on 'H NMR, IR, and x-ray analysis they proposed that the

compound was best represented as a w-cyclohexadienyl metal complex bear-

ing an exocyclic carbon-oxygen double bond (Figure XXI). The 'H NMR

spectrum obtained in benzene-d shows the expected 1:2:2 splitting of
6
the i-cyclohexadienyl protons. However, in chloroform-d the ortho and

para protons are observed as one peak. This difference was attributed

to preferential solvation of one portion of the molecule. The IR spec-
-l
trum shows a carbon-oxygen double bond stretch at 1577 cm- and the x-

ray structure of the above compound with a molecule of methanol sol-

vated to it reveals that the carbon-oxygen bond is intermediate in

length relative to carbon-oxygen single and double bonds found in model









PhOH < E
*Ru
PICH3 Ph3P\/ v H

Ph3


HOPh
+ Ru *HOH
Ph3 H
Ph3p


FIGURE XXI


<0

0


0Cl/CrA
0


FIGURE XXII


,?-(CH)5C,5Rh(ACETONE)3 + PhOH


Na2CO3


3+

<^3)OH -j--
Rh Rh

"CH 315 4~; H-5
L J1


Rh'

(CH3) 5


FIGURE XXIII


R PPh3)4









compounds. Finally, as in other cyclohexadienyl systems, they noted

the nonplanarity of the ring. An exocyclically double-bonded chromium
47
complex was reported in 1977 by Trahanovsky and Hall. They reported

that (T8-allyl phenyl ether)dicarbonylchromium isomerized in benzene to

(n3-allyl)(l-5-rs-cyclohexadienyl-6-one)chromium dicarbonyl complex at

room temperature (Figure XXII). The 'H NMR spectrum of this complex has

two i-cyclohexadienyl peaks of relative area 3:2 and a carbonyl absorp-
-1
tion is observed at 1597 cm in the infrared spectrum.
48
In 1977 White, Thompson, and Maitlis reported that the reaction

of trisacetone- 5-pentamethylcyclopentadienylrhodium(II I) hexafluoro-

phosphate with phenol produces a hydrogen bridged dimer. When this

dimer was treated with sodium carbonate, (l-5-n5-cyclohexadienyl-6-one)-

n5-pentamethylcyclopentadienylrhodium(II ) cation was obtained (Figure
-l
XXIII). The carbon-oxygen stretch in the infrared is observed at 1630 cm

and the 'H NMR spectrum is similar in nature to the 'H NMR spectrum ob-

tained by Cole-Hamilton et al. with the previously discussed ruthenium
46
complex in chloroform-d

The preceding paragraphs present an overview of the variety of arenes

that, upon coordination to a metal, show an increase in acidity of alpha

protons. Depending on the particular arene and-the metal to which it is

coordinated, deprotonation can be effected during the preparation with-

.out the addition of a base. In other cases, strong bases such as po-

tassium t-butoxide was needed to effect deuterium exchange or deproton-

ation. For the deprotonation of the 7-arene-r-cyclopentadienyliron(Il)

cations prepared in this study a variety of bases were found to give

satisfactory results. The base chosen to react with a particular com-

plex was selected primarily to allow an easy workup. This assumes that









the base was of sufficient strength to remove the alpha protons. To

prepare the neutral iron cyclopentadienyl complexes of fluorene, diphenyl-

amine, carbazole, and triphenylmethane, sodium amide in liquid ammonia

was used (Table III).22 Typically, 2-3 mmoles of the iron salt was re-

acted with a 10-fold excess of sodium amide in 90 mL of liquid ammonia

for 1-3 hours. While the deprotonated r-fluorene complex could be pre-

pared in liquid ammonia with sodium amide, this complex was best pre-

pared in benzene or diethyl ether using sodium hexamethyldisilylamide

(Figure XXIV). When the reaction was run in liquid ammonia, an anoma-

lous temperature dependent peak appeared in the 'H NMR spectrum that

shifted from 3.26 at 400C to 4.26 at -60C. A possible explanation for

this peak might be a coordinated molecule of ammonia, but infrared anal-

ysis of the compound never showed peaks characteristic of ammonia. To

avoid this complication the reaction was run in benzene or diethyl ether.

A 43% yield of an intense blue-green product was obtained. Chemical

shift data in benzene-d for this compound are presented in Table IV
45
and are consistent with the spectrum reported by Johnson and Treichel.

However, the 'H NMR spectrum of this neutral complex shows a remarkable

downfield shift of the cyclopentadienyl absorption from 3.376 to 4.026

when acetone-d is the solvent. Similar solvent effects were noted

previously and were attributed to preferential solvation at a particular
46
portion of the molecule. The blue-green color of the deprotonated r-

fluorene complex is unusual in light of the red or red-orange colors seen

for the other neutral complexes prepared in this work. Blue-green com-

plexes for these r-arene-r-cyclopentadienyliron(ll) cations have been
49
reported for the reduced species, and in other metal arene systems
50
charge transfer complexes are known. The 'H NMR spectrum shows no








NoNH F
Fe+PF- NH2 Fe+ 'NH3
M;3


NoNSi2(CH3)6

BENZENE


Fe


FIGURE XXIV


< >-NHPh
Fe+PFE





H


Fe+ PF6-







K<^HPh2
Fe PFg-
4 6


NaNH2

NH3






NaNH2
---------->





NH3






NoNH2
NH3--------
NH3






NaNH2

NH3


Ph

Fe








Fe+








Fe


FIGURE XXV









TABLE III

Deprotonated n6-Arene-15-cyclopentadienyl


ARENE


Aniline

Carbazole

Diphenylamine

Fluorene


Phenol


YIELD (%)


79.1

90

85

43


Thiophenol 55.3


Triphenyl- 89
methane


iron(II) Complexes


DECOMPOSITION BASE AND SOLVENT
POINT USED FOR PREPARATION

140-1460C sodium amide, dichloromethane

176-1780C sodium amide, ammonia

122-124C sodium amide, ammonia

105-107C sodium hexamethyldisilylamide,
diethyl ether

112-1140C sodium hydroxide,
aqueous acetone

183C sodium hydrogen sulfide,
acetonitrile

115-117C sodium amide, ammonia







TABLE IV

'H NMR Spectral Data for Deprotonated
16-Arene-n5-cyclopentadienyliron(II) Complexes


CYCLOPENTA- CYCLOHEXADIENYL OR
DIENYL COORDINATED ARENE

4.53(s,5) 5.30-5.63(m,3)
4.83-5.05(m,2)


Carbazole 4.13(s,5) 6.93-7.18(m,2)
5.50-5.82(m,2)


Diphenylamine 4.48(s,5)


4.58-4.82(m,2-under
cyclopentadienyl peak)
5.17-5.55(m,2)
5.57-5.83(m,1)


4.02(s,5) 6.75-7.13(m,2)
6.38-6.57(m,1)
5.27-5.72(m,2)

3.37(s,5) 4.16-4.75(m,2)
5.67-6.16(m,3)

4.57(s,5) 4.67-4.97(m,2)
5.33-5.63(m,3)


UNCOORDINATED
ARENE


8.07-8.27(m,l)
7.27-7.80(m,2)
6.65-6.93(m,l)

6.72-7.00(m,3)
7.05-7.42(m,2)



8.08-8.33(m,l)
7.13-7.68(m,2)


7.50-8.25(m,4)


Thiophenol 5.02(s,5) 6.00-6.37
(s, broad, 3),
6.37-6.84
(s, broad, 2)
f
Thiophenol 4.70(s,5) 5.37-6.00
(s, broad, 3),
6.00-6.34
(s, broad, 2)

Triphenyl- 4.55(s,5) 4.12-4.38(m,2)
methane 4.82-5.11(m,2)
5.94-6.25(m,l)


6.88-7.46(m,10)


a.Chemical shift is expressed in delta(S) values relative to internal tetramethyl-
silane. 'H NMR results are expressed as chemical shift (splitting pattern, rela-
tive area).
'With few exceptions all spectra were obtained using acetone-d .

C.Acetonitrile-d was used as the solvent.
3
d'Overlapping multiplets of one coordinated and one uncoordinated arene ring proton.

e.Benzene-d6 was used as the solvent.

fChloroform-d was used as the solvent.


a,b
ARENE


Anilinec


OTHER


NH, not
observed


Fluorene


e
Fluorene


Phenol


vinyl CH
5.75(s,l)


vinyl CH
6.13(s,1)









broadening that might be expected for such an intramolecular complex

charge transfer, but the charge transfer hypothesis is intriguing. Fi-

nally, the deprotonated r-fluorene complex was not observed to isomerize

even when heated to 78C in benzene. This is consistent with the results

reported by Treichel and Johnson.45

The remaining three salts were reacted with sodium amide in liquid

ammonia to give intense red solids after workup (Figure XXV). The

85-90% yields were substantially higher than those from the reaction of

the i-fluorene complex with the silicon amide. The 'H NMR spectra for

the three deprotonated w-complexes show a 0.37-0.53 ppm upfield shift

of the cyclopentadienyl protons when compared to the spectra of their

respective cations (Table IV). The 'H NMR spectrum of the deprotonated

r-carbazole complex is analogous to that of the neutral fluorene complex,

while the deprotonated, neutral r-diphenylamine and i-triphenylmethane

complexes show the expected 1:2:2 splitting of coordinated cyclohexadi-

enyl protons.

High temperature 'H NMR spectra of the deprotonated i-triphenyl-

methane complex were obtained in hopes of observing fluxional behavior

in the complex. However, when a tetrachloroethylene solution of the de-

protonated complex was prepared and a series of spectra was obtained

from 40C to 90C this was not observed. Interestingly, in contrast

with most deprotonated T-complexes the deprotonated f-triphenylmethane

complex did not decompose at elevated temperatures.

The spectra are similar to those of other r-cyclohexadienyl systems

such as the nucleophilic addition products of T-arene-i-cyclopentadienyl-
14,21
iron(II) cations 1 and to the reported 'H NMR spectrum of (1-5-q5-6-
46
phenyliminocyclohexadienyl)manganese(l) tricarbonyl. However, in 1970

Schrock and Osborn5 reported that the reaction of rhodium(Ill) chloride









with 1,5-cyclooctadiene and sodium tetraphenylborate gave n6-tetraphenyl-

borate-n4-(l,5-cyclooctadiene)rhodium(I). This compound is a bona fide

zwitterionic complex and yet, the 'H NMR spectrum also shows a 1:2:2

splitting of the arene protons. Also, in the w-cyclohexadienyl-r-cyclo-

pentadienyliron(ll) complexes formed by nucleophilic addition (Figure VI)

the protons ortho to the site of attack are at substantially higher

field than observed here. This may be a reflection of less than com-

plete exocyclic double bond character and hence, lower charge density in

the ring.52 Therefore, the assignment of i-cyclohexadienyl structures

to the deprotonated r-diphenylamine and n-triphenylmethane iron complexes

is not justified on the basis of 'H NMR spectral evidence alone.

Analysis of the infrared spectra of these complexes does provide

an insight into the nature of these complexes. In all cases the neutral

complexes show the loss of the hexafluorophosphate anion by a dramatic

decrease in intensity of the peak at 840 cm Also, there is a loss of

the N-H stretch in going from the i-diphenylamine and i-carbazole cations

to their respective neutral n-complexes upon deprotonation. Correspond-

ingly, the i-fluorene and r-triphenylmethane complexes show the similar

loss of the aliphatic C-H stretch after deprotonation has occurred. Con-

current with the loss of the N-H stretch following the deprotonation of the

r-diphenylamine complex there is the formation of a strong absorption at
-I
1562 cm This absorption is consistent with the formation of a nitro-

gen to carbon double bond. In the analogous compound, (l-5-n5-6-phenyl-

iminocyclohexadienyl)manganese(l) tricarbonyl, the carbon to nitrogen

double bond stretch is reported as a strong peak at 1584 cm -. It is

unfortunate that a similar absorption can not be unambiguously assigned

to carbon-nitrogen double bond stretch in the deprotonated i-carbazole









-l
complex, although medium intensity absorptions are noted at 1544 cm
-1
and 1513 cm The low intensities and position are not expected for a

carbon-nitrogen double bond stretch. This leads to the conclusion that

the deprotonated r-carbazole complex has a substantial degree of di-

polar character. A carbon-carbon double bond stretching frequency can-

not be unambiguously identified in the infrared spectra of the neutral

T-triphenylmethane complex because of strong absorptions in the 1620-
-1
1580 cm region in both the starting material and the product. How-

ever, the similarity of the 'H NMR spectra of the deprotonated r-tri-

phenylmethane and deprotonated r-diphenylamine complexes suggests a

similarity in their structure. Furthermore, the neutral triphenylmeth-

ane complex is the only neutral complex prepared in which the ortho pro-

tons are at higher field than the r-cyclopentadienyl protons. This is

consistent with a high degree of exocyclic double bond character in this

compound.

The mass spectra for the deprotonated r-fluorene, w-diphenylamine,

T-carbazole, and i-triphenylmethane all give a parent peak with the cor-

rectexact mass. Usually the base peak for these complexes was attribut-

able to the respective arene and large peaks were observed at 186, 121,

and 65, corresponding to ferrocene, cyclopentadienyliron, and cyclopenta-

dienyl cations. The first three compounds give parent peaks with ex-

- tremely low intensity parent peaks while the neutral T-triphenylmethane

parent peak has an intensity of 93% of the base peak. That such a large

peak is seen only in the mass spectrum of the deprotonated i-triphenyl-

methane complex suggests a higher stability for this complex than for

the other deprotonated complexes. This may be attributable to substan-

tially more cyclohexadienyl character in the deprotonated i-triphenylmethane









complex. Exact masses were obtained for the parent ions in these 4 de-

protonated i-complexes. This data was used along with satisfactory iron

analyses in place of carbon-hydrogen elemental analyses to fully charac-

terize the deprotonated t-complexes. Satisfactory carbon-hydrogen ele-

mental analyses were not obtained for these complexes, and this is per-

haps due to their low stability.

Three other iron salts were successfully reacted with various bases

to give neutral complexes. The i-aniline, r-phenol, and i-thiophenol

complexes were deprotonated to provide the respective neutral complex.

As mentioned in the preceding section the reaction of n6-chlorobenzene-

n5-cyclopentadienyliron(llI) hexafluorophosphate with sodium hydroxide

or sodium hydrogen sulfide produced the respective deprotonated complex

by an initial nucleophilic attack of hydroxide or hydrosulfide and loss

of chloride ion followed by the reaction of the resulting complexes with

another equivalent of hydroxide or hydrosulfide to form the deprotonated

complexes (Figure XXVI).

The deprotonated i-phenol complex was prepared by reaction of the

r-chlorobenzene complex with hydroxide ion in 50% aqueous acetone, a

medium in which the w-chlorobenzene complex is soluble. After the re-

action had proceeded at room temperature for 18-24 hours, the acetone

was removed by vacuum evaporation. The deprotonated w-phenol complex

was found to be soluble in the resulting aqueous solution while the un-

reacted i-chlorobenzene complex was not and was easily removed by fil-

tration. Evaporation of this solution under vacuum followed by extrac-

tion with dichloromethane and recrystallization afforded the product in

82.7% yield (Table 11).









< CI
Fe+ PF-


2 NoOH
i


CH3I '< OCH3
S Fe+.-


F- CI
Fe+PF


2 NoHS "- Fe


CH31 < -SCH3
) Fe*l-


H2Z
Fe+ PF-




NaNH2


H
Fe

1 I


I. NaNH2

3. AgNO3


2. CH31


4. NH4PF6


CH31 NHCH3
) Fe+ I-

III


II + III


FIGURE XXVI


Fe+ PF 6




CH31


Fe


2~ I


Fe







The deprotonated i-thiophenol complex was prepared similarly. 16-

Chlorobenzene-n5-cyclopentadienyliron(lI) hexafluorophosphate and sodium

hydrogen sulfide were reacted in acetonitrile for 20 minutes and then fil-

tered. Separation of the desired deprotonated --complex from unreacted

starting material was accomplished by passing the filtrate through an

alumina column. The compound was recrystallized to afford a 55.3% yield

of the deprotonated i-thiophenol complex (Table III). The deprotonation

of the n-phenol and r-thiophenol complexes was also possible by the re-

action of ammonia, sodium amide, or alumina with n6-phenol-r5-cyclopenta-

dienyliron(II) hexafluorophosphate or n6-thiophenol-n5-cyclopentadienyl-

iron(ll) hexafluorophosphate, but the direct preparation from n6-chloro-

benzene-n5-cyclopentadienyliron(ll) hexafluorophosphate produced the de-

sired product in much higher yield.
46
As was noted previously by Cole-Hamilton et al. and by Trahanovsky
47
and Hall a 1:2:2 ratio of the r-arene protons is not necessary for a

complex to be consistent with an exocyclically double-bonded cyclohexa-

dienyl system. When the 'H NMR spectrum of the chromium complex was ob-

tained in acetone-d and the 'H NMR spectrum of the ruthenium complex was
6
obtained in chloroform-d, a 3:2 ratio of cyclohexadienyl protons was ob-

served. However, in benzene-d Cole-Hamilton et al. reported that the
6
ruthenium complex had the expected 1:2:2 ratio for the cyclohexadienyl

protons. The 'H NMR spectra of the deprotonated i-phenol and i-thio-

phenol complexes show a 0.2-0.6 ppm upfield shift of the w-cyclopenta-

dienyl protons when compared to those to the respective starting

cations (Table IV). The i-cyclohexadienyl protons of the deproton-

ated complexes are also shifted upfield 0.5-1.7 ppm. There is a concomi-

tant splitting of the r-cyclohexadienyl protons to give a 3:2 splitting









pattern in both cases when acetone-d is used as a solvent. The 'H NMR

spectrum of the deprotonated T-thiophenol complex in chloroform-d did not

show a 1:2:2 splitting pattern of the i-cyclohexadienyl protons, but

changing the solvent did give an upfield shift of 0.30 ppm for the T-

cyclopentadienyl and i-cyclohexadienyl protons. Conceivably, the 'H NMR

spectra of the deprotonated i-thiophenol or deprotonated T-phenol com-

plexes would exhibit a 1:2:2 splitting of their cyclohexadienyl protons

in benzene-d Unfortunately, neither I-complex is soluble in that sol-
6
vent.

The structures of the deprotonated i-phenol and i-thiophenol com-

plexes cannot be assigned unambiguously from the 'H NMR spectra alone,

but as before infrared analysis is extremely helpful. Upon deprotona-

tion of the i-phenol complex there is a loss of the OH stretch and the
-1
PF stretch with a concurrent gain of a peak at 1661 cm attributable

to a carbonyl stretch. The infrared spectrum of the deprotonated r-thio-

phenol complex is instructive since it shows neither an SH stretch nor
-1
a PF stretch but does show a strong absorption at 1082 cm correspond-

ing to a thione stretch. Therefore, the appearance of thione or carbonyl

absorptions in the infrared spectra and the 'H NMR spectra are consis-

tent with the proposed T-cyclohexadienyl structures bearing exocyclic

double bonds.

Mass spectral data were obtained for both of the deprotonated T-

phenol and i-thiophenol complexes. In each case large peaks were ob-

tained at m/e 186, 121, 66, and 56 corresponding to the cations of

ferrocene, cyclopentadienyliron, cyclopentadiene, and cyclopentadienyl.

The deprotonated i-phenol complex showed an unusually large parent peak

at 214 which was 36.2% of the base peak, phenol. In contrast, the parent









peak of the deprotonated i-thiophenol complex at 230 was only 3.5% of

the base peak, ferrocene. Other strong peaks in the mass spectrum of

the deprotonated i-thiophenol complex were at m/e 186 (C 2HI S), 185

(C12H S), 184 (C2HS), 110 (C H S), and 77 (C H ). As was the case

with the deprotonated r-complexes of fluorene, carbazole, diphenylamine,

and triphenylmethane, it was not possible to obtain satisfactory carbon-

hydrogen elemental analyses for these compounds. However, satisfactory

iron analyses and an exact mass for the parent ion of both of the depro-

tonated i-complexes were obtained.

Methyl iodide was added to separate solutions of both the depro-

tonated 7-thiophenol and r-phenol complexes (Figure XXVI). The forma-

tion of the T-thioanisole complex was confirmed by 'H NMR and IR spectra

and by elemental analysis. The formation of the w-anisole complex was

confirmed by comparing its 'H NMR spectrum to that of an authentic sample

prepared directly from anisole, ferrocene, and aluminum chloride. The

reaction of methyl iodide with both of the deprotonated complexes was

rather slow, requiring an hour for the T-phenol complex. It was also

found that methyl iodide did not react either with the w-phenol or r-

thiophenol complexes to give the corresponding r-anisole or r-thioanisole

complexes. Interestingly, Trahanovsky and Hall47 reported that methyl

iodide did not react with the (n3-allyl)(l-5-ns-cyclohexadienyl-6-one)-

-chromium(0) dicarbonyl complex. On the basis of the lower carbonyl fre-

quency observed for the chromium complex in comparison to that of the

deprotonated T-phenol iron(II) complex, a lower bond order of the carbon-

oxygen bond would be expected in the chromium complex. This in turn

would indicate that the chromium complex should react faster with

methyl iodide than the deprotonated w-phenol complexes prepared here.









Therefore, from this work and from the work by Trahanovsky and Hall it

can be seen that a direct comparison of ground state electron densities

and reaction rates is not valid. This direct relation was assumed in-

correctly by Johnson and Treichel45 in their discussion of the structures

of the i-fluorenyl complexes of cyclopentadienyliron(II) and manganese(l)

tricarbonyl. As seen above this is not valid; however, Johnson and

Treichel do not use this point as the only basis for assigning the

structures of the i-fluorenyl complexes.

The deprotonated w-aniline complex was obtained by reacting n6-

aniline-n5-cyclopentadienyliron(ll) hexafluorophosphate with sodium amide

in dichloromethane for 1.5 hours (Figure XXVI). The compound was ob-

tained as a red solid which decomposed in acetone-d but was stable in

acetonitrile-d3 to give a 'H NMR spectrum similar to that of the depro-

tonated w-phenol and r-thiophenol complexes (Table IV). The deprotonated

t-aniline complex could also be prepared in acetonitrile using NaCH CH.

However, purification of the resulting deprotonated complex was not

possible. The 'H NMR spectrum of the complex shows a number of small

peaks in the 1-36 region which were not removed upon recrystallization.

Also, there were 3 peaks in the IR spectrum in the N-H stretch region

which was indicative of a mixture of deprotonated T-aniline complex and

unreacted T-aniline complex. An attempt was made to prepare the com-

pound in liquid ammonia with sodium amide, but the initial red solution

turned black in less than one minute and no product was obtained. Simi-

lar results were obtained when the reaction was run in tetrahydrofuran

except that the complex did not decompose as fast.

The infrared spectrum of this complex shows only one N-H absorp-

tion; this is consistent with a deprotonated complex. However, there









is still a strong absorption at 840 cmI indicative of the presence of

hexafluorophosphate anion and hence unreacted starting material. The

'H NMR spectrum of this complex shows the expected upfield shift of the

coordinated arene and cyclopentadienyl protons relative to the starting

salt. There is also an observed splitting of the coordinated arene pro-

tons into multiplets of relative area 3:2. The mass spectrum of this

complex was obtainable at 800 C in contrast to that of the i-aniline com-

plex for which a temperature of 150 C was needed. Hence, there are fewer

high mass peaks of large intensity. At 800C large peaks were observed at

m/e 186, 121, 93, 66, 65, and 56 which correspond, respectively, to the

cations of ferrocene, cyclopentadienyliron, aniline, cyclopentadiene,

cyclopentadienyl, and iron. The parent peak, 3.5% of the base peak, was

observed at 213. A correct exact mass for this peak was also obtained.

Correct exact masses were also obtained for peaks at 372 and 307 corres-

ponding to C2 H Fe and C H Fe2. A structure for the first complex

is not immediately obvious, but the second ion is consistent with a

triple decker sandwich, a structure unreported for iron and cyclopenta-

dienyl ligands (Figure XXVII), although triple decker sandwiches are

known for other metals.

While the 'H NMR, infrared, and mass spectra of this deprotonated

complex were consistent with the proposed structure, several iron analy-

ses were unsatisfactory. The complex appears to be sufficiently suscep-

tible to decomposition that quantitative analyses are very difficult.

A derivative of this compound was prepared by deprotonating the i-aniline

salt in dichloromethane and then adding excess methyl iodide. The pre-

viously unknown r-N,N-dimethylaniline complex was obtained in 29.7%

yield after workup and provided satisfactory 'H NMR and IR spectral data.











Fe

Fe


FIGURE XVII


NoNH2

NH3


Ph Ph



Fe


LB


FIGURE XXVIII


<> CH2Ph
Fe*PF6C


CFe Ph
"NH
Fe


Fe +


C/Ph
'Ph









Carbon, hydrogen, and iron analytical data also confirmed the proposed

structure. A rational mechanism to explain why the T-N,N-dimethylaniline

complex was obtained instead of the i-N-methylaniline complex is shown in

Figure XXVI. Addition of the methyl iodide to a solution of the depro-

tonated i-aniline complex would form the expected r-N-methylaniline com-

plex. This complex can then react with an additional deprotonated t-aniline

complex to obtain the deprotonated r-N-methylaniline complex and the i-

aniline complex. The new deprotonated r-N-methylaniline complex could

then react with another equivalent of methyl iodide to form the observed

i-N,N-dimethylaniline complex. This mechanism is supported by the ob-

servation of the i-aniline complex in the 'H NMR and IR spectra of the

crude product. Furthermore, this reaction is analogous to the addition

of acetyl chloride to the deprotonated i-fluorene complex. Johnson and
45
Treichel reported that that reaction gave a mixture of the r-fluorene

complex and a i-diacetylated fluorene complex. The possibility that the

methylated complex was obtained directly from the t-aniline complex may

be ruled out as the i-aniline complex was found to be unreactive with

methyl iodide under similar conditions. An alternate explanation for the

formation of the t-N,N-dimethylaniline complex rests on the proposal that

under the initial reaction conditions both alpha protons on the i-aniline

complex were removed. Methylation would then occur twice without the need

.for a two step methylation as shown in Figure XXVI. This seems unlikely,

but the possibility was checked by adding excess acetic acid-d to separate

acetone-d solutions of the deprotonated i-aniline complex and the it-

aniline complex. Addition of the acetic acid-d to each complex forms the

i-anilinium complex. This is seen by the downfield shift of the N-H pro-

tons from 5.76 in the i-aniline complex to 9.406 in the i-anilinium complex.









The r-anilinium complex formed from the deprotonated i-aniline complex has

a relative area ratio of 1:5:5 for the N-H, coordinated arene, and cyclo-

pentadienyl protons respectively. In contrast, the --anilinium complex

formed from the i-aniline complex has a relative area ratio 2:5:5 for the

N-H, coordinated arene, and cyclopentadienyl protons respectively. The

observation of a peak of relative area two in the 'H NMR spectrum of the

T-anilinium complex obtained from the i-aniline complex shows that pro-

ton exchange does not occur under these conditions. The lack of observ-

able proton exchange and the observation of only one N-H proton in the

'H NMR spectrum of the T-anilinium complex obtained from the deprotonated

w-aniline complex are consistent with the structure shown in Figure XXVI

for the deprotonated i-aniline complex.

Four other i-arene-i-cyclopentadienyliron(II) cations were reacted

with strong bases such as sodium amide or sodium hexamethyldisilylamide.

It was found the i-toluene, i-hexamethylbenzene, and I-1,3,5-triisopro-

pylbenzene complexes did not react to any extent with sodium amide in

liquid ammonia. Presumably this was due to an extremely low acid strength

of the alpha protons even in a cationic metal complex.

When n6-diphenylmethane-n5-cyclopentadienyliron(II) hexafluorophos-

phate was reacted with sodium amide in ammonia or with sodium hexamethyl-

disilylamide in benzene or diethyl ether, an intense red solution forms

immediately. Workup of these solutions in the normal way was done to

obtain a red solid. When an 'H NMR spectrum of the product was obtained,

there was observed a single peak for the arene protons at 7.276 and the

cyclopentadienyl protons were split into two peaks centered at 4.006 and

4.136 of relative area 2:3. The peak at 4.136 was a multiple while the

peak at 4.006 was a singlet. Also, a singlet of relative area 1 was ob-

served at 3.836. The structures in Figure XXVIII are consistent with the









singlet observed for the arene protons. However, this does not explain

the splitting of the cyclopentadienyl protons nor the absence of the re-

quired iron-hydrogen absorption in the infrared spectrum in the region
-I 54
2100-1700 cm Furthermore, the reaction of the deprotonated complex

with hydrochloric acid is inconsistent with such a structure. Acidifica-

tion of the deprotonated T-complex prepared in situ resulted in the re-

covery of the starting r-diphenylmethane cation as evidenced by 'H NMR

spectral analysis. The possibility that fluxional behavior was respon-

sible for the observed spectra was discounted when the 'H NMR spectrum

was found to remain constant down to -600C. No consistent explanation

of these results is yet possible and further research on the system is

needed.

The deprotonation of r-arene-r-cyclopentadienyliron(l ) salts can

give products that show definite exocyclic double bond character. How-

ever, the results upon deprotonation of either the i-fluorene or i-car-

bazole complexes are not as definite. Johnson and Treichel believe that

the reactivity, 1C NMR spectrum, and the x-ray structure all argue for

the deprotonated r-fluorene complex being best represented as a dipolar
43-45
structure. They do point out that there is some shortening of the

carbon-carbon bond from C(9) to the coordinated portion of the molecule

and it is also noted that the dihedral angle between the planes described

by C(9), C(10), and C(ll) and C(1)-C(4) is 10-11 C. This is discounted

as small in comparison to the dihedral angle of 430 seen for bis(6-tert-

butyl-l,3,5-trimethylcyclohexadienyl)iron(lI)55 (Figure XXIX). Their

explanation for this small dihedral angle and hence the dipolar nature

of the deprotonated i-fluorene complex is based upon steric constraints

in the fused ring system.









9









R
S 430


Fe



R t-bu


FIGURE XXIX




0



Ru




O=/(1.4 (0.7) 1.277

8 =sin-' 0.409/ (1.277 + a)] I\


FIGURE XXX









On the basis of these arguments and the fact that the deprotonated

i-fluorene and i-carbazole complexes obtained in this work have similar

'H NMR spectra, a dipolar structure could be suggested for both com-

plexes. However, two points need further consideration. There has been

one other deprotonated r-complex prepared in which a partial x-ray struc-
46
ture was reported. Cole-Hamilton et al. reported that the exocyclic

carbon-oxygen bond for (l-5-n5-cyclohexadienyl-6-one)bis(triphenylphos-

phine)ruthenium(ll) hydride is 1.277A, or midway in length between a

carbon-oxygen single or double bond. Furthermore, they report that the

oxygen atom is 0.409A above the plane of the r-cyclohexadienyl ligand.

Unfortunately, they do not report the actual dihedral angle, but assum-

ing a C(l)-C(6) bond length of 1.4A and using the information given it

is possible to calculate a dihedral angle of 9.90 (Figure XXX). This
43-45
compares favorably with the results obtained by Johnson and Treichel.44

However, in this case there are no steric requirements that hinder the

formation of a larger dihedral angle. The second point to consider

when reflecting upon the importance of the size of the dihedral angle in

these complexes is the report by Hoffmann and Hoffman52 in which they

suggest that the large dihedral angle observed in i-cyclohexadienyl sys-

tems in which the 6-position is saturated is due, in part, to an endo H-

iron orbital interaction, but the major contribution to the formation of

a non-planar system results from an antisymmetruc orbital interaction.

In i-cyclohexadienyl systems formed with an exocyclic double bond, there

would be no endo H-iron orbital interaction and the antisymmetric orbital

interaction would be minimized. Therefore, the size of the dihedral angle

should be smaller. The fact, then, that dihedral angles of 9-110 have

been observed in two deprotonated rr-arene systems is consistent with the









formation of an exocyclic double bond. In this work we have demonstrated

that the deprotonated i-complexes of triphenylmethane, diphenylamine,

phenol, and thiophenol have substantial exocyclic double bond character.

X-ray structural data for these complexes would be helpful in determining

the merit of assuming exocyclic double bond character for a dihedral angle
0
of 9-11.

The deprotonated w-phenol, i-thiophenol, and i-diphenylamine have

the exocyclically double bonded structure. The 'H NMR spectra are con-

sistent with a cyclohexadienyl structure and the infrared spectra have

absorptions corresponding to the expected C=0, C=S, or C=N stretches.

The lack of an assignable C=C stretch for the exocyclic double bond of

the deprotonated i-triphenylmethane complex in the infrared spectrum and

the existence of a strong absorption in the infrared spectrum of n6-anil-

ine-n5-cyclopentadienyliron(ll) hexafluorophosphate that obscures the

C=N region for the deprotonated i-aniline complex does not allow the

unambiguous assignment of these structures. However, the deprotonated

T-triphenylmethane and i-aniline complexes were assigned the exocyclic

double bonded structure based on the similarity of their 'H NMR spectra

to those of the preceding three deprotonated i-complexes.

The preceding five complexes are substantially exocyclic double

bonded complexes, but some degree of dipolar character probably is present

in each. The infrared spectrum of the deprotonated n-diphenylamine com-
-l
plex shows a C=N stretch that is 22 cm- lower frequency than the C=N

stretch reported for (l-5-n5-6-phenyliminocyclohexadienyl)manganese(I)
39
tricarbonyl by Pauson and Segal. This shift to lower frequency for

the C=N stretch is consistent with a lower bond order in the exocyclic

double bond of the iron complex when compared to the manganese complex.









The deprotonated i-phenol complex has a C=0 stretch in the infrared
-1
spectrum that is 30-60 cm higher than any of those previously reported

deprotonated i-phenol complexes.44 50, 51 THis suggests a higher bond

order for the exocyclic carbon-oxygen bond and hence a lower degree of

dipolar character. A greater electron withdrawing effect of the iron

would tend to favor the exocyclic double bond form. This might be re-

flected in a larger dihedral angle between the carbonyl plane and the cyclo-

hexadienyl plane resulting in less polarization of the double bond and

hence giving a higher frequency in the IR spectrum. Conceivably, the

still unknown (1-5-n5-cyclohexadienyl-6-one)manganese(I) tricarbonyl

should show a higher v than the iron complex.
c=o
Finally, there have been no previous reports of a i-coordinated

metal complex containing an exocyclic sulfur to carbon double bond. How-

ever, the fact that the v is observed at the high frequency end of
c=s
the v region suggests that the complex has a strong carbon-sulfur
c=S
exocyclic double bond and that the complex exists primarily in the r-

cyclohexadienyl form.

There remains a substantial amount of work to be done in this field.

The series of exocyclic double bonded compounds should be extended by

changing the metal to which the arene is coordinated and by changing

the arene. Also, reactions that would take advantage of an exocyclic

double bond such as (2+4) cycloadditions, Wittig reactions, or Simmons-

Smith reactions could have some synthetic utility and should be attempted.















EXPERIMENTAL


All reactions were run under a N2 atmosphere. Except for acetone

or methylene chloride, which were reagent grade, solvents used in the

preparation and purification of air-sensitive compounds were distilled

from lithium aluminum hydride or calcium hydride slurries. The prepara-

tion and purification of the hexafluorophosphate salts were done with re-

agent grade solvents. Reagents in all reactions were reagent grade and

were not purified before use. Yields are based on the iron-containing

reagent exclusively. All reactions to prepare the r-arene-r-cyclopenta-

dienyliron(ll) hexafluorophosphate salts from ferrocene and the appro-

priate arene were run in three neck,round-bottom flasks that had been

dried by repeatedly evacuating the system, flame drying, and filling

with nitrogen. All reagents were added to the system against a counter-

current of nitrogen. When the reagents had been added, the reaction

vessel was stoppered and the reaction run under the conditions described.

Stirring of the reaction mixture was accomplished by the use of a mech-

anical stirrer.

Melting points of the compounds were obtained in sealed evacuated

capillaries and are uncorrected. Either a Buchi or Mel-temp melting

point apparatus was used to obtain the melting points.

'H NMR spectra were obtained on a Varian A-60, Varian XL-100, or

Jeol PMX-60 spectrometer. Chemical shift is expressed in delta(6)

values relative to internal tetramethylsilane. 'H NMR results are ex-

pressed as chemical shift (splitting pattern, relative number of protons,









coupling constant in hertz). Splitting pattern abbreviations, s, d, t, and

m, correspond to singlet, doublet, triplet, and multiple, respectively.

Infrared spectra were obtained on a Beckman IR-10 instrument and

mass spectra were obtained on an AEI MS-30 spectrometer.

Microanalyses were obtained from Galbraith Laboratories, Inc.,

Knoxville, Tennessee; PCR, Inc., Gainesville, Florida; or Atlantic

Microlab, Inc., Atlanta, Georgia. Iron analyses were obtained on a

Perkin-Elmer 290B atomic absorption spectrophotometer.


66-Anisole-n5-cyclopentadienyliron(II) Hexafluorophosphate19

Ferrocene (10.67 g, 0.0575 mol), anisole (100 mL)a, aluminum chlor-

ide powder (15.90 g, 0.119 mol), and aluminum powder (2.00 g, 0.0741 mol)

were added to the reaction vessel and heated to 140 C with stirring. At
o
110 C the solution turned black. However, no oil or solid coated the

walls of the reaction flask as was normally observed. After 4 hours the

mixture was cooled to 00C and a 25% aqueous methanol solution (130 mL)

was added slowly to give a yellow aqueous layer. This was filtered and

separated from the orange organic layer. The aqueous portion was washed

twice with 30 mL portions of diethyl ether and then slowly added to a

solution of ammonium hexafluorophosphate (3.0 g, 0.0126 mol) in 2 mL of

water. The yellow precipitate which immediately formed was filtered

and washed with water. This damp solid was dissolved in 40 mL dichloro-

methane. After filtration the solution was added slowly to diethyl

ether. A fine yellow powder was obtained which weighed 3.99 g (18.6%).

m.p. 135-136 C (lit. m.p. 137.5-138.50C).


a.This reagent, used both as a solvent and reactant, was present in
great excess with respect to the limiting reagent, ferrocene.









NMR (acetone-d ): cyclopentadienyl, 5.15 (s,5); coordinated arene,

6.34 (m,5); methyl, 4.04 (s,3).


IR (KBr): 3125 (s), 2960 (w), 1625 (w), 1539 (s), 1471 (s), 1447 (s),

1424 (s), 1388 (s), 1260 (s, broad), 1021 (s), 995 (s), 850 (vs, broad),
-1
666 (m), 557 (s, broad), 490 (s), 479 (s), 462 (s) cm


n6-Chlorobenzene-n5-cyclopentadienyliron( I) Hexafluorophosphatel8, 20

Ferrocene (10.25 g, 0.0551 mol), chlorobenzene (100 mL),a and alum-

inum chloride powder (23.84 g, 0.179 mol) were added to the reaction

vessel and heated to 1300C with stirring. The solution turned black at

1100C. After 4 hours the solution was cooled to 00C and 150 mL of ice

water was slowly added. A green aqueous layer and an orange organic

layer were obtained. The resulting mixture was filtered and the layers

were separated. To the aqueous solution ascorbic acid was added (to re-

duce any ferricenium ion) and a yellow solution was obtained. The aque-

ous solution was refiltered, washed with diethyl ether (2 x 30 mL), and

then slowly added to a solution of ammonium hexafluorophosphate (5.00 g,

0.0321 mol) in 5 mL of water. The resulting yellow precipitate was

washed with water to remove excess ammonium hexafluorophosphate and with

diethyl ether to remove entrapped water and promote drying. The compound

was recrystallized from acetone and diethyl ether to give 5.91 g (28.3%)

of a fine yellow powder, dec. 224-2280Cc (no lit. m.p. reported).


a.This reagent was used both as solvent and reactant and was present
in large excess with respect to the limiting reagent, ferrocene.
'Aluminum powder was not added to this reaction in order to avoid the
problem of dehalogenation of chlorobenzene and subsequent formation of
n6-benzene-q5-cyclopentadienyliron(II) cation along with the desired
product.
'.Product decomposed in this temperature range over 1.5 hours. The
melting point for the PF6- salt has not been previously reported.









NMR (acetone-d ): cyclopentadienyl, 5.35 (s,5); coordinated arene,

6.40-6.90 (m,5).


IR (KBr): 3131 (s), 3110 (s), 1504 (m), 1445 (s), 1424 (s), 1099 (s),
-1
848 (s, broad), 711 (s), 555 (s), 510 (s), 467 (s, broad) cm .
27
T6-Diphenylmethane-n5-cyclopentadienyliron(ii) Hexafluorophosphate

Ferrocene (19.45 g, 0.104 mol), diphenylmethane (34 mL, 0.183 mol),

aluminum chloride powder (28.36 g, 0.212 mol), aluminum powder (3.10 g,

0.115 mol), and decahydronaphthalene (95 mL) were added to the reaction

flask and stirred at 145-150C for 4 hours. As the mixture was heated

to the reaction temperature, a black oil formed on the sides of the ves-

sel and the organic solution changed from orange to colorless. After 4

hours the reaction mixture'was cooled to 0C and 150 mL of 50% aqueous

methanol was added slowly with concurrent loss of the black oil. The

mixture was filtered and the aqueous layer was separated from the orange

organic layer. The water layer was added slowly to an aqueous solution

of ammonium hexafluorophosphate (7.0 g, 0.044 mol). A yellow precipi-

tate formed immediately and was collected by filtration and dried under
a
vacuum. The product was recrystallized from acetone and diethyl ether

or dichloromethane and diethyl ether to give 17.9 g (38.6%) of a fine

yellow powder, m.p. 136-1380C (dec.) (no lit. m.p. reported).


NMR (acetone-d6): cyclopentadienyl 5.22 (s,5), coordinated arene -

6.47 (s,5), uncoordinated arene 7.37 (s,5), methylene 4.25 (s,2).



a.If the product is not completely dry at this point, oiling will
occur and good recrystallized product is almost impossible to obtain.
If dichloromethane is used to dissolve the salt, any water remaining
is easily removed by using a separatory funnel. The use of dichloro-
methane allows a more rapid work-up as it obviates the tedious step
of completely drying out the product before recrystallization.









IR (KBr): 3128 (m), 1603 (w), 1532 (w), 1498 (m), 1468 (m), 1458 (m),
-1
1425 (s), 845 (vs, broad), 740 (s), 710 (s), 559 (s), 497 (s), 477 (s) cm-

27
16-Fluorene-T5-cyclopentadienyliron(Il) Hexafluorophosphate

Ferrocene (8.0 g, 0.043 mol), fluorene (10.0 g, 0.0603 mol), alumin-

um chloride powder (11.65 g, 0.0875 mol), aluminum powder (3.29 g, 0.122
a
mol), and decahydronaphthalene (40 mL) were added to the reaction ves-

sel and stirred at 1550C. At 1200C a black oil began forming and even-

tually completely coated the sides of the container. After 4 hours the

stirring was stopped and the mixture was cooled to -78C and 20 mL of

10% aqueous methanol was added to the frozen mixture. The mixture was

allowed to warm to 00C and 200 mL of water was added to complete the hy-

drolysis. The resulting orange organic layer and yellow aqueous layer

were filtered and separated. The aqueous layer was added slowly to an

aqueous solution of ammonium hexafluorophosphate containing 3.2 g (0.0201

mol). The immediate yellow precipitate which formed was filtered, washed

with water and dried under vacuum. The product was recrystallized from

acetone and diethyl ether to give 6.34 g (34.1%) of a yellow powdery

product, dec. 163-165C (no lit. m.p. reported).


NMR (acetone-d ): cyclopentadienyl, 4.85 (s,5); coordinated arene, 6.35-

6.63 (m,2), 6.93-7.38 (m,2); uncoordinated arene, 7.38-7.82 (m,3), 8.00-

.8.33 (m,l); exo CH, 4.23 (d,l,J=23Hz); endo CH 4.50 (d,l,J=23Hz).


IR (KBr): 3135 (w), 2950 (broad, w), 1612 (w), 1437 (s), 1422 (s),
-1
1388 (s), 840 (broad, vs), 777 (s), 730 (s), 557 (s), 496 (s), 481 (s) cm


aHexane was used as a solvent in one reaction. The reaction was run
with a similar molar ratio of reactants in refluxing hexane. The pro-
duct was worked up without cooling to -780C and gave a 31.8% yield.








20
n6-Hexamethylbenzene-n5-cyclopentadienyliron( I) Hexafluorophosphate

Ferrocene (10.18 g, 0.0547 mol), hexamethylbenzene (19.18 g, 0.118

mol), aluminum chloride powder (32.0 g, 0.240 mol), aluminum powder (5.00

g, 0.185 mol), and decahydronaphthalene (100 mL) were added to the reac-
o o
tion vessel and stirred at 150-160 C for 4 hours. At 115 C a brown vis-

cous oil appeared on the walls of the container and remained until the
o
reaction mixture was cooled to 0 C and 100 mL of 20% aqueous-methanol was

slowly added. The mixture was filtered and the yellow aqueous layer was

separated from the organic layer. The aqueous layer was then added to

an aqueous solution of ammonium hexafluorophosphate (5.00 g, 0.0321 mol).

The immediately-formed yellow precipitate was filtered, washed with water,

and dried under vacuum. The product was recrystallized from acetone and

diethyl ether to give 14.00 g (59.8%) of a yellow powdery product, dec.

280-2830C (complex originally reported as a BF4- salt).


NMR (acetone-d ): cyclopentadienyl, 4.74 (s,5); methyl, 2.58 (s,18).


IR (KBr): 3014 (vw), 2933 (vw), 1448 (w), 1418 (s), 1382 (s), 1076 (w),
-1
837 (vs, broad), 553 (s), 500 (w), 461 (m) cm

26
n6-Toluene-n5-cyclopentadienyliron(II) Hexafluorophosphate

Ferrocene (19.07 g, 0.103 mol), toluene (150 mL),a aluminum chloride

powder (27.06 g, 0.203 mol), and aluminum powder (5.15 g, 0.191 mol) were

added to the reaction vessel and stirred at 1200C. After 5 hours the re-
o
action mixture was cooled to 0 C and 100 mL of 5% aqueous-methanol was

slowly added. The resulting yellow aqueous and orange organic solutions

were filtered and subsequently separated. The resulting solution was ad-

ded to an aqueous solution of ammonium hexafluorophosphate (7.0 g, 0.0440 mol).



This reagent was used as solvent and reactant and was present in excess.









The immediate yellow precipitate which formed was filtered, washed with

water, and dried under vacuum. The product was recrystallized from di-

chloromethane and diethyl ether to obtain 13.29 g (36.0%) of a fine

yellow powder, dec. 286-2890Ca (lit. m.p. 165C).


'H NMR (acetone-d ): cyclopentadienyl, 5.18 (s,5); coordinated arene,

6.40 (s,5); methyl, 2.58 (s,3).


IR (KBr): 3122 (m), 1530 (w), 1464 (s), 1421 (s), 1393 (s), 1387 (s),
-l
843 (vs, broad), 552 (s), 495 (m), 469 (s, broad) cm-

19
n6-Aniline-rn-cyclopentadienyliron(ll) Hexafluorophosphate9

Ferrocene (34.27 g, 0.184 mol), aniline (30.26 g, 0.163 mol), al-

uminum chloride powder (94.00 g, 0.707 mol),c aluminum powder (15.04 g,

0.557 mol), and decahydronaphthalene (140 mL) were added to the reaction

vessel and stirred at 2000C for 5 hours. During the heating a black oil

coated the sides of the vessel and remained until the system was cooled

to 00C and 200 mL of 10% aqueous methanol was slowly added. The resulting

orange aqueous layer, after filtering, was separated from the orange or-

ganic layer and washed twice with 30 mL portions of diethyl ether. The

resulting solution was slowly added to a solution containing 10.0 g

(0.0629 mol) of ammonium hexafluorophosphate. The immediate yellow-

orange precipitate which formed was filtered, washed with water and di-


a'The melting point capillary was placed inthe bath at 2800C.

b.The aniline was added last via an addition funnel very slowly to
avoid a violent reaction with aluminum chloride.
C'Excess aluminum chloride (approximately four-fold excess with re-
spect to the ferrocene) was used to insure that after the aniline was
added and had reacted with the aluminum chloride there still remained
sufficient aluminum chloride to react with the ferrocene.









ethyl ether, and allowed to dry. The compound was recrystallized from

acetone and diethyl ether to give 21.58 g (36.9%) of a powdery orange

product, dec. 2420Ca (lit. dec. 2500C).


NMR (acetone-d ): cyclopentadienyl, 4.98 (s,5); coordinated arene, 5.80-

6.38 (m,5); NH2, 5.57-5.80 (broad, s,2).


IR (KBr): 3499 (s), 3401 (s), 3248 (w), 3135 (w), 3096 (w), 1634 (s),

1554 (s), 1472 (s), 1421 (s), 1389 (s), 1305 (s), 1162 (s), 845 (vs,
-l
broad), 740 (s), 556 (s), 475 (s), 461 (s) cm-1


Mass Spectrum, 1500C probe temperature (m/e, relative intensity): 372,

1.7; 345, 6.7; 307, 0.7; 280, 5.8; 252, 21.0; 214, 0.4; 213, 2.0; 186,

50.4; 179, 100; 121, 30.6; 107, 34.6; 93, 32.6; 92, 15.0; 91, 39.0;

66, 62.5; 65, 53.2; 23.6, 64; 56, 17.9; 51, 8.2.


n6-Carbazole-n5-cyclopentadienyliron(II) Hexafluorophosphate

Carbazole (9.56 g, 0.0572 mol), ferrocene (5.23 g, 0.028 mol), al-

uminum chloride powder (7.70 g, 0.0579 mol), aluminum powder (2.19 g,

0.081 mol), and decahydronaphthalene (50 mL) were added to the reaction

flask and stirred at 1450C for 4 hours. During the heating to reaction

temperature the sides of the reaction vessel became covered with a black

solid which prevented any observation of solution color. After 4 hours

the system was cooled to 00C and 60 mL of 16% aqueous-methanol solution

was slowly added. The solid dissolved after the subsequent addition of


a.In this particular reaction pure nr-aniline-n5-cyclopentadienyliron(lI)
hexafluorophosphate was obtained. However, in one reaction a mixture
of T-complexes of aniline and tetralin was obtained. These compounds
were easily separated by column chromatography on alumina. After elu-
tion of the r-tetralin complex with dichloromethane, the n-aniline com-
plex was removed with acetone.









100 mL of water and boiling the solution for 5 minutes. After the mix-

ture was filtered, a yellow aqueous layer was separated from an orange

organic layer. The aqueous layer was washed with 75 mL (3 x 25 mL) of

diethyl ether and the resulting aqueous layer slowly poured into an aque-

ous solution containing 2.35 g (0.0144 mol) of ammonium hexafluorophos-

phate. An immediate orange precipitate formed which was filtered,

washed with water, and dried under vacuum. The product was recrystal-

lized from dichloromethane and diethyl ether to give 5.21 g (42.8%) of

an orange powder, dec. 204-2060C.


'H NMR (acetone-d ): cyclopentadienyl, 4.62 (s,5); coordinated arene, 6.07-
6
6.42 (m,2); uncoordinated arene, 8.28-8.50 (m,l), 7.07-7.78 (m,5) ; NH,

10.38 (s,l, broad).

IR (KBr): 3417 (s), 3097 (m), 1627 (s), 1567 (s), 1498 (s), 1443 (s),
-I
1376 (s), 1331 (s), 1249 (s), 838 (s), 759 (s) cm-1

For C7 H 4NFePF6 calculated: C, 47.14%, H, 3.26%; N, 3.23%; Fe, 12.9%.

Found: C, 47.15%; H, 3.63%;N, 3.23%; Fe, 12.9%.


Mass Spectrum, 1500C probe temperature (m/e, relative intensity): 288,

0.1; 186, 50.5; 168, 46.9; 167, 100; 166, 59.4; 139, 37.5; 121, 25.0;

107, 23.8; 85, 43.8; 66, 12.5; 65, 9.4; 56, 18.8.


n6-Diphenylamine-ns-cyclopentadienyliron(II) Hexafluorophosphate

Ferrocene (10.03 g, 0.0549 mol), diphenylamine (18.30 g, 0.108 mol),

aluminum chloride powder (14.20 g, 0.107 mol), aluminum powder (4.34 g,

0.160 mol), and decahydronaphthalene (100 mL) were added to the reaction


a.overlapping multiplets of 2 coordinated and 3 undoordinated arene
protons.









vessel and stirred at 155-1800C. During the course of the reaction

a gray-green solid formation formed on the walls of the vessel. After

3.5 hours the mixture was cooled to 00C and 80 mL of 12% aqueous methanol

was slowly added. The resulting mixture was filtered and the layers

were separated. The orange, aqueous layer was added to a solution con-

taining ammonium hexafluorophosphate (4.00 g, 0.0188 mol) in 10 mL of

water. An orange precipitate formed immediately and was filtered,

washed with water, and dried under vacuum. To increase the yield, the

organic layer was washed with water (2 x 30 mL) and the washings were

added to a solution of ammonium hexafluorophosphate (1.00 g, 0.0063 mol)

in water (5 mL). The orange precipitate which formed was filtered,

washed with water, and dried under vacuum. The two products were com-

bined, dissolved in 60 mL of dichloromethane, and recrystallized by slow

addition of the solution to 250 mL of diethyl ether. A yield of 1.37 g

(5.8%) of orange needles, m.p. 180-181C, was obtained after filtering

and drying.


NMR (acetone-d6): cyclopentadienyl, 4.95 (s,5); coordinated arene, 5.85-

6.35 (m,5); uncoordinated arene, 7.02-7.62 (m,5); NH, 7.87 (s,l,broad).


IR (KBr): 3389 (s), 3117 (w), 1548 (s), 1469 (m), 1386 (s), 839 (s,
-1
broad), 557 (s), 467 (m) cm


Calculated for C H NFePF : C, 46.92%; H, 3.71%; N, 3.22%, Fe, 12.8%.
17 16 6
Found: C, 47.16%;H,3.88%;N, 3.16%; Fe, 12.2%


Mass Spectrum, 150C probe temperature (m/e, relative intensity): 290,

>0.1; 186, 4.8; 170, 13.1; 169, 100; 168, 75.0; 167, 43.2; 121, 3.5;

83, 26.9; 66, 15.5; 65, 11.6; 51, 21.1.









n6-1,3,5-Triisopropylbenzene-n5-cyclopentadienyliron(lI) Hexafluorophosphate

Ferrocene (9.00 g, 0.0484 mol), 1,3,5-triisopropylbenzene (20.31 g,

0.0996 mol), aluminum chloride powder (23.00 g, 0.173 mol), aluminum pow-

der (1.97 g, 0.0744 mol), and decahydronaphthalene (75 mL) were added

to the reaction vessel and stirred at 155 C. During the reaction a

brown oil formed on the sides of the vessel. After 4.5 hours the reac-

tion was stopped by cooling the mixture to 0 C and slowly adding 150 mL

of 20% aqueous methanol. A green aqueous layer and an orange organic

layer were formed. The mixture was filtered and the aqueous layer was

separated from the organic layer and slowly added to an aqueous solution

of ammonium hexafluorophosphate (5.00 g, 0.0321 mol). A yellow precipi-

tate formed immediately and was filtered, washed with water and dried

under vacuum. The product was recrystallized from acetone and diethyl

ether to give 10.41 g (45.8%) of a fine, yellow powder, dec. 264-265 C.


'H NMR (acetone-d6); cyclopentadienyl, 5.07 (s,5); coordinated arene, 6.27

(s,3); methyl, 1.42 (d,18, J = 6.5 Hz); methine, 3.22 (m,3, J = 6.5 Hz).

IR (KBr): 3125 (w, broad), 2967 (vs), 2940 (s), 2882 (m), 1624 (w),

1529 (s), 1466 (s), 1421 (s), 1384 (s), 844 (vs, broad), 554 (s), 489
-I
(w), 452 (w) cm-1


Calculated for C20H 29FePF6: C, 51.08%; H, 6.22%; Fe, 11.88%.

Found: C, 51.16%; H, 6.26%; Fe, 11.6%.


a'The sample was placed in the apparatus at 2520C and the yellow com-
pound turned orange immediately. There was a slow loss of color, with
concomitant oil vaporization until 2640C, at which time rapid decom-
position occurred.







Mass Spectrum, 1400C, probe temperature (m/e, relative intensity): 325,

0.5; 204, 26.3; 190, 15.6; 189, 100; 186, 26.4; 161, 58.1; 147, 12.0;

133, 14.1; 121, 15.1; 119, 15.2; 107, 25.7; 105, 25.9; 91, 36.7; 66,

11.1; 65, 13.3; 56, 12.0.


q6-Thiophenol-15-cyclopentadienyliron(lI) Hexafluorophosphate

n6-Chlorobenzene-nS-cyclopentadienyliron(II) hexafluorophosphate

(2.32 g, 0.00612 mol), sodium hydrogen sulfide (2.16 g, 0.047 mol), and

acetonitrile (20 mL) were added to a dry 100 mL Schlenk tube against a

countercurrent of nitrogen. Upon stirring an immediate reaction took

place as evidenced by the formation of an intense red solution. After

20 minutes the solution was filtered and the blue-green residue was

washed with four 30 mL portions of acetonitrile and filtered. The com-

bined filtrates were taken to dryness under vacuum to give an intense red

solid. An alumina column was prepared by dry packing the column and evacu-

ating the resulting column to remove any entrapped oxygen. The product

from the above reaction was dissolved in 100 mL acetone and then added

to the column against a countercurrent of nitrogen. A yellow band eluted

with acetone was discarded. An orange band remained on the column and was

eluted with 30% aqueous acetone. Acidification of this solution with con-

centrated hydrochloric acid resulted in a color change from red-orange to

yellow. Excess ammonium hexafluorophosphate was added and the yellow solu-

tion was taken to dryness under vacuum. The residue was recrystallized from

acetone and diethyl ether and then dichloromethane and diethyl ether (both

under nitrogen) to give 1.16 g (50.4%) of an olive green product, dec. 236-

2380C.


'H NMR (acetone-d ): cyclopentadienyl, 5.20 (s,5); coordinated arene,

6.28-6.78 (m,5); SH 3.6-3.92 (s, broad, 1).


IR (KBr): 3124 (m), 2584 (w), 1625 (w), 1505 (w), 1451 (m), 1421 (s),









1408 (m), 1385 (m), 1099 (m), 1011 (w), 835 (vs, broad), 703 (w), 552
-1
(s), 503 (m), 467 (s, broad ) cm


Calculated for C11 H SFePF6: C, 35.13%; H, 2.95%; Fe, 14.85%.

Found: C, 35.25%; H, 2.96%; Fe, 14.41%.


Mass spectrum, 140C probe temperature (m/e, relative intensity): 231,

0.2; 230, 1.5; 187, 9.7; 186, 72.4; 185, 10.9; 184, 10.3; 121, 39.8; 110,

93.4; 109, 36.4; 107, 100; 104, 39.9; 88, 27.5; 85, 34.2; 84, 19.6; 78,

11.3; 77, 21.3; 69, 25.1; 66, 76.6; 65, 49.2; 56, 22.5; 51, 21.2.


n6-Phenol-n5-cyclopentadienyliron(II) Hexafluorophosphate

n6-Chlorobenzene-n5-cyclopentadienyliron(II) hexafluorophosphate

(4.54g, 0.0120 mol) and sodium hydroxide (5.00 g, 0.125 mol) were added

to a 100 mL round bottom flask that had been previously flushed with N .
2
After the addition of 50% aqueous acetone (80 mL), the reaction mixture

was stirred under nitrogen. During the reaction two layers formed, an

orange aqueous layer and an intense brown organic layer. After 36 hours

the organic layer was evaporated under vacuum without prior separation.

With the evaporation of the organic layer substantial amounts of solid

formed in the flask. The residual slurry was filtered to give an orange

filtrate and a yellow solid which was washed with water. Acidication of

the filtrate with concentrated hydrochloric acid produced a color change

from orange to yellow and concomitant formation of a small amount of yel-

low solid. This mixture was then extracted with five 40 mL portions of

dichloromethane. The fractions were combined and blown dry under N

A brown, wet product was obtained which was dissolved in acetone, fil-

tered, and precipitated by the addition of diethyl ether to give golden

needles (2.24 g). Further purification was achieved by dissolving the









product in acetone and chromatographing it on an alumina column. A

yellow band moved rapidly down the column with acetone and was discarded.

An orange band remained at the top of the column. This was eluted with

a 30% aqueous acetone solution and collected when it came off the column.

The solution was acidified with concentrated HCI and then taken to dry-

ness under vacuum by heating to 450C. The resulting yellow product was

dissolved in dichloromethane and precipitated with diethyl ether to

give 1.65 g (36.3%) of a fine yellow powder, dec. 204-206C.


'H NMR (acetone-d ): cyclopentadienyl, 5.12 (s,5); coordinated arene,

6.28 (s, broad, 5); OH, 9.05 (s, broad, 1).


IR (KBr): 3495 (s, broad), 3121 (s, broad), 1548 (s), 1518 (m), 1469

(m), 1444 (s), 1420 (m), 1276 (m), 1218 (m, broad), 1833 (vs, broad),
-1
551 (s), 523 (m), 476 (s), 452 (m) cm


Calculated for C H OFePF : C, 36.70%; H, 3.08%; Fe, 15.51%.
11 11 6
Found: C, 36.95%; H, 3.15%; Fe, 15.55%.


Mass Spectrum, 110C probe temperature (m/e, relative intensity): 252,

9.4; 232, 3.8; 215, 2.0; 214, 10.8; 200, 7.2; 186, 71.8; 178, 17.2; 121,

47.4; 107, 59.9; 94, 100; 77, 34.2; 76, 25.8; 66, 41.6; 65, 39.7; 56,

25.5.


n6-N,N-dimethylanil ine-n5-cyclopentadienyliron(Il) Hexafluorophosphate

n6-Aniline-n5-cyclopentadienyliron(II) hexafluorophosphate (1.05 g,

0.00292 mol) and sodium amide (1.49 g, 0.0382 mol) were added to a dry,

nitrogen filled 250 mL round bottom flask against a countercurrent of

nitrogen. The addition of dichloromethane (100 mL) produced a red solu-

tion upon stirring. After 1.75 hours the solution was filtered and methyl








iodide (1.5 mL) was added. There was an initial formation of a small

amount of yellow precipitate followed by a slower change in the color

of the solution from red to yellow. The resulting solution was taken to

dryness under vacuum to give an orange oil. The oil could not be crys-

tallized by trituration with pentane, benzene, or diethyl ether. Further-

more, recrystallization using chloroform and pentane produced only an

oil. The product was obtained by the addition of 0.1M aqueous silver

nitrate to precipitate any silver iodide. There was an immediate forma-

tion of a white solid with a concomitant yellow solution. The solution

was filtered and ammonium hexafluorophosphate (0.50 g, 0.0031 mol) added

to obtain an orange-yellow precipitate. This was extracted with dichloro-

methane (20 mL) and precipitated with pentane to obtain 0.38 g

(29.7%) of an impure brown solid. Purification of the product was ac-

complished by dry packing the product on a 1 x 20 cm column of silica

gel. The product was eluted as a broad yellow band with 250 mL of di-

chloromethane. This was blown dry under nitrogen to give orange needles.

The orange needles were recrystallized from dichloromethane and pentane

to give a flocculant orange powder, m.p. 153-156.50C.


'H NMR (acetone-d ): cyclopentadienyl, 5.11 (s,5); coordinated arene,

6.37-6.08 (m,3), 6.08-5.83 (m,2); methyl, 3.21 (s,6).


IR (KBr); 3126 (m), 2925 (w), 1564 (s), 1493 (w), 1447 (m), 1431 (m),

1367 (m), 1235 (m), 1193 (m), 1005 (w), 830 (s, broad), 656 (m), 556 (s),
-l
475 (m), 463 (m) cm-1


For C 3H 6FeNPF6 calculated: C, 40.31%; H, 4.13%; Fe, 14.5%.

Found: C, 40.23%; H, 4.11%; Fe, 14.7%.









Mass Spectrum, 110C probe temperature (m/e, relative intensity): 239,

15.2; 228, 0.2; 227, 1.8; 187, 13.5; 186, 100; 184, 10.9; 121, 99.4;

120, 91.9; 107, 56.5; 77, 17.5; 56, 8.4.


n6-Thioanisole-i5-cyclopentadienyliron(II) Iodide

To a dry, nitrogen filled Schlenk recrystallization tube (1-5-n5-

cyclohexadienyl-6-thione)-n5-cyclopentadienyliron(Il) (0.51 g, 0.0022 mol)

was added against a countercurrent of nitrogen. The complex was dissolved

in chloroform (30 mL), filtered, and then reduced in volume under vacuum

to 15 mL. To this solution methyl iodide (0.8 mL, 0.0084 mol) was added

and stirred, producing a slow change in color from red to orange. After

30 minutes diethyl ether (80 mL) was added to produce an initial orange

oil which slowly crystallized. After sitting for 1 hour, the yellow-

orange solid was filtered and dried under vacuum to give 0.58 g (70.8%),

dec. 119-1220C.


'H NMR (chloroform-d): cyclopentadienyl, 5.28 (s,5); coordinated arene,

6.65 (s,5); methyl, 2.81 (s,3).


IR (KBr): 3073 (m), 3035 (m), 1630 (w), 1503 (m), 1444 (s), 1418 (s),

1387 (m), 1159 (m), 1146 (w), 1089 (s), 1006 (m), 979 (w), 854 (s),
-1
.689 (m), 500 (m), 469 (s, broad) cm

Calculated for C H FeSI: C, 38.74%; H, 3.52%; Fe, 15.0%.
12 13
Found: C, 38.65%; H, 3.42%; Fe, 15.2%.


Mass Spectrum, 1200C probe temperature (m/e, relative intensity): 187,

13.0; 186, 100; 124, 99.2; 123, 20.5; 121, 44.9; 109, 42.5; 91, 29.2;

-65, 20.3; 56, 24.3.









n6-Anisole-n5-cyclopentadienyliron( I) Iodide

The title compound was prepared by treating the filtrate obtained

from recrystallizing a sample of (l-5-n5-cyclohexadienyl-6-one)-r5-cyclo-

pentadienyliron(ll) from dichloromethane and diethyl ether. Methyl iodide

(1 mL, 0.017 mol) was added and the solution was stirred for 1 hour.

The red solution slowly changed to a cloudy orange solution during this

period. The solvent was then evaporated under a stream of nitrogen and

the resulting orange residue was recrystallized from chloroform and di-

ethyl ether, dec. 143-146C.


'H NMR (chloroform-d): cyclopentadienyl, 5.25 (s,5); coordinated arene,

6.50 (s,5); methyl, 4.08 (s,3).


n6-Carbazolyl-n5-cyclopentadienyliron(II)

n6-Carbazole-n5-cyclopentadienyliron(II) hexafluorophosphate (0.89 g,

0.00205 mol) and sodium amide (0.84 g, 0.0215 mol) were added to a dry,

nitrogen filled 100 mL Schlenk tube. Ammonia (90 mL) was condensed into

the system to produce an immediate reaction as seen by the formation of

a dark red solution. After stirring for 3 hours the ammonia was evapor-

ated and the resulting red solid was extracted with four 90 mL portions

of benzene. These extracts were filtered and dried under vacuum to give

0.70 g (90%) of a red solid. The product was recrystallized from benzene

and diethyl ether at O0C to give a fine red powder, dec. 176-1780C.


'H NMR (acetone-d ): cyclopentadienyl, 4.13 (s,5); coordinated arene,

6.93-7.18 (m,2); 5.50-5.82 (m,2); uncoordinated arene, 8.07-8.27 (m,l),

7.27-7.80 (m,2), 6.65-6.93 (m,l).


IR (KBr): 1613 (m), 1544 (m), 1477 (m), 1434 (s), 1389 (s), 1330 (s),

1287 (m), 1234 (s), 1007 (m), 844 (s), 757 (s), 559 (m), 428 (s) cm-1









Calculated for C7 H Fe: Fe, 19.4%.

Found: Fe, 19.8%.


Mass Spectrum, 200C probe temperature (m/e, relative intensity): 287,

0.1; 186, 17.9; 168, 13.4; 167, 100; 166, 22.4; 139, 11.1; 121, 8.9; 107,
12 1 14 56
13.1; 83.5, 11.4. Parent peak exact mass calculated for C H N Fe,
17 13
287.0396; found, 287.0373.


n6-Fluorenyl-n5-cyclopentadienyliron(II)

n6-Fluorene-n5-cyclopentadienyliron(II) hexafluorophosphate (0.98 g,

0.00226 mol) and sodium hexamethyldisilylamide (1.29 g, 0.00705 mol) were

added to a dry, nitrogen filled 100 mL Schlenk tube. Dry diethyl ether

(70 mL) was added and led to an immediate reaction as seen by the forma-

tion of a blue-green solution. The solution was stirred for 1.5 hours,

then filtered and taken to dryness under vacuum. A deep green residue

weighing 0.27 g (43%) was obtained. The compound was recrystallized from

diethyl ether and pentane at -78C, dec. 105-1070C.


'H NMR (acetone-d ): cyclopentadienyl, 4.02 (s,5); coordinated arene,
6
6.75-7.13 (m,2)a, 6.38-6.57 (m,l), 5.27-5.72 (m,2); uncoordinated arene,

8.08-8.33 (m,1), 7.13-7.68 (m,2); vinyl CH, 5.75 (s,l).


IR (KBr): 1602 (s), 1532 (s), 1470 (s), 1386 (s), 1338 (s), 1226 (s),
-1
1005 (m), 835 (s), 752 (s), 472 (s) cm


Calculated for C H 4Fe: Fe, 19.5%.
18 14
Found: Fe, 19.8%.


a.
overlapping multiplets of one coordinated and one uncoordinated
arene ring proton.









Mass Spectrum, 1400C probe temperature (m/e, relative intensity): 286,

1.0; 186, 69.8; 166,49.7; 165, 100; 121, 33.3. Parent peak exact mass

calculated for 1218 H 56Fe, 286.0444; found, 286.0435.


(1-5-n56-Diphenylmethylenecyclohexadienyl)-n5-cyclopentadienyliron(II)

n6-Triphenylmethane-n5-cyclopentadienyliron(l ) hexafluorophosphate

(0.54 g, 0.00103 mol) and sodium amide (0.40 g, 0.0106 mol) were added

to a dry, N filled, 100 mL Schlenk tube. Ammonia (90 mL) was condensed
2
into the Schlenk tube to produce an immediate reaction as seen by the

formation of an intense red solution. The solution was stirred for 4

hours with no further visible changes. The ammonia was then allowed to

evaporate leaving a red solid. The solid was extracted with three 90 mL

portions of diethyl ether, filtered, and then dried under vacuum to

yield 0.34 g (89%) of a brick-red product. The product was recrystal-

lized at -780C using ether and pentane to give a red powder, dec. 115-

117C.


'H NMR (acetone-d ): cyclopentadienyl, 4.55 (s,5); coordinated arene,

4.12-4.38 (m,2); 4.82-5.11 (m,2), 5.94-6.25 (m,l); uncoordinated arene,

6.88-7.46 (m,10).


SIR(KBr): 3058 (m), 1588 (s), 1423 (s), 1270 (s), 837 (s), 758 (s), 697
-1
(s), 598 (s), 468 (m), 303 (s) cm


Calculated for C H Fe: Fe, 15.4%.
24 20
Found: Fe, 15.2%


Mass Spectrum, 1300C probe temperature (m/e, relative intensity): 364,

93.0; 298, 23.3; 243, 60.4; 186, 21.5; 165, 100; 121, 39.5. Parent peak
12 1 56
exact mass calculated for C H Fe, 364.0910; found, 364.0920.
24 20








(1-5- -6-Phenyliminocyclohexadienyl)-n5-cyclopentadienyliron(II)

n6-Diphenylamine-n5-cyclopentadienyliron(lI) hexafluorophosphate

(0.56 g, 0.00128 mol) and sodium amide (0.75 g, 0.0192 mol) were added

to a dry, nitrogen filled, 100 mL Schlenk tube. Ammonia (90 mL) was

condensed into the Schlenk tube to produce an immediate reaction as seen

by the formation of an intense red solution. The solution was stirred

for 3 hours with no further visible change and then the ammonia was al-

lowed to evaporate leaving a red-orange solid. This was extracted with

two 90 mL portions of benzene, filtered, and dried under vacuum. The

resulting product was recrystallized from benzene and pentane at 00C

to give 0.31 g (85%) of a red powder, dec. 122-1240C.

'H NMR (acetone-d ): cyclopentadienyl, 4.48 (s,5); coordinated arene,

4.58-4.82 (m, 2-under cyclopentadienyl peak), 5.17-5.55 (m,2), 5.57-5.83

(m,l); undoordinated arene, 6.72-7.00 (m,3); 7.05-7.42 (m,2).

IR(KBr): 3066 (m), 1596 (m), 1562 (s), 1494 (s), 1475 (s), 1421 (m),

1386 (s), 1217 (m), 814 (m), 764 (m), 717 (m), 667 (m), 498 (m), 484 (m),
-l
466 (m) cm1


Calculated for C H NFe: Fe, 19.3%.
17 15
Found: Fe, 19.3%.

Mass Spectrum (no heat) (m/e, relative intensity): 289, 1.4; 186, 31.5;

169, 100; 168, 55.9; 167, 37.3; 121, 15.0. Parent peak mass calculated
12 1 14 56
for C H N Fe, 289.0552; found, 289.0552.
17 15

(1-5-n5-Cyclohexadienyl-6-one)-n5-cyclopentadienyllron(ll)

n6-Chlorobenzene-n5-cyclopentadienyliron(l I) hexafluorophosphate

(5.41 g, 0.0135 mol) and potassium hydroxide (4.18 g, 0.0745 mol) were









added to a nitrogen-flushed 250 mL round bottom flask. To the reactants

120 mL of 50% aqueous acetone was added to produce a yellow-brown solu-

tion which slowly turned orange-brown over 1 hour. After the reaction

mixture was stirred for 18 hours under nitrogen, the acetone was re-

moved by evaporation under vacuum. The mixture was rapidly filtered in

air to remove insoluble matter. An orange solution was obtained which

was pumped to dryness under vacuum at room temperature to give a damp

brown solid. The solid was extracted with two 100 mL portions of dichloro-

methane, filtered, and then taken to dryness under vacuum. The resulting

orange solid was recrystallized from dichloromethane and ether to give

2.39 g (82.7%) of an orange powder, dec. 112-114C.


'H NMR (acetone-d ): cyclopentadienyl, 4.57 (s,5), coordinated arene,

4.67-4.97 (m,2), 5.33-5.63 (m,3).


IR (KBr): 3060 (w), 1661 (m), 1535 (s, broad), 1471 (s), 1417 (m),

1387 (m), 1346 (s), 1142 (m), 1112 (m), 1047 (w), 1005 (m), 841 (s),
-l
689 (s), 542 (s), 476 (s) cm .


Calculated for C H FeO: Fe, 26.1%.
II 10
Found: Fe, 26.6%.


Mass Spectrum, 110C probe temperature (m/e, relative intensity): 214,

36.2; 186, 49.4; 121, 36.3; 94, 100; 66, 24.3; 65, 28.7; 56, 19.3; 51,
12 1 56 16
85. Parent peak exact mass calculated for C H Fe 0, 214.00800;
11 10
found, 214.00791.


(l-5-ns-Cyclohexadienyl-6-thione)-T5-cyclopentadienyliron(II)

n6-Chlorobenzene-ns-cyclopentadienyliron(II) hexafluorophosphate

(7.07 g, 0.0187 mol), sodium hydrogen sulfide (4.19 g, 0.0748 mol), and









acetonitrile (70 mL) were added to a dry 100 mL Schlenk tube against a

countercurrent of nitrogen. Upon stirring, an intense red solution de-

veloped. After 20 minutes the solution was filtered and the remaining

blue-green solid was washed with ten 20 mL portions of acetonitrile.

The combined filtrates were reduced in volume to approximately 150 mL

and 100 g of alumina was added against a countercurrent of nitrogen.

The mixture was then taken to dryness under vacuum. This alumina was

then added, against a countercurrent of nitrogen, to a 5 x 40 cm column

previously packed with alumina dried under vacuum. A small yellow band

was eluted with acetone and was discarded. The desired product was eluted

as a broad red-orange band using 20% aqueous acetone. The red solution

was dried under vacuum to give a red-brown product. To remove the last

traces of water it was necessary to heat to 450C. This product was re-

crystallized from dry dichloromethane and dry diethyl ether. The resulting

red-orange powder was dried under vacuum to give 2.38 g (55.3%). Darkens

at 125-1280C. Decomposes at 1830C.


'H NMR (acetone-d6): cyclopentadienyl, 5.02 (s,5); coordinated arene,

6.00-6.37 (s, broad, 3), 6.37-6.84 (s, broad, 2).


'H NMR (chloroform-d): cyclopentadienyl, 4.70 (s,5); coordinated arene,

5.37-6.00 (s, broad, 3), 6.00-6.34 (s, broad, 2).


IR (KBr): 3025 (w), 1628 (w), 1487 (s), 1418 (s), 1399 (m), 1383 (m),
-l
1082 (s), 842 (s), 711 (m), 641 (m), 463 (s) cm .


Calculated for Cl H FeS: Fe, 24.3%.

Found: Fe, 24.0%.








Mass Spectrum, 100*C probe temperature (m/e, relative intensity): 230,

3.5; 186 (C H S), 85.1; 186 (C H Fe), 100; 185 (C H S), 56.0; 184
12 10 10 10 12 9
(C 2H S), 36.1; 154, 20.5; 152, 10.0; 121, 31.4; 1l10, 15.9; 77, 14.2;

66, 52.1; 65, 31.5. Parent peak exact mass calculated for 12C 1H
56 32 11 10
5Fe 32S, 229.98520; found, 229.98663.


(l-5-ns-6-lminocyclohexadienyl)-ns-cyclopentadienyliron(II)

a. Reaction in Dichloromethane

n6-Aniline-n5-cyclopentadienyliron(II) hexafluorophosphate (1.05 g,

0,00292 mol), sodium amide (1.34 g, 0.0344 mol), and dichloromethane

(90 mL) were added to a dry, 100 mL Schlenk tube against a countercurrent

of nitrogen. Stirring this solution produced an intense red color.

After 1.75 hours the solution was filtered and taken to dryness under

vacuum. The residue was recrystallized from dichloromethane and ether

to give 0.49 g (79.1%) of a red solid, dec. 140-1460C.


'H NMR (acetonitrile-d ): cyclopentadienyl, 4.53 (s,5); coordinated
3
arene, 5.30-5.63 (m,3), 4.83-5.05(m,2); NH, not observed.


IR (KBr): 3458 (m), 3400 (m), 3100 (m), 1631 (m), 1551 (s), 1469 (s),

1469 (m), 1342 (m), 1149 (w), 1005 (w), 838 (s, broad), 661 (m), 557 (s),
-1
472 (m, broad) cm


Mass Spectrum, 80C probe temperature (m/e, relative intensity): 372,

0.1; 307, >0.1; 213, 4.5; 186, 35.7; 121, 28.8; 93, 100; 92, 12.1; 66,

89.6; 65, 50.5; 56, 16.3. Exact mass calculated for 12C20 H20 Fe2'

372.02620; found, 372.02633. Exact mass calculated for 12C15 H15 56Fe2'

306.9871; found, 306.9894. Parent peak exact mass calculated for
121 H 14N 56Fe,213.02390;found,213.02352.
C H N6Fe, 213.02390; found, 213.02352.









b. Reaction in Ammonia

n6-Aniline-n5-cyclopentadienyliron(ll) hexafluorophosphate (2.04 g,

0.00568 mol) and sodium amide (4.64 g, 0.119 mol) were added to a dry,

nitrogen filled 100 mL Schlenk tube against a countercurrent of nitrogen.

Ammonia (90mL) was condensed into the system to produce an initial in-

tense red solution which turned black. After 3 hours the ammonia was

evaporated to leave a tacky black solid. Extraction of this solid with

dichloromethane produced only a pale yellow solution.


c. Reaction in Tetrahydrofuran

n6-Aniline-n5-cyclopentadienyliron(ll) hexafluorophosphate (1.09 g,

0.00301 mol) and sodium amide (2.00 g, 0.0513 mol) were added to a dry,

nitrogen filled, 100 mL.Schlenk tube against a countercurrent of nitro-

gen. Tetrahydrofuran (90 mL) was added to produce an intense red solu-

tion which then turned black in less than ten minutes. The sample was

not worked up.


d. Reaction in Acetonitrile

Acetonitrile (80 mL) was added to a dry, nitrogen filled, 250 mL

round bottom flask against a countercurrent of nitrogen. To the aceto-

nitrile sodium metal (0.90 g, 0.039 mol) was slowly added to produce

rapid bubbling and a white precipitate. After 10 minutes the bubbling

stopped and a pale yellow solution was noted. r6-Aniline-r5-cyclopenta-

dienyliron(ll) hexafluorophosphate (2.53 g, 0.00707 mol) was added to

the mixture resulting in the formation of an intense red solution.

After stirring for 3 hours, the solution was filtered and reduced in vol-

ume to 10 mL under vacuum. Diethyl ether (300 mL) was added to obtain

a red-orange precipitate. This was filtered and dried under vacuum and

was identical with the compound prepared in CH2Cl2.









Reaction of n6-Diphenylmethane-n5-cyclopentadienyliron(lI) Hexafluoro-
phosphate with Sodium Amide

n6-Diphenylmethane-n5-cyclopentadienyliron(l ) hexafluorophosphate

(1.06 g, 0.00245 mol) and sodium amide (1.07 g, 0.0274 mol) were added to

a dry, nitrogen-filled, 100 mL Schlenk tube against a countercurrent of

nitrogen. The system was sealed and ammonia (90 mL) was condensed into

the system to give an intense red solution. After stirring for 2 hours,

the ammonia was evaporated and the resulting red residue was dried under

vacuum. The mixture was then extracted with four 100 mL portions of

diethyl ether. These were filtered, combined and taken to dryness to

give 0.25 g (36%) of a red-brown powder. The product was recrystallized

at -78C from diethyl ether and pentane, dec. 66-680C.


'H NMR (acetone-d ): cyclopentadienyl, 4.00 (s,2), 4.13 (m,3); arene,

7.27 (s,10); methine, 3.83 (s,l).


IR (KBr): 3021 (w), 2903 (vw), 1597 (m), 1576 (m), 1490 (m), 1448 (m),

1411 (m), 1383 (s), 1259 (m), 1105 (s), 998 (m), 804 (s), 698 (s), 470 (m)
-1
cm


Regeneration of 16-Diphenylmethane-n5-cyclopentadienyliron(l I) Cation

A sample of the deprotonated r-diphenylmethane complex was prepared

as described previously. To the filtered red solution of the complex in

benzene 4 M hydrochloric acid (4 mL) was added. There was an immediate

reaction to give a yellow solution. This was filtered and the product

was precipitated by the addition of ammonium hexafluorophosphate. The

resulting yellow precipitate was filtered and dried. The 'H NMR spectrum

of this compound was identical to that of the starting material.








Attempted Reaction of Methyl Iodide with Certain t-Arene-T-cyclopenta-
dienyliron(ll) Salts

a. n6-Aniline-r5-cyclopentadienyliron( I) Hexafluorophosphate

n6-Aniline-n5-cyclopentadienyliron(II) hexafluorophosphate (0.30 g,

(0.00084 mol) and methyl iodide (0.46 g, 0.0032 mol) were stirred to-

gether for 30 minutes in dichloromethane (15 mL). The solvent was then

removed under vacuum and 'H NMR spectral analysis of the resulting

showed it to be exclusively the starting cation.


b. T6-Phenol-n5-cyclopentadienyliron(ll) Hexafluorophosphate

n6-Phenol-ns-cyclopentadienyliron(II) hexafluorophosphate (0.10 g,

0.00028 mol) was stirred with methyl iodide (2.0 g, 0.014 mol) in dichloro-

methane (5 mL) for 40 minutes. The solvent was then removed under vacuum

and 'H NMR spectral analysis of the resulting residue showed it to be

exclusively the starting cation.


c. n6-Thiophenol-n5-cyclopentadienyliron(I I) Hexafluorophosphate

n6-Thiophenol-n5-cyclopentadienyliron(l I) Hexafluorophosphate (0.10 g,

0.00027 mol) was stirred with methyl iodide (2.0 g, 0.014 mol) in chloro-

form (5 mL) for 40 minutes. The solvent was then removed under vacuum

and 'H NMR spectral analysis of the resulting residue showed it to be

exclusively the starting cation.


Preparation of _6-anilinium-n5-cyclopentadienyliron(ll) Cation

a. From n6-Aniline-n5-cyclopentadienyliron(II) Hexafluorophosphate

n6-Aniline-n5-cyclopentadienyliron( I) hexafluorophosphate (approx-

imately 0.05 g) was dissolved in acetone-d6 (0.50 mL) and filtered into

a 'H NMR tube. Acetic acid-d (0.2 mL) was added to the solution and the

'H NMR spectrum of the sample was obtained.








'H NMR (acetone-d ): cyclopentadienyl, 4.85 (s,5, broad); coordinated

arene, 5.85 (s,5, broad); N-H, 9.40 (s,2, broad).


b. From (1-5-15-6-lminocyclohexadienyl)-n5-cyclopentadienyliron(II)

(l-5-n5-6-Iminocyclohexadienyl)-n5-cyclopentadienyliron(II) (approx-

imately 0.10 g) was dissolved in acetone-d6 (0.50 mL) and filtered into

a 'H NMR tube. Acetic acid-d (0.2 mL) was added to produce an immediate

change from a red to an orange solution. A 'H NMR spectrum was obtained

of this solution.


'H NMR (acetone-d ): cyclopentadienyl, 4.85 (s,5, broad); coordinated

arene, 5.85 (s,5, broad); N-H, 9.90 (s,l, broad).


Preparation of n6-Benzene-ns-cyclopentadienyliron(lI) Hexafluorophosphate
in Ethylpyridinium Bromide-Aluminum Chloride Eutectic

Ferrocene (10.0 g, 0.0538 mol), aluminum powder (2.00 g, 0.074 mol),

and benzene (100 mL) were added to a 250 mL 3-neck round bottom flask

equipped with a mechanical stirrer and a condenser. The flask had been

previously charged with approximately 100 mL of the ethylpyridinium
a
bromide-aluminum chloride eutectic. The brown solution was heated with

stirring at 80C and as the temperature approached 80C, the solution be-

came black. After 3 hours the system was cooled and then rapidly poured

over 200 g of ice. A vigorous exothermic reaction occurred and was fol-

lowed by the formation of an orange organic layer and a yellow aqueous



a'Koch, Miller, and Osteryoung31 reported that the preparation of the
eutectic was accomplished by the addition of ethylpyridinium bromide
to aluminum chloride in a 1:2 mole ratio. For the preparation of the
eutectic in this experiment ethylpyridinium bromide (59 g, 0.31 mol)
was carefully mixed with aluminum chloride (65 g, 0.49 mol) to produce
a brown liquid. The mole ratio was considerably less than that reported
by Koch et al. It was found that addition of the full two equivalents
of aluminum chloride resulted in the solidification of the eutectic.









layer. After filtration of the mixture, the aqueous layer was separated

from the organic layer and washed with two 40 mL portions of diethyl

ether. The aqueous solution was then slowly added to an aqueous solution

of ammonium hexafluorophosphate (4.0 g, 0.025 mol). A yellow precipitate

formed immediately and was filtered and dried under vacuum. The product

was recrystallized from acetone and diethyl ether to give 14.7 g of a

pale yellow solid. 'H NMR showed the solid to be a mixture of the ex-

pected product and ethylpyridinium bromide. The mixture could not be

separated by recrystallization from acetone, dichloromethane, or acetoni-

trile. Chromatographic separation of a column prepared with either al-

umina or silica gel was not possible either. Under no conditions could

any discernable difference in R values be discerned.


Reaction of Hydroiodic Acid with n6-Anisole-n5-cyclopentadienyliron(ll)
Hexafluorophosphate

n6-Anisole-n5-cyclopentadienyliron(ll) hexafluorophosphate (0.35 g,

0.00094 mol) and hydroiodic acid (20 mL, 57%) were added to a 50 mL

nitrogen-flushed, round bottom flask. The mixture was heated to 150C

for 45 minutes. It was then cooled and neutralized with solid sodium

carbonate and the resulting yellow solid was collected. A 'H NMR spec-

trum showed the solid to contain only starting material. In a similar

reaction in which the mixture was kept at 150C for 16 hours decomposi-

tion occurred as evidenced by the formation of a brown solid. When the

system was worked up in a fashion similar to that above, neither the

starting material nor the desired T-phenol complex was obtained.


Attempted Synthesis of n6-Phenol-n5-cyclopentadienyliron(ll) Hexafluoro-
phosphate

Ferrocene (10.33 g, 0.0555 mol), phenol (50.5 g, 0.537 mol), aluminum









chloride powder (15.17 g, 0.114 mol), and aluminum powder (4.53 g, 0.168

mol) were added to the reaction flask against a countercurrent of nitro-

gen. A vigorous reaction occurred immediately upon addition of the re-

gents to produce a green, grainy solution. Heating the mixture to 1000C

caused continued bubbling and the solution became a black viscous oil.

After 1 hour the reaction was cooled to 0C and 100 mL of water was

added. A black, oily organic layer and a colorless aqueous layer was

obtained. No further workup was attempted.


Attempted Synthesis of n6-Aniline-n5-cyclopentadienyliron(lI) Hexafluoro-
phosphate Using Anilinium Sulfate

Ferrocene (10.42 g, 0.056 mol),anilinium sulfate (27.92 g, 0.109 mol),

aluminum chloride powder (14.41 g, 0.108 mol), aluminum powder (6.00 g,

0.222 mol), and decahydronaphthalene (120 mL) were added to a reaction

flask against a countercurrent of nitrogen and stirred. The reaction

mixture was rapidly heated to 155C with the original pale green solution

turning black at 130C and forming a black oil on the sides of the reac-

tion flask. After 1 hour at 155C the black oil solidified and broke

away from the sides of the flask to reveal that a white insoluble mater-

ial was also present. After 3 hours the reaction was cooled to 00C and

150 mL of 15% aqueous-methanol solution was slowly added to produce an

orange organic layer and a green aqueous layer. After filtration of the

mixture, the aqueous layer was separated from the organic layer and

washed with 50 mL of diethyl ether. The aqueous solution was then added

to an aqueous solution of ammonium hexafluorophosphate (4.0 g, 0.025 mol).

The green precipitate which formed immediately was then filtered and

dried under vacuum to give 1.28 g of a green powder. 'H NMR spectra

showed the product to be the iron cyclopentadienyl complex of tetralin.









Sodium Hexamethyldisilylamide5

Sodium amide (11.3 g, 0.290 mol), hexamethyldisilazane (48.0 g,

0.298 mol), and benzene (150 mL) were added to a dry three-neck 250 mL

flask and stirred under nitrogen for 60 hours at 780C. After 60 hours

no ammonia was detected at the pressure release bubbler using damp red

litmus paper. The hot solution was then rapidly filtered through filter

paper and a yellow filtrate was collected. It was necessary to wash the

solid remaining on the filter paper with hot benzene to obtain the maxi-

mum of product. The combined filtrates were evaporated under vacuum to

give 43.0g(8l.0%) of a white powder.

57
Sodium Hydrogen Sulfide

Sodium metal (12.61 g, 0.548 mol) and absolute ethanol (200 mL)

were added against a countercurrent of nitrogen to a dry 3-neck 1 L

round bottom flask equipped with a condenser and a gas dispersion tube.

Addition of the ethanol produced vigorous bubbling from H evolution and
2
refluxing ethanol. After 45 minutes the sodium had completely reacted to

give a colorless solution. Hydrogen sulfide (26.0 g, 0.706 mol) was then

slowly bubbled into the system. After 2 hours the hydrogen sulfide gas

was stopped and 750 mL of dry diethyl ether was added. An immediate

fine, white precipitate which formed was subsequently filtered under

nitrogen and washed with diethyl ether to give 23.2 g (75.4%) of sodium

hydrogen sulfide.















REFERENCES


1. T. J. Kealy and P. L. Pauson, Nature, 168, 1039 (1951).

2. a. M. L. H. Green, "Organometallic Compounds," Vol. II, "The Transi-
tion Elements," 3rd Ed., Methuen and Co., Ltd., London, 1968.
b. J. Organometal. Chem., 100 (1975).
c. R. F. Heck, "Organotransition Metal Chemistry, A Mechanistic Ap-
proach," Academic Press, New York, 1974.

3. C. W. Bird, "Transition Metal Intermediates in Organic Synthesis,"
Logos Press, London, 1967.

4. H. Alper, Ed., "Transition Metal Organometallics in Organic Synthesis,"
Vol. I, Academic Press, New York, 1976.

5. H. C. Longuet-Higgins and L. E. Orgel, J. Chem. Soc., 1959 (1969).

6. 0. L. Chapman, C.-C. Chang, J. Kalc, M. E. Jung, J. A. Lowe, T. J.
Barton, M. L. Tumey, J. Amer. Chem. Soc., 98, 7844 (1976).

7. H. Sakurai, Y. Kamiyama, Y. Nadadaira, J. Amer. Chem. Soc., 98,
7453 (1976).

8. E. 0. Fischer and K. Ofele, Chem. Ber., 90, 2532 (1957).

9. H. J. Dauben, Jr. and L. R. Honnen, J. Amer. Chem. Soc., 80, 5570 (1958).

10. A. Steitwieser, Jr., and U. Muller-Westerhoff, J. Amer. Chem. Soc.,
90, 7364 (1968).

11. G. Wilkinson, M. Rosenblum, M. C. Whiting, and R. B. Woodward, J.
Amer. Chem. Soc., 74, 2125 (1952).

12. H. H. Zeiss and W. Herwig, J. Amer. Chem. Soc., 78,5959 (1956).

13. a. T. H. Coffield, V. Sandel, and R. D. Closson, J. Amer. Chem. Soc.,
79, 5826 (1957).
b. T. H. Coffield and R. D. Closson, U. S. Patent 3130214 (1964)
Chem. Abstr., 61, 4397g (1964).

14. M. L. H. Green, L. Pratt and G. Wilkinson, J. Chem. Soc., 989 (1960).

15. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl. Akad.
Nauk SSSR, 149, 615 (1963).

16. D. Astruc and R. Dabard, J. Organometal. Chem., iii, 339 (1976).








17. R. G. Sutherland, J. Organometal. Chem. Library, 3, 311 (1977).

18. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl. Akad.
Nauk SSSR, 166, 607 (1965).

19. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl. Akad.
Nauk SSSR, 175, 606 (1967).

20. 1. U. Khand, P. L. Pauson, and W. E. Watts, J. Chem. Soc., C., 2261
(1968).

21. D. Jones, L. Pratt, and G. Wilkinson, J. Chem. Soc. 4458 (1962)

22. J. F. Helling and G. Cash, J. Organometal. Chem., 73, CIO (1973).

23. E. 0. Fischer and H. Brunner, Z. Naturforsch., B., 16, 406 (1961)

24. A. N. Nesmeyanov, N. A. Vol'kenau, and E. I. Sirotkina, Izv. Akad.
Nauk SSSR, Ser. Khim., 1170 (1967).

25. E. 0. Fischer and W. Hafner, Z. Naturforsch., 106, 665 (1955).

26. D. Astruc and R. Dabard, C. R. Acad. Sci. Ser. C., 272, 1337 (1971).

27. C. C. Lee, R. G. Sutherland, and B. J. THompson, J. Chem. Soc., C.,
907 (1972).

28. I. U. Khand, P. L. Pauson, and W. E. Watts, J. Chem. Soc., C.,
2257 (1968).

29. V. R. Koch, L. L. Miller, and R. A. Osteryoung, J. Amer. Chem. Soc.,
98, 5277 (1976).

30. J. F. Helling and W. A. Hendrickson, J. Organometal. Chem., 141,
99 (1977).

31. R. L. Burwell, Jr., Chem. Rev., 54, 615 (1954).

32. D. E. Games, A. H. Jackson, L. A. P. Kane-Maguire, and K. Taylor,
J. Organometal. Chem., 88, 345 (1970).

33. J. W. Kang and P. M. Maitlis, J. Organometal. Chem. 30, 127 (1971).

34. A. Wu, E. R. Biehl, and P. C. Reeves, J. Chem. Soc., Perkin II,
449 (1972).

35. W. S. Trahanovsky and R. J. Card, J. Amer. Chem. Soc., 94, 2897 (1972).

36. G. Jaouen, A. Meyer, and G. Simmonneaux, J. Chem. Soc., C.,813 (1975).

37. P. M. Treichel and J. W. Johnson, J. Organometal. Chem., 88, 207 (1975).

38. C. White, S. J. THompson, and P. M. Maitlis, J. Chem. Soc., C., 409
(1976).








39. P. L. Pauson and J. A. Segal, J. Chem. Soc., D., 1677 (1975).

40. J. E. Sheats, W. Miller, and T. Kirsch, J. Organometal. Chem., 91,
97 (1975).

41. J. E. Sheats, W. Miller, M. D. Rausch, S. A. Gardner, P. S. Andrews,
and F. A. Higbie, J. Organometal. Chem., 96, 115 (1975).

42. R. Markby, H. W. Sternberg, and I. Wender, Chem. and Ind. (London),
1381 (1959).

43. J. W. Johnson and P. M. Treichel, J. Chem. Soc., C., 688 (1976).

44. P. M. Treichel and J. W. Johnson, Inorg. Chem., 16, 749 (1977).

45. J. W. Johnson and P. M. Treichel, J. Amer. Chem. Soc., 99, 1427 (1977).

46. D. J. Cole-Hamilton, R. J. Young, and G. Wilkinson, J. Chem. Soc.,
D, 1995 (1976).

47. W. S. Trahanovsky and R. A. Hall, J. Amer. Chem. Soc., 99, 4850 (1977).

48. C. White, S. J. Thompson, and P. M. Maitlis, J. Organometal. Chem.,
127, 415 (1977).

49. A. N. Nesmeyanov, N. A. Vol'kenau, L. S. Shilovtseva, and N. A.
Petrakova, J. Organometal. Chem., 61, 329 (1973).

50. a. J. W. Fitch, III and J. J. Lagowski, J. Organometal. Chem., 5,
480 (1966).
b. G. Hunter, E. 0. Fischer, and C. Elschenbroich, J. Organometal.
Chem., 3, 330 (1965).
c. G. Hunter, E. 0. Fischer, R. D. Rischer, 0. L. Carter, A. T.
McPhail, and G. A. Sim, J. Organometal. Chem., 6, 288 (1966).
d. H. Kobayashi, M. Kobayashi, and Y. Kaizu, Bull. Chem. Soc.
Jap., 46, 3109 (1973).

51. R. R. Schrock and J. A. Osborn, Inorg. Chem., 9, 2339 (1970).

52. a. R. Hoffmann and P. Hoffman, J. Amer. Chem. Soc., 98, 598 (1976).
b. R. C. Haddon, Aust. J. Chem., 30, I (1977).

53. H. Werner, Angew. Chem., 16, I (1977).

54. K. Nakamoto in "Characterization of Organometallic Compounds, Part I,"
M. Tsutsui, Ed., Interscience Publishers, New York, 1969, p. 126.

55. M. Mathew and G. Palenik, Inorg. Chem. II, 2809 (1972).

56. C. P. KrUger and H. Niederprum in "Inorganic Synthesis," VIII,
H. F. Holtzclaw, Jr., Ed., McGraw-Hill, New York, 1966, p. 15.

57. R. E. Eibeck in "Inorganic Synthesis," VII, J. Kleinberg, Ed.,
McGraw-Hill, New York, 1963, p..128.















BIOGRAPHICAL SKETCH


William A. Hendrickson was born on 2 July 1952 in Shreveport,

Louisiana. In June 1970 he graduated from John F. Kennedy Senior High

School in Bloomington, Minnesota. In May 1974 he received a Bachelor of

Arts degree with majors in math and chemistry from Saint Olaf College

in Northfield, Minnesota. In September 1974 he enrolled in the Gradu-

ate School of the University of Florida where he studied under a Gradu-

ate School Fellowship (1974-1975) and teaching assistantships from both

the Department of Chemistry and the Graduate School (1975-1978) toward

the degree of Doctor of Philosophy.

William A. Hendrickson married the former Lee Joan Davidson on

17 August 1974.









I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.



John F. Helling, Chairma~
Professor of Physical Science
and Chemistry



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.



Merle A. Battiste
Professor of Chemistry



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.



Jures A. Deyrup
P ofessor of Chemistry



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.



Robert C. Stoufer
Associate Professor of Chemi ry

PAGE 1

PREPARATION AND DEPROTONATIOti OF CERTAIN P I -ARENE-P I -CYCLOPENTAD I ENYL I RON ( I I ) SALTS By WILLIAM A. HENDRICKSON A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REOUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1978

PAGE 2

ACKNOWLEDGMENTS I would like to thank Dr. John F. Helling, my research chairman, whose guidance has made this dissertation possible. I would also like to recognize the great support of Dr. Jackie Dugan who provided me with excellent mass spectra of my compounds. Finally, I wish to thank Lee, my wife, for typing this dissertation,

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii LIST OF TABLES vi ABSTRACT vi i INTRODUCTION 1 PREPARATION OF n^-ARENE-n^-CYCLOPENTADIENYLIRONd I) SALTS 15 DEPROTONATION OF n^-ARENE-n^-CYCLOPENTADIENYLlRON(ll) SALTS .... 28 EXPERIMENTAL 59 n^-Anisole-n^-cyclopentadienyl iron(il) Hexaf luorophosphate . . 60 n^-Chlorobenzene-n^-cyc1opentadienyl iron(! I) Hexaf luorophosphate 61 n^-Dlphenylmethane-ri^-cyc1opentadienyl i ron(l I) Hexaf luorophosphate 62 n^-Fluorene-Ti^-cyclopentadienyl iron(l I) Hexaf luorophosphate . . 63 Ti^-Hexamethylbenzene-ri^-cyclopentadienyliron(l I) Hexaf luorophosphate 64 n^-Toluene-n^-cyclopentadienyl iron(l I) Hexaf luorophosphate . . 6k n^-Aniline-n^-cyclopentadienyl iron(l I) Hexaf luorophosphate . . 65 n^-Carbazole-n^-cyclopentadienyl iron(l I) Hexaf luorophosphate . 66 n^-D!phenylamine-n^-cyclopentadlenyl iron(l i) Hexaf luorophosphate 67 n^-i ^3^5-jri Isopropylbenzene-n^-cyclopentadienyriron(l I) Hexaf luorophosphate 69 n^-Thiophenol-n^-cyclopentadienyl iron( 11) Hexaf luorophosphate . 70 n^-Phenol-Ti^-cyclopentadienyliron(l I) Hexaf luorophosphate ... 71 ill

PAGE 4

TABLE OF CONTENTS (continued) Page il^-N,N-dimethylani 1 ine-r\^-cyc1opentadienyl i ron(l I) Hexafluorophosphate 72 n^-Thioanisole-n^-cyclopentadienyl i ron(lO Iodide 7^ n^-Anlsole-n^-cyclopentadienyl iron(l I) Iodide 75 n^-Carbazolyl-n^-cyclopentadienyl i ron(! l) 75 n^-Fluorenyl-n^-cyclopentadienyl i ron(l I) 76 ( l-5-n^-6-Di phenyl methyl enecyclohexadi enyl)-ri^ -eye lopentadienyl i ron( I I) 77 (l-5-n^"6-Phenyl iminocyclohexadienyl ) -n ^ -eye 1 open tadienylironCl I) 78 (l-5-ri^-Cyclohexadienyl-6-one)-n^-cyclopentadienyl i ron(l I) ... 78 {l-5-n^-Cyclohexadienyl-6-thione)-n^-cyclopentadienyl i ron(l I) . 79 (l-5-n^-6-lniinocycIohexadIenyl)-ri^-cyclopentadienyl I ron(i I) . . 8l a. Reaction in Dichloromethane 81 b. Reaction in Ammonia 82 c. Reaction in Tetrahydrofuran 82 d. Reaction in Acetonitrile '• 82 Reaction of n^-Di phenylmethane-n^-cyclopentad ienyl i ron( I I) Hexaf luorophosphate with Sodium Amide 83 Regeneration of n^-Di phenylmethane-n^-cyclopentadienyl i ron ( I! ) Cation 83 Attempted Reaction of Methyl Iodide with Certain ir-AreneTT-cyclopentadienyl i ron( M ) Salts 8^ a. n^-Ani 1 ine-n^-cyclopentadienyl i ron(ll) Hexaf luorophosphate 84 b. n^-Phenol-n^-cyclopentadienyl iron(li) Hexaf luorophosphate 84 c. n^-Thiophenol-n^-cyclopentadienyl i ron(l I) Hexaf luorophosphate 84 Preparation of n^-Ani 1 inium-n^-cyclopentadienyl i ron ( II) Cation 84 a. From n^-Ani 1 ine-n^-cyclopentadienyl i ron(l I) Hexaf luorophosphate 84

PAGE 5

TABLE OF CONTENTS (continued) Page b. From (l-5-Ti^-6-lmlnocyclohexadlenyl)-ri^-cyclopentadienyl i ron( I I) 85 Preparation of n^-Benzene-n^-cyclopentadienyl i ron(l I) Hexaf luorophosphate in Ethylpyridi nium Bromide-Aluminum Chloride Eutectic 85 Reaction of Hydroiodic Acid with n^-Anisole-n^-cyclopentadienyl i ron{ll) Hexaf luorophosphate 86 Attempted Synthesis of n^-Phenol-n^-cyclopentadienyl i ron(ll) Hexaf luorophosphate 86 Attempted Synthesis of n^-Ani 1 ine-n^-cyclopentadienyl i ron( I l) Hexaf luorophosphate Using Anilinium Sulfate 87 Sodium Hexamethyld i si lyl amide 88 Sodium Hydrogen Sulfide 88 REFERENCES 89 BIOGRAPHICAL SKETCH 92

PAGE 6

LIST OF TABLES Table Page I. n^-Arene-n^-cyclopentadienyl i rond I ) Salts 17 II. 'H NMR Spectral Data for n^-Arene-n^-cyclopentadienyl iron( 1 I ) Salts 19 III. Deprotonated n^-Arene-n^-cyclopentadienyl i ron ( II ) Complexes ^' IV. 'H NMR Spectral Data for Deprotonated n^-Arenen^-cyclopentadienyl I ron( I l) Complexes 40

PAGE 7

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 PREPARATION AND DEPROTONATION OF CERTAIN P I -ARENE-P I -CYCLOPENTAD I ENYL I RON ( I I ) SALTS By William A. Hendrickson August 1978 Chairman: Dr. John F. Helling Major Department: Chemistry A number of known ir-arene-Tr-cyclopentadienyl i ron( 1 1) salts have been prepared including the ir-complexes of benzene, toluene, mesitylene, diphenylmethane, fluorene, hexamethyl benzene, anisole, aniline, and chlorobenzene by the direct reaction of the appropriate arene, ferrocene, aluminum chloride, and aluminum at elevated temperatures. Also, new ucomplexes of carbazole, di phenyl amine, triphenylmethane, and 1,3,5-trii sop ropy 1 benzene were prepared directly from ferrocene and the respective arene as above. Finally, the ir-phenol and ir-thiophenol complexes were prepared by the reaction of n^-chlorobenzene-n^-cyclopentadienyliron(ll) hexafluorophosphate with potassium hydroxide or sodium hydrogen sulfide respectively. The Ti-complexes of fluorene, triphenylmethane, di phenyl amine, carbazole, phenol, thiophenol, and aniline were deprotonated with appropriate bases including sodium amide, sodium hexamethyldisi lylamide, potassium hydroxide, alumina, or the sodium salt of acetoni tri le. The TT-fluorenyl and ir-carbazolyl complexes seem to possess a greater degree of dipolar character than the others. The remaining complexes resulting from deprotonation have properties consistent with a structure containing an exocyclic double-bonded cyclohexadienyl unit.

PAGE 8

INTRODUCTION With the increased interest in organometal 1 ic chemistry following the discovery of ferrocene slightly more than twenty-five years ago, the field has developed into a major research area. ' This rapid development was, in part, promoted by the unique aspects of the field. The interface of inorganic and organic chemistry has led to a number of exciting reactions and compounds of unique structure and bonding. In an organometal lie complex the interaction of the metal and the organic 1 igand changes the properties of both the 1 igand and the metal. The quintessence, then, of organometal 1 ic chemistry is the way in which this interaction manifests itself in the reactions and structure of a particular compound . The field of catalytic chemistry is, to a large extent, dependent on the favorable interaction between metal atoms or ions and organic 3 molecules. The formation of either a transient or a stable organometal lie species allows the organic 1 igand to undergo reactions that are not normally possible under similar conditions in the absence of a metal atom. Synthetic organic chemistry is also heavily dependent upon k metal-organic compounds. V/hether it is the formation of alcohols from ketones via a Grignard reagent or the preparation of aldehydes via the Wacker process, metal atoms have become deeply entwined with the field of synthetic organic chemistry. Finally, formation of metal to carbon bonds has allowed the isolation and characterization of organic species that are normally considered unstable or transient intermediates. A

PAGE 9

classical example of this is the formation of stable metal complexes with cyclobutadiene such as n'*-tetramethylcyclobutadienenickel ( I I ) chloride dimer. Cyclobutadiene and its derivatives are so reactive that their isolation is not normally possible except by using matrix 6 isolation techniques. A more recent example of this is the isolation 7 of a metal-coordinated allyl system containing a silicon atom. Compounds with silicon-carbon double bonds or even partial double bonds are too reactive to isolate and this represents the first case of this structure being successfully isolated. While the field of organometal 1 ic chemistry may be subdivided according to reaction or structural type, it has also been classified according to ligand type such as complexes of carbon monoxide, alkenes, dienes, all
PAGE 10

/\ CI CI \ / ^ >k;i <^^ c c <^? Fe J^o+ BF.r7 \ c^ C^ figure FeClg

PAGE 11

Following the discovery and characterization of ferrocene, Zeiss and 12 Herwig showed that the previously reported reaction of phenyl magnesium bromide with chromium( I I I ) chloride did not give sigma complexes, but rather bis (n^-arene)chromium(O) compounds (Figure II). From this modest beginning, the number of compounds and the scope of the field have expanded tremendously. The sheer number of compounds runs into the thousands and includes most transition metals. 13 In 1957 Coffield, Sandel , and Closson reported the first preparation of a transition metal sandwich compound containing both a cyclopentadienyl and an arene ligand (Figure III). By reacting n^-cyclopentadienyl i ron dicarbonyl chloride or bis (n^"CyclopentadIenyl )di i ron tetracarbonyl with aluminum chloride in refluxing mesitylene, followed by hydrolysis in the workup, they were able to isolate n^-mesi tylenen^-cyclopentadienyl i ron ( I I ) iodide. Following this initial preparation, 13 I't Coffield et al . and others prepared approximately 15 compounds with similar structures. In 1963 Nesmeyanov et al . reported that it was possible to prepare the n^"benzene-n^-cyclopentad ieny 1 i ron ( I I ) tetraphenylborate by the reaction of ferrocene and aluminum chloride in the presence of aluminum powder and refluxing benzene. A mechanism for this reaction reported in 1976 by Astruc and Dabard proposed the initial 16 step to be the complexation of aluminum chloride and ferrocene. The aluminum chloride can complex at the cyclopentadienyl ring which leads to product or the aluminum chloride can complex at the iron atom to give an insoluble complex which does not react further (Figure IV) . Since the first paper by Nesmeyanov et_ aj|_. , reporting this reaction in 1963, this procedure has been the preferred synthetic route with well over 100 new -rr-arene-ir-cyclopentad ienyl i ron( I I } salts reported.

PAGE 12

CpFeCCOgCI + AICI3 41. A 2. H^ 3. Nal ^ « Fe* ICp^Fe^CCO)^ + AICI3 + '2' '•^^--'4 I. A 2. HgO 3. Nal FIGURE AlCI ^^ Fe — AlCI. ^^ ^^? F« 4AlCJ, 4Phrt -h Al 1. A 2. H^ ^ 3. NH4BF4 Fe+ BF4 Aia. AlCI. "^^^ LPhH -> F« " FIGURE IV

PAGE 13

While the reaction of ferrocene and aluminum chloride has great utility in the preparation of these mixed sandwich compounds, the use of aluminum chloride does put some restraints on the type of arene that can be reacted successfully. Since aluminum chloride is a very strong Lewis acid, the coordination of aluminum chloride with basic functional groups on substituted arenes has interfered with the desired reaction. 1 O IQ Nesmeyanov et^ aj_, 'reported that no n-arene-Tr-cyclopentadienyl i ron( 1 1) complexes were obtained when aniline, phenol, thiophene, or benzophenone were reacted with ferrocene and aluminum chloride (Figure V) . Subse19 quently, the aniline complex was prepared by an interesting route that will be discussed later while the others are yet to be reported in the literature. Chlorobenzene and bromobenzene also have the ability to act as Lewis bases toward aluminum chloride as they both have the requisite unshared pair of electrons. However, their reaction with aluminum chloride is less a concern than their reaction with the aluminum 20 powder in the reaction mixture to form benzene (Figure V). Over 50-^ of the product in some reactions of bromobenzene and ferrocene is the n^-benzene-n^-cyclopentadienyl i ron(l!) salt. This is especially unfortunate, not only with respect to the yield, but because the separation of the resulting products is almost impossible. Even a cursory inspection of the structure of these ir-arene-Tr, cyclopentadienyl i ron( I I ) complexes should provide some insight into the types of reactions that should be possible. An analogy might be drawn to reactions of 2,4-dini trochlorobenzene and anisole. Since anisole is an electron-rich system, electrophi 1 ic substitution on the ring should be facilitated. Conversely, 2 ,4-dinitrochlorobenzene is an electronpoor system and the same reaction would be unlikely; nucleophilic sub-

PAGE 14

r ANILINE r^ PHENOL r^ <^ THIOPHENE Pe + AICI, + Al BENZOPHENONE v_^ CHLOROBENZENE ^^-^ NO REACTION NO REACTION NO REACTION NO REACTION Fe* + F«* BROMOBENZENE Fe* FIGURE V

PAGE 15

stitution should be favored. Likewise, the -rr-arene-ir-cyclopentadienyliron(ll) cation should favor reactions with electron-rich species. Coupling this with a transition metal atom's known ability to stabilize normally unstable species leads one to propose the following four ligand reactions as viable possibilities: nucleophilic addition, nucleophilic substitution, ring metalation, and removal of an acidic alpha proton using a strong base (Figure Vl). Although this discussion is specifically about iron sandwich complexes, such reactions have also been described for complexes of other metals and ligands. The first of these reactions, nucleophilic addition, was discovered shortly after the initial preparation of these compounds. Green, Pratt, 1^ and Wilkinson reported that it was possible to react sodium borohydride with various Tr-arene-ir-cyclopentadienyl i ron( I I ) salts. The resulting product is viewed as a Meisenheimer intermediate stabilized by coordination to the iron atom (Figure Via). While the use of hydride as a nucleophile does not allow distinction between exo or endo attack, the ]k 21 ^ 22 use of sodium borodeuter ide, phenyl lithium, and sodium cyanide showed that attack is always exo . It should be noted that oxygen or nitrogen nucleophiles did not react with these ir-arene-ir-cyclopentadi22 enyliron(ll) cations as was also the case with the uncoordinated arenes. The second reaction, nucleophilic substitution, is quite analogous to nucleophilic addition with respect to the nature of the first step, addition of the nucleophile, and with respect to the structure of the resulting intermediate. However, for nucleophilic substitution to occur, a good leaving group such as bromide or chloride should be present which allows the formation of a coordinated arene by expulsion of the

PAGE 16

Nu F?^ + Nu> Fe Nu= H", D", /^^ R, CNA. NUCLEOPHILiC ADDITION Fe-*+ Nu> Fe* Nu= RO-,RS4:^^ ^^^^ RHN-, NH " B. NUCLEOPHILIC SUBSTITUTION ^ Fe-" + MB > , F«*^ + HB C. PROPOSED RING METALATION Fe* + B > Fe* + HB* ^^ ^^^ 0. OEPROTDNATIOW AT THE ALPHA POSITION FIGURE VI

PAGE 17

10 ^9 chloride or bromide ion. Nesmeyanov et^ a_l_. first reported the preparation of the n^-chlorobenzene-ri^-cyclopentadienyl irond I) cation and its reactions with nucleophi les. Under conditions that were remarkably mild Nesmeyanov et al . were able to react n^-chlorobenzene-n^-cyclopentadienyl irond I) salts with methoxide, phenoxide, thiophenoxide, amide, alkyl amide and phthalimide anions to obtain the substituted products in 50-90% yield (Figure VIb). The reactions were typically performed at 50° C for 1-2 hours in ethanol . The free chlorobenzene was usually unreactive under these conditions. Interestingly, it was reported that it was not possible to prepare the ir-phenol complex via 19 this route. it should be noted that the removal of an arene proton has not been reported in sandwich complexes where Iron is the metal. However, 23 it has been shown that metalation of bis (n^-benzene)chromium(O) is possible. Only in the past five years has experimental evidence of increased 22 acidity for alpha protons of the iron-coordinated arenes been available. Yet, previous to 1973, there was strong evidence to suggest that the idea was not without merit. During the course of Nesmeyanov's initial studies of the TT-arene-rr-cyclopentadienyl i ron ( I I ) salts it was discovered that these mixed ligand sandwich compounds are extremely stable 2k -' toward oxidation and reduction at the iron atom. This is rather surprising as the analogous ferrocene compounds are easily oxidized at the metal. Furthermore, if the ir-arene-ir-cyclopentadienyl i ron( I I ) salt was substituted with alkyl substituents on either ring, attempted oxidations oxidized the alkyl substituents to the corresponding carboxylic acids (Figure VI i). This reaction is interesting in light of the differences

PAGE 18

n Fe* 4KMn04 Pe-^ -jKMn04 •CH, Fe -f KMn04 FIGURE VI •COOH Fe P" Fe*V 4LiNH« or LiOMa > ^Z^ll^ I FIGURE VI I!

PAGE 19

12 observed with respect to ferrocene, but nrore pertinent to this discussion was the increase in acidity of the coordinated acids with respect 24 to the uncoordinated acids. The first evidence that arene alpha protons in iron systems might have enhanced acidity with respect to the uncoordinated arene alpha 22 protons was presented in 1973 by Helling and Cash. In an attempt to extend the variety of nucleophiles that could react with the coordinated arenes.they reacted methoxide and amide anions with bis (n^-mesi tylene)iron(ll) hexafluorophosphate. The expected reaction was the formation of a bis(n^-cyclohexadienyl)iron(ll) complex obtained by nucleophilic addition of either methoxide or amide ion. Instead, regardless of the nucleophile, the structure shown in Figure VIM was obtained. To explain the formation of this product it was necessary to consider the possibility that the nucleophiles were acting as bases to initially remove an alpha proton from one coordinated mesitylene. This produces a new nucleophile which adds readily to another unreacted bis (n^-mesi tyl ene)iron(li) cation. The compound then partially decomposes to give the product obtained. The key step in this reaction is the removal of an alpha proton by a strong base. As a check on the possibility that the iron was substantially affecting the acidity of the alpha protons, a reaction was carried out in which bis (n^-hexamethylbenzene) i ron(l I) chloride was dissolved in D with a catalytic amount of trisethylenediamine. After Ih hours at room temperature Sh% of the alpha protons had exchanged with deuterium. A similar experiment with uncomplexed hexamethyl benzene showed no deuterium exchange (Figure IX) . The greatly increased acidity of the alpha protons could be attributable to two effects promoted by the iron atom. The first is a simple indue-

PAGE 20

13 <^^ 24 HOURS < ^^ ^^:^Me)g DgO. CgH,^3 ^ ^^ REACTION 24 HOURS FIGURE IX F« H3 3ASE ^<^2><^«2 I I FIGURE X > Fe* < > F«

PAGE 21

1^ tive effect as seen with the carboxylic acids. Also, and perhaps more importantly, there exists the possibility of resonance interaction of the anion with the coordinated ring to form an exocycl ic double bond and cyclohexadienyl structure similar to the type seen with nucleophilic addition (Figure X) . A number of attempts were made to isolate these exocycl ic double-bonded compounds. However, their reactivity and instability precluded this possibility. Due to the greater stability of ir-arene-ir-cyclopentadienyl i ron( II ) salts with respect to the bis (n^-arene) i ron( I I) salts, the possibility of isolating a stable, exocycl ic double-bonded, cyclohexadienyl complex was pursued using these as reagents. This dissertation will deal with the preparation and characterization of these complexes.

PAGE 22

PREPARATION OF n^-ARENE-n^-CYCLOPENTADI ENYLI RON( I l) SALTS Since the initial preparation of n^-mesi tylene-n^-cyclopentadienyl13 iron(ll) iodide by Coffield, Sandel , and Closson and the subsequent 15 preparation of similar salts by Nesmeyanov , Vol'kenau, and Bolesova two general procedures have been used. The most widely used is the Fischer-Hafner synthesis with ferrocene as the iron-containing starting material. The second method involves the conversion of an existing n^-arene-n^-cyclopentadlenyl i ron(ll ) salt to a new u-arene salt by one of the methods described in the introduction, such as nucleophilic substitution or oxidation of alkyl groups on either of the rings. Two methods were used in this study to obtain good yields of these ir-areneir-cyclopentadienyl i ron(l I) salts. A number of other methods were attempted and these will be briefly described. 15 The Fischer-Hafner synthesis, as modified by Nesmeyanov et al . was the method used to prepare a wide variety of alkylated and substi26 27 27 tuted iT-coordinated arenes. Toluene, diphenylmethane, fluorene, 28 19 hexamethyl benzene, anisole, and 1 ,3 ,5-tri isopropylbenzene were reacted with ferrocene and aluminum chloride to obtain the respective ttarene-ir-cyclopentadienyl i ron(l I ) cations (Figure Xl). Of the preceding cations only the last, the tt-1 ,3, 5-tri isopropylbenzene complex, is a new cation. Excluding yields, the methods of preparation and characterization data obtained are generally consistent with the reported values (Table I). The sole important exception to this is the melting point of n^-toluene-ri^-cyclopentadienyl iron(i I) hexaf luorophosphate 15

PAGE 23

<^S>-CH, Fe* PFgFe* PFg ^HgPr OMe Fe^PFg<^ <^S> AMeh F«* PFg Fe* PFg ^ FIGURE XI "^^S^ I. AICI3Et Pyr Br' Fe + PhH •»• AI f ^ 2.H2O, NH^PFq Fe^P Fg" FIGURE XM

PAGE 24

17 TABLE I n^-Arene-n^-cycl open tad ienyl iron (I l) Salts ARENE YIELD Aniline 36.9 Anisole 18.6 b Anisole Carbazole A2.8 Chlorobenzene 28.3 Di phenyl amine 5-8 Diplienyl me thane 38.6 Fluorene 3'».1 Hexamethyl benzene 59.8 N,N-dimethylaniline 29.7 Phenol 36.3 Thioanisole 70.8 Thiophenol 50.4 Toluene 36.0 1 ,3,5-Tri isopropyl45.8 benzene DECOMPOSITION POINT In this work Reported in lit, 242° C 250 C 135-136°C 137.5-138.5 C 1A3-146°C 204-206 C 224-228° C 180-181°C 1 36-1 38° C 1 63-1 65° C 280-283° C 153-156. 5°C 204-206° C 119-122 C 236-238° C 286-289° C 264-265° C 165°C METHOD OF PREPARATION Fischer-Hafner Fischer-Hafner Methylation Fischer-Hafner Fischer-Hafner Fischer-Hafner Fischer-Hafner Fischer-Hafner Fischer-Hafner Methylation Nucleophi 1 ic Sub. c sti tution Methylation Nucleophi 1 ic Substitution Fischer-Hafner Fischer-Hafner 'With two exceptions all salts are hexaf luorophosphate salts. 'Complexes obtained as iodide salts. 'Complexes prepared using the tt -chlorobenzene salt as the starting compound,

PAGE 25

26 which was reported to be 165°C by Astruc and Dabard, but which was found here to be 286-289°C (decomposition). This discrepancy is explainable by assuming that Astruc and Dabard used an open melting point capillary instead of a sealed evacuated capillary tube. The 'H NMR spectra of the preceding complexes show absorption in the ^.625.356 region, corresponding to the cyclopentadienyl protons. The coordinated arene protons show absorption in the range of 5.80-7.3o6. The chemical shift data for each compound are tabulated in Table II. As referred to in the introduction, reactions of arenes substituted with heteroatoms in the Fischer-Hafner synthesis are complicated by the possibility of substituent reactions with either the aluminum chloride or aluminum present in the reaction mixture. Nesmeyanov et^a_l_.,in their preparation of n^"chlorobenzene-n^-cyclopentadienyl i ron ( I I ) tetrafluoroborate, reported that dehalogenation of the chlorobenzene starting material occurred under the conditions used and subsequently, a mixture a of TT-biphenyl and Tr-chlorobenzene cations was obtained. Nesmeyanov and co-workers reported further that it is possible to separate these two products on alumina. In this work the Russian synthesis was modified to avoid the dehalogenation problem. The preparation of pure n^" a. It should be noted that Nesmeyanov determined the presence or ttbiphenyl by the oxidative cleavage of his product and analysis of the arenes obtained. In a similar reaction with bromobenzene Nesmeyanov reports the reaction proceeds to give exclusively irbenzene and not TT-biphenyl. 'H NMR analysis of the product mixture obtained here when n^-chlorobenzene-n^-cyclopentadienyl i ron ( I I ) cation was prepared in the presence of aluminum powder shows that the product obtained by dehalogenation is n^"benzene-Ti^-cyclopentadienyl iron(l I) cation and not n^-biphenyl -r)^-cyclopentadienyl i ron( I I) cation. However, Khand et_ aj_. reported that under conditions similar to that reported by Nesmeyanov dehalogenation was not a problem. This is in contrast to the work by Nesmeyanov and the work presented here. There does not appear to be a rational explanation for the difference.

PAGE 26

19 TABLE I I b,c ARENE

PAGE 27

20 TABLE I I (continued) a.b ARENE CYCLOPENTACOORDINATED UNCOORDINATED OTHER DIENYL ARENE ARENE CH^, 3.21 (s.6) OH, 9.05 (s.l) CH^, 2.81 (s.3) SH, 3.60-3.92 (s.l) CH^, 2.58 (s,3) CH, 3.22 (m,3, J=6.5) CH ].k2 (d.l8, J=6.5) 'With two exceptions all salts were hexaf luorophosphate salts and spectra were obtained using acetone-d,. 'Salts were obtained as iodides and spectra were obtained using chloroform-d as the solvent. N.N-dimethyl-

PAGE 28

21 chlorobenzene-ri^-cyclopentadienyl 1 ron ( I I ) hexaf luorophosphate was accomplished by doing the reaction in the absence of aluminum powder. Consequently, it was necessary to reduce the ferricenium ion formed during the reaction prior to the precipitation of the product with ammonium hexaf luorophosphate. Significantly, the yield of product was not adversely affected by the absence of aluminum as seen on Table I. The 'H NMR data for this compound are given in Table II and are consistent 20 with the data provided by Khand et al . While these ir-arene-iT-cyclopentad ienyl i ron ( I I ) cations were being 29 prepared, the re was a report by Koch, Miller, and Osteryoung about electroinitiated Friedel-Craf ts reactions in the room temperature eutectic, ethylpyridinium bromide-aluminum chloride. The paper was pertinent to this research as both benzene and ferrocene are soluble in the eutectic. 16 In the mechanism proposed by Astruc and Dabard for the ligand exchange reaction they attributed the low yields to the formation of an insoluble complex of aluminum chloride and ferrocene. It was hoped that the use of this eutectic would obviate this solubility problem and hence greater yields would be possible. When the ethylpyridinium bromide-aluminum chloride eutectic was prepared and benzene, ferrocene, and aluminum were added and then heated at 80°C for 3 hours, n^-benzene-n^-cyclopentadienyl i ron( I I) hexaf luorophosphate was obtained (Figure Xli). However, through 'H NMR spectral analysis the product was shown to be a mixture of the desired hexaf luorophosphate salt and a salt of ethylpyridinium bromide. The mixture was recrystal 1 ized and chromatographed. However, no separation was observed. Further reactions using this solvent were attempted, but no product at all was obtained when either diphenylamine or anisole was used as the arene. This reaction potentially

PAGE 29

22 has great promise. However, the purification and general procedure needs substantially more work. The preparation of nitrogen, oxygen, and sulfur substituted tt18, 19 coordinated arenes has been explored primarily by Nesmeyanov et al . They reported that significant limitations were encountered when arenes substituted with oxygen, sulfur, or nitrogen were used in the FischerHafner synthesis. It was possible to prepare Tr-complexes of anisole and other aromatic ethers by this method, but it was not possible to pre19 pare a ir-complex of phenol by this method. Also, Nesmeyanov et al . reported that while the ir-acetanl 1 ide complex could be prepared directly, 18 that of aniline could not. The low reactivity of phenol and aniline reflects the ability of their substituent groups to coordinate with aluminum chloride. Our interest in the preparation of these complexes was stimulated by observations made in the early stages of this work which suggested that a series of coordinated cyclohexadienyl complexes bearing exocyclic C=N, C=0, or C=S double bonds might be synthesized by the reaction of a suitable base with the ir-complexes of anilines, N-alkyl anilines, phenols, or thiophenols. The reaction of n^~fluorenen^-cyclopentadienyl i ron ( II) hexaf luorophosphate with sodium hexamethyldisilylamide was found to give the relatively stable, neutral depro30 tonated complex. The formation of a stable deprotonated ir-complex from the n^~f luorene-n^"cyclopentadienyl i ron( I I ) cation provided the impetus to attempt the preparation of the analogous cattonic complex 18 derived from carbazole. On the basis of the work by Nesmeyanov et al . there was considerable doubt as to the chance of success and consequently, the high yield, k2.B%, obtained when carbazole, ferrocene, and aluminum chloride were reacted was unexpected. Subsequent reactions of diphenyl-

PAGE 30

23 amine and aniline, however, were only partially successful. The reaction of diphenylamine with ferrocene and aluminum chloride provided the cyclopentadienyli ron complex of diphenylamine in only a 6^ yield. Initially, the reaction of aniline to prepare the corresponding ir-complex was unsuccessful. A reaction of anillnlum sulfate, ferrocene, and aluma Inum chloride also failed. For aniline this problem was overcome by Increasing the reaction temperature. Excess aluminum chloride was necessary to compensate for the coordination of aniline and aluminum chloride, while the high temperature offset the lowered reactivity of the aniline due to its complexatlon with the aluminum chloride. When these changes were made, r\^anil ine-n^-cyclopentad ienyl i ron( I I ) hexaf luorophosphate was obtained in 36.9^ yield. The 'H NMR spectra obtained for these compounds is consistent with the proposed structures (Table II). The 'H NMR spectrum of the Irani line salt Is the same as that reported by Nesmeyanov et_ aj_. The IR spectra of the above compounds show in all cases an absorption at ap-1 proximately 840 cm corresponding to the hexaf luorophosphate anion. An N-H or NH stretch is seen at 3389 cm" , 3^17 cm" , or 3^99 and 3AOI cm for the diphenylamine, carbazole, and aniline complexes, respectively. When the conditions successfully used to prepare the ir-anlllne complex were applied to a reaction of phenol, ferrocene, and aluminum chloride, only a black viscous oil was obtained. The reaction of sodium hy*The reaction of anillnlum sulfate, ferrocene, and aluminum chloride was attempted as a possible way to avoid coordination of the aniline and aluminum chloride. The failure of this reaction may be attributed to the low solubility of the anillnlum sulfate In the decahydronaphthalene and the positive charge of the arene salt.

PAGE 31

2i^ droxide and a n^-chlorobenzene-n^-cyclopentadieny1 i ron( 1 1 ) salt would seem to be a reasonable method for preparing the coordinated phenol. However, when Nesmeyanov and co-workers attempted this reaction in water 19 at 50°C, only extensive decomposition occurred. Due to this report another route was sought for the preparation of the Tr-phenol complex. The preparation of phenol by the reaction of anisole and hydrogen iodide is a well-known reaction. However, the reaction of n^-anisole-n^cyclopentadlenyl iron(l I) hexaf luorophosphate with hydrogen iodide did not give the desired product; only decomposition was noted. Since it was not possible to prepare the u-phenol complex either directly or by using hydrogen iodide and the ir-coordinated anisole complex, attention was refocused on the reaction of the ir-chlorobenzene complex and hydroxide. In retrospect, the use of elevated temperatures was the probable reason that decomposition occurred when the reaction was attempted by Nesmeyanov et a1_. In a strongly basic media any iron-coordinated phenol formed would immediately react with another equivalent of base to form the deprotonated ir-phenol complex. In this work the deprotonated ttcomplexes were found to be unstable in solution at elevated temperatures. Therefore, when n^-chlorobenzene-n^-cyclopentadienyl i ron ( II ) hexaf luorophosphate was reacted in 50^ aqueous acetone at room temperature for 24 hours, followed by acidification and recrystal 1 ization , n^-phenol-n^. cyclopentadienyl i ron( I I ) hexaf luorophosphate was obtained in 36.3% yield (Figure XI I la). The 'H NMR spectrum of the proposed coordinated irphenol complex shows the expected absorptions for the cyclopentadienyl group and coordinated arene protons. The phenolic proton is observed as a broad peak at 9.055. Also, there is a strong, broad absorption at -1 3495 cm corresponding to an 0-H stretch.

PAGE 32

25 CI Fe* PFg+ 2 ^JaOH AQUEOUS ACETONE 24 HOURS L HCI Fe 4:^pN. 2. NH^PFg Fe* PFc A. <:Q>ci Fe^ PF Fe^PFg+ 2 NoHS ACETONITRILE 20 MINUTES I. HCI F« Z. NH^Fg > _Fe* PFB. FIGURE XIII

PAGE 33

TF^ One additional iron salt was prepared by an analogous route, n^ Thiophenol-ri^-cyclopentad ienyl i ron( I I ) hexaf luorophosphate was prepared by the reaction of sodium hydrogen sulfide with Ti^"chlorobenzene-Ti^cyclopentadienyl i ron( I I ) hexaf luorophosphate in acetoni tri le. In this reaction the deprotonated u-th iophenol complex was also initially obtained. The hexaf luorophosphate salt was obtained by acidifying the resulting solution with concentrated hydrochloric acid (Figure XI I lb). The 'H NMR spectrum shows an absorption at 3.6-3.926 of relative area corresponding to one proton. The IR spectrum shows the characteristic -1 absorption for an S-H stretch at 2584 cm . Mass spectral data have been obtained for all the new compounds. 32 The spectra are consistent with those reported by Games et^ £j_The parent cation is observed without the hexaf luorophosphate anion and characteristically the base peak is ferrocene or the arene ligand. For the aniline, thiophenol and phenol complexes a substantial number of large peaks of greater mass than the parent peak are observed. Two high mass peaks were observed in the mass spectrum of the ii-aniline complex that deserve special note, A peak at m/e 372 corresponding to C H Fe and 20 20 2 a peak at m/e 307 corresponding to C H Fe„ were seen. These peaks 15 15 2 were also seen in the mass spectrum of the deprotonated ir-aniline complex and will be discussed in the next section. The observation of peaks at a greater mass than the parent peak seems to correlate with the observed method of decomposition. In measuring the decomposition temperatures of a number of the ir-arene— rr-cyclopentadienyl iron( I l) salts two distinct decomposition modes were observed. The more common mode is a darkening of the product as the temperature is raised followed by the formation of a black oil at the decomposition

PAGE 34

27 temperature. An alternative method of decomposition does not form a black oil at the decomposition temperature, but instead turns white with a concurrent formation of a green or red oil at the top of the capillary tube. For those compounds which decomposed in the second mode it was not only difficult to determine a decomposition temperature, but the mass spectra showed the peaks above the parent peak. Perhaps in decomposing these compounds form dimers, trimers, etc., which are observed in the mass spectrum as high mass peaks. The results presented here on the preparation of TT-arene-ir-cyclopentadienyl iron(l I) salts indicate that the limitations imposed on their preparation by the presence of aluminum chloride or aluminum may be overcome by a judicious choice of reaction conditions or by choosing an al, 18, 19 ternate route to their preparation. The work done by Nesmeyanov et al . needs to be reviewed further in light of the results presented here. In the cases where a low yield or no product is reported there is a reasonable chance that those results could be easily changed with slightly different conditions.

PAGE 35

DEPROTONATION OF n^-ARENE-n^-CYCLOPENTADI ENYLI R0N( I I) SALTS 22 The paper by Helling and Cash provided the first evidence of increased acidity of alpha protons in ir-arene iron(ll) salts. However, a substantial amount of work has been reported in which increased acidity of alpha protons of other Tr-arene metal systems is observed. 33 In 1971 Kang and Ma it! is showed that the protons of n^-pentamethylcyclopentadienyl rhod ium( I I I ) chloride dimer could be completely exchanged using sodium deuteroxide in deuterium oxide (Figure XIV). The exchange was effected in 72 hours at 70°C. To explain this facile exchange of the alpha protons they suggested that an inductive and a mesomeric stabilization of the intermediate carbanion must occur. In 1971 two papers were presented that dealt with the T7-arenechromium(0) tricarbonyl systern. Wu, Biehl, and Reeves reported that when pKg values for a series of arene substituted Tr-phenolchromium(O) tricarbonyl complexes were obtained, there was a large increase in acid strength for all the ir-phenol complexes with respect to the uncoordinated phenols. When the pKg values were plotted against a~ values for the corresponding substituent, a good correlation was obtained. The fact that a" and not a values gave the "best correlation indicates that mesomeric interaction is of particular 35 importance in this system. Trahanovsky and Card reported that when n^-1 ,^-diphenylbutanechromium(0) tricarbonyl is reacted with potassium ^-butoxide in DMSO-d , followed by cleavage of the ring from the metal using eerie ammonium nitrate, a 70% yield of 1 ,4-diphenyl -1 , 1 -dideuteriobutane was obtained. Or, if n^-indanechromium(O) tricarbonyl is treated 28

PAGE 36

29 r)^-(CH^)^C^R^^ DgO NaOO FIGURE XIV iSP^b (CD,) 3'5 + < jdI$ >^CHr,)^Ph L KO-t-bu, DMSO-d. .Cr 0^ / Co ace'^' -> < gIi:$ >-CD^(CHg)^Ph ^^y I. KO-t-bu, DMSO-dg 2.Ce' FIGURE XV 0-7 CH2CH3 -IMel KO-tel — DMF

PAGE 37

30 in the same fashion, cis1 ,3~dideuter ioindane is obtained (Figure XV). The deuterium exchange experiments by Trahanovsky and Card and the pKa measurements by Wu et_ aj_. show that increased acidity can be expected with 7r-arenechromium(0) tricarbonyls. This was reiterated in 1975 by 36 Jaouen, Meyer, and Simmonneaux. They reported the reaction of potassium t;-butoxide, methyl Iodide, and a variety of irarenechromium(O) tricarbonyls in DMF. For Instance, the reaction of n^~ethylbenzenechromium(0) tricarbonyl with the above reagents yields n^"isopropylbenzenechromium(0) tricarbonyl and n^-t^-butylbenzenechromi um(0) tricarbonyl (Figure XVl). Under similar conditions the uncoordinated ethylbenzene did not react with methyl iodide. By the addition of hydrochloric acid to bis (n^-indenyl ) I ron (I I ) Johnson and Trelchel were the first to prepare n^~indene-n^-indenyl I ron ( I I ) 37 hexaf luorophosphate. Furthermore, it was possible to regenerate the starting material by addition of n-butyl 1 i thium to the cation. Conceivably, the regeneration of bis (n^-indenyl) i ron( I I ) was brought about by the formation of a ir-cyclohexadienyl Intermediate with an exocyclic double bond similar to the Intermediate proposed by White, Thompson, and Maitlis 38 in 1976 (Figure XVIl). White et_ aj_. reported that when trisacetonen^-pentamethylcyclopentadienylrhodium(l I I) hexaf luorophosphate or the iridium analog is reacted with indene or indole, a complex Is formed In which the llgand and metal are bonded via the six-membered ring of the arene. However, upon standing or upon the addition of a suitable base the complex isomer I zes to a complex in which the metal is coordinated to a f ive-membered ring. White and co-workers proposed that the rearrangement occurred via deprotonat ion to produce an intermediate coordinated cyclohexadienyl complex with an exocyclic double bond. This then iso-

PAGE 38

31 Fe HCl n-buLi ^^ ri75-( CH3)5C5M( ACETONE) J * * MsRhJr /y // X s N , CH ^^^S^ -H" < // (CHj), o C"3^5 FIGURE XVI I NHPh NoOH <^r>^ Ph ->• Mn EtOH C .<^y FIGURE XVIII

PAGE 39

32 merized to the observed product. The preceding 6 papers and the paper 22 by Helling and Cash provide a variety of examples of increased acidity of coordinated arene alpha protons which, in some cases, point to the formation of a ir-cyclohexadienyl system as an Intermediate. 39 In 197^ Pauson and Segal reported the isolation and characterization of (l-5-n^-6-phenyl imlnocyclohexadienyl )manganese (I ) tricarbonyl. They reacted n^-d iphenylaminemanganese( I ) tricarbonyl cation with ethanolic sodium hydroxide to obtain the neutral complex (Figure XVI I I). From 'H NMR and IR spectral analysis they assigned the structure as a ir-cyclohexadienyl system bearing an exocycl ic phenyl imine group. Sheats, Miller, 40 and Kirsh in 1975 also reported a structure that they believed contained an exocycl ic double bond (Figure XIX). The deprotonation of n^[ (di phenyl methyl ) eye 1 open tad ienyl ] -ri^-cycl open tad ienylcobal t ( I I I) cation with ethanolic sodium hydroxide resulted In a complex with substantial solubility differences compared to the starting cation that suggested the exocycllc double bond structure. A second paper on the acidity of alpha protons in cobal ticenium cations was presented by Sheats et al . in which the reaction of n^-cyclopentadienylcobal t dicarbonyl or n^-cyclopentadienyl rhodium dicarbonyl with dimethyl-, diphenyl-, or diperfluorophenyl acetylene provided products that were best represented as n'*-cyclopentadienone-n^-cyclopentadienylcobal t ( I I I ) complexes or the analogous rhodium(lll) complexes rather than zwitterionic structures like that pro42 posed by Markby, Sternberg, and Wender. In 1976 Johnson and Treichel reported the reaction of n fluoreneT)5-cyclopentadienyl i ron ( I I ) hexaf luorophosphate with potassium t^-butoxide in toluene to give a new product that is best represented as a zwitterionic structure. The analogously prepared ,deprotonated ir-f luorenemanganese(l)

PAGE 40

33 < ^^P ^»P^p NqOH Co* ^^ EtOH Co M + RCSCR — C c MsCo, Rh RsMe, Ph,Php FIGURE XIX <^^^^^2> M MsiyS-CgHgFe*, MnCCOj* FIGURE XX

PAGE 41

3A tricarbonyl complex was proposed to be a complex in which the exocyclic double bond is substantially more important. In 1977 further elaboration of this work was published in two separate papers ' in which 13 both compounds are shown by C NMR to have substantial negative charge residing at the 9-position. However, the neutral manganese complex appears to have less charge density at the 9-positlon and hence is less dipolar in nature than the neutral iron complex. The x-ray structure for the neutral iron complex was also obtained and it showed only a slight difference in the bond lengths of the two carbon-carbon bonds to C(9) and only a 11 deviation of C(10) from the plane of the coordinated arene. Finally, the neutral iron complex was much more reactive toward electrophi les than the neutral manganese compound. Therefore, the 13 x-ray structure, the C NMR, and the reactivity all argue for substantially more dipolar character in the neutral iron complex than in the neutral manganese complex (Figure XX) . 46 In 1976 Cole-Hamilton, Young, and Wilkinson reported the preparation of (l-5-n^~cyclohexadienyl-6-one)bis (tri phenyl phosphine) rutheni um( I l) hydride. Based on 'H NMR, IR, and x-ray analysis they proposed that the compound was best represented as a ir-cyclohexadienyl metal complex bearing an exocyclic carbon-oxygen double bond (Figure XXi). The 'H NMR spectrum obtained in benzene-d shows the expected 1:2:2 splitting of 6 the ir-cyclohexadienyl protons. However, in chloroform-d the ortho and para protons are observed as one peak. This difference was attributed to preferential solvation of one portion of the molecule. The IR spectrum shows a carbon-oxygen double bond stretch at 1577 cm and the xray structure of the above compound with a molecule of methanol solvated to it reveals that the carbon-oxygen bond is intermediate in length relative to carbon-oxygen single and double bonds found in model

PAGE 42

35 RuH^PPh3)4 PhOH -> Ru Ru , HOPh HOPh Phs"" 7^" Ph3' Ph3^ 7 " FIGURE XXI FIGURE XXI 17^ (CHjjgCgRhC ACET0NE)3 + PhOH Rh 3^5 Rh CH 3J5 NOg' CO, Rh* FIGURE XXIII

PAGE 43

36 compounds. Finally, as In other cyclohexadienyl systems, they noted the nonplanarity of the ring. An exocycl ical ly double-bonded chromium hi complex was reported in 1977 by Trahanovsky and Hall. They reported that (n -allyl phenyl ether)dicarbonylchromium Isomerized in benzene to (n -allyl) (l-5-Ti^-cyclohexadienyl-6-one)chromium dicarbonyl complex at room temperature (Figure XX I I ) . The 'H NMR spectrum of this complex has two TT-cyclohexadienyl peai
PAGE 44

37 the base was of sufficient strength to remove the alpha protons. To prepare the neutral iron cyclopentadienyl complexes of fluorene, diphenylamine, carbazole, and triphenylmethane, sodium amide in liquid ammonia , . 22 was used (Table III). Typically, 2-3 mmoles of the iron salt was reacted with a 10-fold excess of sodium amide in 90 mL of liquid ammonia for 1-3 hours. V/hile the deprotonated Tr-fluorene complex could be prepared In liquid ammonia with sodium amide, this complex was best prepared in benzene or diethyl ether using sodium hexamethyldls i lylamide (Figure XXIV). When the reaction was run in liquid ammonia, an anomalous temperature dependent peak appeared in the 'H NMR spectrum that shifted from 3.26 at 'tCC to k.2S at -SCC. A possible explanation for this peak might be a coordinated molecule of ammonia, but infrared analysis of the compound never showed peaks characteristic of ammonia. To avoid this complication the reaction was run in benzene or diethyl ether. A k3% yield of an intense blue-green product was obtained. Chemical shift data in benzene-d for this compound are presented in Table IV and are consistent with the spectrum reported by Johnson and Treichel. However, the 'H NMR spectrum of this neutral complex shows a remarkable downfield shift of the cyclopentadienyl absorption from 3.375 to A. 025 when acetone-d is the solvent. Similar solvent effects were noted 6 previously and were attributed to preferential solvation at a particular he portion of the molecule. The blue-green color of the deprotonated irfluorene complex is unusual in light of the red or red-orange colors seen for the other neutral complexes prepared in this work. Blue-green complexes for these Tr-arene-ir-cyclopentadienyl iron( i I ) cations have been ^9 reported for the reduced species, and in other metal arene systems 50 charge transfer complexes are known. The 'H NMR spectrum shows no

PAGE 45

38 NoNH NH^ 2-> Fe + ^ \NaNS»2(CH3)6 BENZENE m-. FIGURE XXIV <^2> NHPh F«*PFs NaNH< NH. <^>NPh Fe ^^ Fe^PFg^^^^^ NaNHr -> Fe^NH, ^ F«» PFgHPh, NaNH2 NH, Fe FIGURE XXV

PAGE 46

39 TABLE I I I Deprotonated n -Arene-n -cyclopentadienyl i ron( I I ) Complexes ARENE YIELD {%) DECOMPOSITION BASE AND SOLVENT POINT USED FOR PREPARATION ]kQ-]k()°C sodium amide, dichloromethane 176-178°C sodium amide, ammonia 122-124''C sodium amide, ammonia 105-107°C sodium hexamethyldisi lylamide, diethyl ether 112-llA°C sodium hydroxide, aqueous acetone 183°C sodium hydrogen sulfide, acetoni tri le 115"'117°C sodium amide, ammonia Aniline

PAGE 47

40 TABLE IV a,b 'H NMR Spectral Data for Deprotonated n -Arene-n^-cyclopentadienyl i ron(l I) Complexes ARENE Ani 1 Ine Carbazole CYCLOPENTACYCLOHEXADIENYL OR DIENYL COORDINATED ARENE 4.53(s,5) 4.13(s,5) 5.30-5.63(m,3) 4.83-5.05(m,2) 6.93-7. I8(m, 2) 5.50-5.82(m,2) UNCOORDINATED ARENE 8.07-8.27(m,l) 7.27-7.80(m,2) 6.65-6.93(m,1) OTHER NH, not observed Diphenylamine 4.48(s,5) 4.58-A.82(ni,2-under 6. 72-7. 00 (m, 3) eye 1 open tad! any 1 peak) 7.05-7.'t2(m,2) 5.17-5.55(m,2) 5.57-5. 83(ni,l) Fluorene e Fluorene Phenol Thiophenol Thiophenol Trl phenylmethane A.02(s,5) 3.37(s,5) A.57(s,5) 5.02(s,5) 4.70(s,5) A.55(s,5) 6.75-7. 13(m, 2) 6.38-6.57(m,I) 5.27-5.72(ni,2) 4.l6-4.75(m,2) 5.67-6. I6(m, 3) 4.67-4.97(m.2) 5.33-5.63(m,3) 6.00-6.37 (s, broad, 3) , 6.37-6.8it (s, broad, 2) 5.37-6.00 (s, broad, 3) , 6.00-6.34 (s, broad, 2) 4.12-4.38(m,2) 4.82-5. n(m, 2) 5.94-6.25(m,l) 8.08-8.33(m,l) vinyl CH 7.13-7.68(m,2) 5.75(s,l) 7.50-8.25(m,4) vinyl CH 6.13(s,l) 6.88-7.46(m,10) a. Chemical shift is expressed in delta(6) values relative to internal tetramethylsi lane. 'H NMR results are expressed as chemical shift (splitting pattern, relative area) . b. d. With few exceptions all spectra were obtained using acetone-d . . 6 'Acetonitri le-d was used as the solvent. Overlapping multiplets of one coordinated and one uncoordinated arene ring proton, 'Benzene-d, was used as the solvent. 'Ch.loroform-d was used as the solvent.

PAGE 48

k] broadening that might be expected for such an intramolecular complex charge transfer, but the charge transfer hypothesis is intriguing. Finally, the deprotonated Tr-fluorene complex was not observed to isomerize even when heated to 7S°C In benzene. This is consistent with the results reported by Treichel and Johnson. The remaining three salts were reacted with sodium amide in liquid anmonia to give intense red solids after workup (Figure XXV). The 85-90^ yields were substantially higher than those from the reaction of the ir-fluorene complex with the silicon amide. The 'H NMR spectra for the three deprotonated ircomplexes show a 0.37-0.53 ppm upfield shift of the cyclopentadienyl protons when compared to the spectra of their respective cations (Table IV). The 'H NMR spectrum of the deprotonated ir-carbazole complex is analogous to that of the neutral fluorene complex, while the deprotonated, neutral ir-di phenyl amine and -n-triphenylmethane complexes show the expected 1:2:2 splitting of coordinated cyclohexadlenyl protons. High temperature 'H NMR spectra of the deprotonated ir-triphenylmethane complex were obtained in hopes of observing fluxional behavior in the complex. However, when a tetrachloroethylene solution of the deprotonated complex was prepared and a series of spectra was obtained from 40°C to 90°C this was not observed. Interestingly, in contrast with most deprotonated ir-complexes the deprotonated ir-triphenylmethane complex did not decompose at elevated temperatures. The spectra are similar to those of other ir-cyclohexadienyl systems such as the nucleophilic addition products of ir-arene-ir-cyclopentadienyl14,21 iron(ll) cations ' and to the reported 'H NMR spectrum of (l-5-ri^~6phenyl iminocyclohexadienyl )manganese (1 ) tricarbonyl . However, in 1970 Schrock and Osborn reported that the reaction of rhodium(lll) chloride

PAGE 49

kz with 1 ,5-cyclooctadiene and sodium tetraphenyl borate gave n^-tetraphenylborate-n'*-(l ,5-cyclooctadiene)rhodlum(l) . This compound is a bona fide zwitterionic complex and yet, the 'H NMR spectrum also shows a 1:2:2 splitting of the arene protons. Also, in the ir-cyclohexadienyl-TT-cyclopentadienyl irond I) complexes formed by nucleophilic addition (Figure VI) the protons ortho to the site of attack are at substantially higher field than observed here. This may be a reflection of less than complete exocyclic double bond character and hence, lower charge density In 52 the ring. Therefore, the assignment of ir-cyclohexadienyl structures to the deprotonated ir-di phenyl amine and ir-triphenylmethane iron complexes is not justified on the basis of 'H NMR spectral evidence alone. Analysis of the infrared spectra of these complexes does provide an insight into the nature of these complexes. In all cases the neutral complexes show the loss of the hexaf luorophosphate anion by a dramatic decrease in Intensity of the peak at 8^0 cm . Also, there is a loss of the N-H stretch in going from the ir-diphenylamine and ir-carbazole cations to their respective neutral Tr-complexes upon deprotonation. Correspondingly, the TT-fluorene and ir-triphenylmethane complexes show the similar loss of the aliphatic C-H stretch after deprotonation has occurred. Concurrent with the loss of the N-H stretch following the deprotonation of the ir-di phenyl amine complex there is the formation of a strong absorption at 1562 cm . This absorption is consistent with the formation of a nitrogen to carbon double bond. In the analogous compound, (l-5-Ti^"6-phenyl39 iminocyclohexadienyl)manganese(l) tricarbonyl , the carbon to nitrogen double bond stretch Is reported as a strong peak at 1584 cm . It is unfortunate that a similar absorption can not be unambiguously assigned to carbon-nitrogen double bond stretch in the deprotonated TT-carbazole

PAGE 50

A3 ,, -1 complex, although medium intensity absorptions are noted at IS'*'* cm -1 and 1513 cm . The low intensities and position are not expected for a carbon-nitrogen double bond stretch. This leads to the conclusion that the deprotonated Tr-carbazole complex has a substantial degree of dipolar character. A carbon-carbon double bond stretching frequency cannot be unambiguously identified in the infrared spectra of the neutral TT-tr iphenylmethane complex because of strong absorptions in the 1620-1 1580 cm region in both the starting material and the product. However, the similarity of the 'H NMR spectra of the deprotonated ir-triphenylmethane and deprotonated ir-d i phenyl amine complexes suggests a similarity in their structure. Furthermore, the neutral tr iphenylmethane complex is the only neutral complex prepared in which the ortho protons are at higher field than the ir-cyclopentadienyl protons. This is consistent with a high degree of exocycl ic double bond character in this compound. The mass spectra for the deprotonated ir-f luorene, ir-di phenyl amine, iT-carbazole, and ir-triphenylmethane all give a parent peak with the correct exact mass. Usually the base peak for these complexes was attributable to the respective arene and large peaks were observed at I86, 121, and 65, corresponding to ferrocene, cyclopentadienyl i ron, and cyclopentadienyl cations. The first three compounds give parent peaks with extremely low intensity parent peaks while the neutral ti -triphenylmethane parent peak has an intensity of 31% of the base peak. That such a large peak is seen only in the mass spectrum of the deprotonated tt -triphenylmethane complex suggests a higher stability for this complex than for the other deprotonated complexes. This may be attributable to substantially more cyclohexadienyl character in the deprotonated ir-tri phenyl methane

PAGE 51

kk complex. Exact masses were obtained for the parent ions in these k deprotonated it -complexes. This data was used along with satisfactory iron analyses in place of carbon-hydrogen elemental analyses to fully characterize the deprotonated ir-complexes . Satisfactory carbon-hydrogen elemental analyses were not obtained for these complexes, and this is perhaps due to their low stability. Three other iron salts were successfully reacted with various bases to give neutral complexes. The ir-aniline, Tr-phenol , and rr-thiophenol complexes were deprotonated to provide the respective neutral complex. As mentioned in the preceding section the reaction of n^-chlorobenzenen^-cyclopentadieny1 i ron(i I) hexaf luorophosphate with sodium hydroxide or sodium hydrogen sulfide produced the respective deprotonated complex by an initial nucleophilic attack of hydroxide or hydrosulfide and loss of chloride ion followed by the reaction of the resulting complexes with another equivalent of hydroxide or hydrosulfide to form the deprotonated complexes (Figure XXVl). The deprotonated TT-phenol complex was prepared by reaction of the TT-chlorobenzene complex with hydroxide ion in 50% aqueous acetone, a medium in which the ir-chlorobenzene complex is soluble. After the reaction had proceeded at room temperature for 18-2A hours, the acetone was removed by vacuum evaporation. The deprotonated TT-phenol complex was found to be soluble in the resulting aqueous solution while the unreacted ir-chlorobenzene complex was not and was easily removed by filtration. Evaporation of this solution under vacuum followed by extraction with dlchloromethane and recrystal 1 ization afforded the product in 82. n yield (Table III).

PAGE 52

A5 Fe^PFg 2NaOH ^=> CH. OCH, I -> _Ee -> Fe-^l< gZl2 >Cl 2 NoHS < £II> S CH,I SCH, Fe*PFQ ^^ -> Fe -> Fe* |<^S>NHp Fe*PFgI. NaNH2 2. CH3I 3. AgNOg 4. NH^PFg <^2>N(CH3)2 NaNH. CH3I Fe CH,I <^2>NHCH3 > _Fe*|Fe II + III ^ FIGURE XXVI

PAGE 53

k6 The deprotonated ir-thiophenol complex was prepared similarly, n Chlorobenzene-n^-cyclopentadienyl irond I) hexaf luorophosphate and sodium hydrogen sulfide were reacted in acetonitrile for 20 minutes and then filtered. Separation of the desired deprotonated ir-complex from unreacted starting material was accomplished by passing the filtrate through an alumina column. The compound was recrystal 1 ized to afford a 55.3% yield of the deprotonated ir-thiophenol complex (Table ill). The deprotonation of the iT-phenol and ir-thiophenol complexes was also possible by the reaction of ammonia, sodium amide, or alumina with n^-phenol -n^-cyclopentadienyl irond I) hexaf luorophosphate or n^-thiophenol -n^-cyclopentadienyliron(li) hexaf luorophosphate, but the direct preparation from n^-chlorobenzene-n^-cyclopentadienyl irond I) hexaf luorophosphate produced the desired product in much higher yield. As was noted previously by Cole-Hamilton et_ aj_. and by Trahanovsky ^7 and Hall a 1:2:2 ratio of the ir-arene protons is not necessary for a complex to be consistent with an exocycl ical 1y double-bonded cyclohexadienyl system. When the 'H NMR spectrum of the chromium complex was obtained in acetone-d and the 'H NMR spectrum of the ruthenium complex was 6 obtained in chloroform-d , a 3:2 ratio of cyclohexadienyl protons was observed. However, in benzene-d Cole-Hamilton et al . reported that the 6 ruthenium complex had the expected 1:2:2 ratio for the cyclohexadienyl protons. The 'H NMR spectra of the deprotonated -rr-phenol and Tr-thiophenol complexes show a 0.2-0,6 ppm upfield shift of the it -eye 1 open tadienyl protons when compared to those to the respective starting cations (Table IV) . The ir-cyclohexadienyl protons of the deprotonated complexes are also shifted upfield 0.5-1.7 ppm. There is a concomitant splitting of the rr-cyclohexadienyl protons to give a 3:2 splitting

PAGE 54

hi pattern in both cases when acetone-d is used as a solvent. The 'H NMR 6 spectrum of the deprotonated Tr-thiophenol complex in chloroform-d did not show a 1:2:2 splitting pattern of the ir-cyclohexadienyl protons, but changing the solvent did give an upfield shift of 0,30 ppm for the ttcyclopentadienyl and Tr-cyclohexadienyl protons. Conceivably, the 'H NMR spectra of the deprotonated ir-thiophenol or deprotonated ir-phenol complexes would exhibit a 1:2:2 splitting of their cyclohexadienyl protons in benzene-d . Unfortunately, neither fr-complex is soluble in that sol6 vent . The structures of the deprotonated ^-phenol and ir-thiophenol complexes cannot be assigned unambiguously from the 'H NMR spectra alone, but as before infrared analysis is extremely helpful. Upon deprotonation of the ir-phenol complex there is a loss of the OH stretch and the -1 PF stretch with a concurrent gain of a peak at 1661 cm attributable to a carbonyl stretch. The infrared spectrum of the deprotonated -rr-thiophenol complex is instructive since it shows neither an SH stretch nor -1 a PF stretch but does show a strong absorption at 1082 cm corresponding to a thione stretch. Therefore, the appearance of thione or carbonyl absorptions in the infrared spectra and the 'H NMR spectra are consistent with the proposed TT-cyclohexadienyl structures bearing exocycl ic double bonds. Mass spectral data were obtained for both of the deprotonated ttphenol and ir-thiophenol complexes. In each case large peaks were obtained at m/e l86, 121, 66, and 56 corresponding to the cations of ferrocene, cyclopentadienyl i ron , cyclopentadiene, and cyclopentadienyl . The deprotonated ir-phenol complex showed an unusually large parent peak at l\h which was 36.2% of the base peak, phenol. In contrast, the parent

PAGE 55

48 peak of the deprotonated ir-thiophenol complex at 230 was only 3-5% of the base peak, ferrocene. Other strong peaks in the mass spectrum of the deprotonated ir-thiophenol complex were at m/e 186 (C.^H S), 185 (C H S), 184 (C,,H„S), 110 (CHS), and 77 (C H ). As was the case 12 9 12 8 6 6 65 with the deprotonated Tr-complexes of fluorene, carbazole, di phenyl amine, and triphenylmethane, it was not possible to obtain satisfactory carbonhydrogen elemental analyses for these compounds. However, satisfactory iron analyses and an exact mass for the parent ion of both of the deprotonated IT -complexes were obtained. Methyl iodide was added to separate solutions of both the deprotonated TT-thiophenol and Tr-phenol complexes (Figure XXVI). The formation of the TT-thioanisole complex was confirmed by 'H NMR and IR spectra and by elemental analysis. The formation of the Tr-anisole complex was confirmed by comparing its 'H NMR spectrum to that of an authentic sample prepared directly from anisole, ferrocene, and aluminum chloride. The reaction of methyl iodide with both of the deprotonated complexes was rather slow, requiring an hour for the ir-phenol complex. It was also found that methyl iodide did not react either with the ir-phenol or irthiophenol complexes to give the corresponding ir-anisole or ir-thioani sole 47 complexes. Interestingly, Trahanovsky and Hall reported that methyl iodide did not react with the (n^-al lyl ) (l-5-ri^-cyclohexadienyl -6-one)-chromium(O) dlcarbonyl complex. On the basis of the lower carbonyl frequency observed for the chromium complex in comparison to that of the deprotonated ir-phenol iron(ll) complex, a lower bond order of the carbonoxygen bond would be expected in the chromium complex. This in turn would indicate that the chromium complex should react faster with methyl iodide than the deprotonated ir-phenol complexes prepared here.

PAGE 56

49 Therefore, from this work and from the work by Trahanovsky and Hall it can be seen that a direct comparison of ground state electron densities and reaction rates is not valid. This direct relation was assumed incorrectly by Johnson and Treichel in their discussion of the structures of the TT-fluorenyl complexes of cyclopentadienyl i ron( II) and manganese(l) tricarbonyl. As seen above this is not valid; however, Johnson and Treichel do not use this point as the only basis for assigning the structures of the ir-fluorenyl complexes. The deprotonated Tr-aniline complex was obtained by reacting n^" aniline-n^-cyclopentadienyl iron(l I) hexaf luorophosphate with sodium amide in dichloromethane for 1.5 hours (Figure XXVI). The compound was obtained as a red solid which decomposed in acetone-d but was stable in acetonitri le-d to give a 'H NMR spectrum similar to that of the deprotonated TT-phenol and ir-thiophenol complexes (Table IV) . The deprotonated ir-aniline complex could also be prepared in acetonitri le using NaCH^CH. However, purification of the resulting deprotonated complex was not possible. The 'H NMR spectrum of the complex shows a number of small peaks in the 1-36 region which were not removed upon recrystal 1 ization. Also, there were 3 peaks in the IR spectrum in the N-H stretch region which was indicative of a mixture of deprotonated Tr-aniline complex and unreacted tr-aniline complex. An attempt was made to prepare the compound in liquid ammonia with sodium amide, but the initial red solution turned black in less than one minute and no product was obtained. Similar results were obtained when the reaction was run in tetrahydrofuran except that the complex did not decompose as fast. The infrared spectrum of this complex shows only one N-H absorption; this is consistent with a deprotonated complex. However, there

PAGE 57

50 is still a strong absorption at SkO cm indicative of the presence of hexafluorophosphate anion and hence unreacted starting material. The 'H NMR spectrum of this complex shows the expected upfield shift of the coordinated arene and cyclopentadienyl protons relative to the starting salt. There is also an observed splitting of the coordinated arene protons into multiplets of relative area 3:2. The mass spectrum of this complex was obtainable at 80 C in contrast to that of the Tr-aniline complex for which a temperature of 150°C was needed. Hence, there are fewer high mass peaks of large intensity. At 80°C large peaks were observed at m/e 186, 121, 93, 66, 65, and 56 which correspond, respectively, to the cations of ferrocene, cyclopentadienyl i ron, aniline, cyclopentadiene, cyclopentadienyl, and iron. The parent peak, 3.5^ of the base peak, was observed at 213. A correct exact mass for this peak was also obtained. Correct exact masses were also obtained for peaks at 372 and 307 corresponding to C H Fe and C H Fe . A structure for the first complex ^ ^ 20 20 2 15 15 2 is not immediately obvious, but the second ion is consistent with a triple decker sandwich, a structure unreported for iron and cyclopentadienyl ligands (Figure XXVI I), although triple decker sandwiches are , 53 known for other metals. While the 'H NMR, infrared, and mass spectra of this deprotonated complex were consistent with the proposed structure, several iron analyses were unsatisfactory. The complex appears to be sufficiently susceptible to decomposition that quantitative analyses are very difficult. A derivative of this compound was prepared by deprotonating the ir-aniline salt in dichloromethane and then adding excess methyl iodide. The previously unknown TT-N,N-dimethylani 1 ine complex was obtained in 29.7^ yield after workup and provided satisfactory 'H NMR and IR spectral data.

PAGE 58

51 Fe Fe FIGURE XVI I < Sll2 >-CHpP>> NaNHg Fe-*-P Ffi> Fe 'Ph O^ NH. Ph Ph •^ H"^ ^Ph Ph Ph ?57

PAGE 59

52 Carbon, hydrogen, and iron analytical data also confirmed the proposed structure. A rational mechanism to explain why the TT-N,N-dimethylani 1 ine complex was obtained instead of the ir-N-methylani 1 ine complex is shown in Figure XXVI. Addition of the methyl iodide to a solution of the deprotonated Tv-aniline complex would form the expected ir-N-methylani 1 ine complex. This complex can then react with an additional deprotonated ir-aniline complex to obtain the deprotonated TT-N-methylani 1 ine complex and the Irani line complex. The new deprotonated ir-N-methylani 1 ine complex could then react with another equivalent of methyl iodide to form the observed Tr-N,N-dimethylaniline complex. This mechanism is supported by the observation of the ir-aniline complex in the 'H NMR and IR spectra of the crude product. Furthermore, this reaction is analogous to the addition of acetyl chloride to the deprotonated ir-fluorene complex. Johnson and Treichel reported that that reaction gave a mixture of the ir-fluorene complex and a ir-diacetylated fluorene complex. The possibility that the methylated complex was obtained directly from the ir-aniline complex may be ruled out as the ir-aniline complex was found to be unreactive with methyl iodide under similar conditions. An alternate explanation for the formation of the ir-N,N-dimethylani 1 ine complex rests on the proposal that under the initial reaction conditions both alpha protons on the ir-aniline complex were removed. Methyl at ion would then occur twice without the need ,for a two step methyl at ion as shown in Figure XXVI. This seems unlikely, but the possibility was checked by adding excess acetic acid-d to separate acetone-d solutions of the deprotonated ir-aniline complex and the ir6 aniline complex. Addition of the acetic acid-d to each complex forms the ir-anilinium complex. This is seen by the downfield shift of the N-H protons from 5.76 in the ir-aniline complex to 9.^06 in the ir-anilinium complex.

PAGE 60

53 The ir-anilinium complex formed from the deprotonated Tr-anlline complex has a relative area ratio of 1:5:5 for the N-H, coordinated arene, and cyclopentadienyl protons respectively, in contrast, the ir-anilinium complex formed from the Tr-aniline complex has a relative area ratio 2:5:5 for the N-H, coordinated arene, and cyclopentadienyl protons respectively. The observation of a peak of relative area two in the 'H NMR spectrum of the ir-anilinium complex obtained from the Tr-aniline complex shows that proton exchange does not occur under these conditions. The lack of observable proton exchange and the observation of only one N-H proton in the 'H NMR spectrum of the ir-anilinium complex obtained from the deprotonated ir-aniline complex are consistent with the structure shown in Figure XXVI for the deprotonated ir-aniline complex. Four other ir-arene-ir-cyclopentadienyl i ron(l I) cations were reacted with strong bases such as sodium amide or sodium hexamethyldisi lylamide. It was found the ir-toluene, ir-hexamethyl benzene, and ir-1 ,3,5-tri isopropylbenzene complexes did not react to any extent with sodium amide in liquid ammonia. Presumably this was due to an extremely low acid strength of the alpha protons even in a cationic metal complex. When n^-di phenyl methane-ri ^-cyclopentadienyl iron(l i) hexaf 1 uorophosphate was reacted with sodium amide in ammonia or with sodium hexamethyldisi lylamide in benzene or diethyl ether, an intense red solution forms immediately. Workup of these solutions in the normal way was done to obtain a red solid. When an 'H NMR spectrum of the product was obtained, there was observed a single peak for the arene protons at 1.21^ and the cyclopentadienyl protons were split into two peaks centered at A.00<5 and 4.136 of relative area 2:3The peak at 4.136 was a multiplet while the peak at 4.006 was a singlet. Also, a singlet of relative area 1 was observed at 3.836. The structures in Figure XXVI 11 are consistent with the

PAGE 61

5^ singlet observed for the arene protons. However, this does not explain the splitting of the cyclopentadienyl protons nor the absence of the required iron-hydrogen absorption in the infrared spectrum in the region -1 5^ 2100-1700 cm . Furthermore, the reaction of the deprotonated complex with hydrochloric acid is inconsistent with such a structure. Acidification of the deprotonated ir-complex prepared in situ resulted in the recovery of the starting ir-di phenyl methane cation as evidenced by 'H NMR spectral analysis. The possibility that fluxional behavior was responsible for the observed spectra was discounted when the 'H NMR spectrum was found to remain constant down to -60 C. No consistent explanation of these results is yet possible and further research on the system is needed . The deprotonation of Tr-arene-7r-cyclopentadienyl i ron( I I) salts can give products that show definite exocyclic double bond character. However, the results upon deprotonation of either the T7-fluorene or ir-carbazole complexes are not as definite. Johnson and Treichel believe that 1 3 the reactivity, C NMR spectrum, and the x-ray structure all argue for the deprotonated Tr-fluorene complex being best represented as a dipolar structure. They do point out that there is some shortening of the carbon-carbon bond from C(9) to the coordinated portion of the molecule and it is also noted that the dihedral angle between the planes described by C(9), C(10), and C(ll) and C(l)-C(i») is 10-11°C. This is discounted as small in comparison to the dihedral angle of 43 seen for bis(6tertbutyl-1 ,3,5-trimethylcyclohexadienyl) iron(l I) (Figure XXIX). Their explanation for this small dihedral angle and hence the dipolar nature of the deprotonated T:-fluorene complex is based upon steric constraints in the fused ring system. . .

PAGE 62

55 Fe 4 8 ^13^ 6 R l-bu FIGURE XXIX 9 ssin'' [o.409/(l.277+a)] FIGURE XXX

PAGE 63

56 On the basis of these arguments and the fact that the deprotonated i:-fluorene and ir-carbazole complexes obtained in this work have similar 'H NMR spectra, a dipolar structure could be suggested for both complexes. However, two points need further consideration. There has been one other deprotonated u-complex prepared in which a partial x-ray structure was reported. Cole-Hamilton et^ aj_reported that the exocyclic carbon-oxygen bond for (l-5-n^-cyclohexadienyl-6-one)bis(triphenylphosphlne)ruthenium(!i) hydride is 1.277^, or midway in length between a carbon-oxygen single or double bond. Furthermore, they report that the oxygen atom is 0.409^ above the plane of the ir-cyclohexadienyl ligand. Unfortunately, they do not report the actual dihedral angle, but assuming a C(l)-C(6) bond length of \ .k% and using the information given it is possible to calculate a dihedral angle of 9-9 (Figure XXX). This compares favorably with the results obtained by Johnson and Treichel. However, in this case there are no steric requirements that hinder the formation of a larger dihedral angle. The second point to consider when reflecting upon the importance of the size of the dihedral angle in 52 these complexes is the report by Hoffmann and Hoffman in which they suggest that the large dihedral angle observed in ir-cyclohexadienyl systems in which the 6-position is saturated is due, in part, to an endo Hiron orbital interaction, but the major contribution to the formation of a non-planar system results from an anti symmetruc orbital interaction. in iT-cyclohexadienyl systems formed with an exocyclic double bond, there would be no endo H-iron orbital interaction and the antisymmetric orbital interaction would be minimized. Therefore, the size of the dihedral angle should be smaller. The fact, then, that dihedral angles of 9-11 have been observed in two deprotonated rr-arene systems is consistent with the

PAGE 64

57 formation of an exocyclic double bond. In this work we have demonstrated that the deprotonated ir-complexes of triphenylmethane, di phenyl amine, phenol, and thiophenol have substantial exocyclic double bond character. X-ray structural data for these complexes would be helpful in determining the merit of assuming exocyclic double bond character for a dihedral angle of 9-11 . The deprotonated TT-phenol , ir-thiophenol , and u-diphenylamine have the exocycl ical ly double bonded structure. The 'H NMR spectra are consistent with a cyclohexadienyl structure and the infrared spectra have absorptions corresponding to the expected C=0, C=S, or C=N stretches. The lack of an assignable C=C stretch for the exocyclic double bond of the deprotonated Tr-triphenylmethane complex in the infrared spectrum and the existence of a strong absorption in the infrared spectrum of n^-aniline-n^-cyclopentadienyl iron(i 1) hexaf luorophosphate that obscures the C=N region for the deprotonated ir-aniline complex does not allow the unambiguous assignment of these structures. However, the deprotonated Tr-triphenylmethane and iv-aniline complexes were assigned the exocyclic double bonded structure based on the similarity of their 'H NMR spectra to those of the preceding three deprotonated -rr-complexes. The preceding five complexes are substantially exocyclic double bonded complexes, but some degree of dipolar character probably is present in each. The infrared spectrum of the deprotonated ir-diphenylamine complex shows a C=N stretch that is 22 cm lower frequency than the C=N stretch reported for (l-5-n^-6-phenyl iminocyclohexadienyl )manganese( I ) 39 tricarbonyl by Pauson and Segal. This shift to lower frequency for the C=N stretch is consistent with a lower bond order in the exocyclic double bond of the Iron complex when compared to the manganese complex.

PAGE 65

58 The deprotonated ir-phenol complex has a C=0 stretch in the infrared -1 spectrum that is 30-60 cm higher than any of those previously reported !i4 50, 51 deprotonated tr-phenol complexes. ' ' THis suggests a higher bond order for the exocyclic carbon-oxygen bond and hence a lower degree of dipolar character. A greater electron withdrawing effect of the iron would tend to favor the exocyclic double bond form. This might be reflected in a larger dihedral angle between the carbonyl plane and the cyclohexadienyl plane resulting in less polarization of the double bond and hence giving a higher frequency in the IR spectrum. Conceivably, the still unknown (l-5-n^-cyclohexadienyl-6-one)manganese( I) tricarbonyl should show a higher v than the iron complex. c=o Finally, there have been no previous reports of a iicoordinated metal complex containing an exocyclic sulfur to carbon double bond. However, the fact that the v is observed at the high frequency end of c=s the V region suggests that the complex has a strong carbon-sulfur c=s exocyclic double bond and that the complex exists primarily in the ttcyclohexadienyl form. There remains a substantial amount of work to be done in this field. The series of exocyclic double bonded compounds should be extended by changing the metal to which the arene is coordinated and by changing the arene. Also, reactions that would take advantage of an exocyclic double bond such as (2+4) cycloaddi tions, Wittig reactions, or SimmonsSmith reactions could have some synthetic utility and should be attempted.

PAGE 66

EXPERIMENTAL All reactions were run under a N atmosphere. Except for acetone or methylene chloride, which were reagent grade, solvents used in the preparation and purification of air-sensitive compounds were distilled from lithium aluminum hydride or calcium hydride slurries. The preparation and purification of the hexaf luorophosphate salts were done with reagent grade solvents. Reagents in all reactions were reagent grade and were not purified before use. Yields are based on the iron-containing reagent exclusively. All reactions to prepare the ir-arene-ii-cyclopentadienyl iron( I I) hexaf luorophosphate salts from ferrocene and the appropriate arene were run in three neck , round -bottom flasks that had been dried by repeatedly evacuating the system, flame drying, and filling with nitrogen. All reagents were added to the system against a countercurrent of nitrogen. When the reagents had been added, the reaction vessel was stoppered and the reaction run under the conditions described. Stirring of the reaction mixture was accomplished by the use of a mechanical stirrer. Melting points of the compounds were obtained in sealed evacuated capillaries and are uncorrected. Either a Buchi or Mel-temp melting point apparatus was used to obtain the melting points. 'H NMR spectra were obtained on a Varian A-60, Varian XL-100, or Jeol PMX-60 spectrometer. Chemical shift is expressed in delta(6) values relative to internal tetramethylsi lane. 'H NMR results are expressed as chemical shift (splitting pattern, relative number of protons, 59

PAGE 67

60 coupling constant in hertz). Splitting pattern abbreviations, s, d, t, and m, correspond to singlet, doublet, triplet, and multiplet, respectively. Infrared spectra were obtained on a Beckman IR-10 instrument and mass spectra were obtained on an AEI MS-30 spectrometer. Microanalyses were obtained from Galbraith Laboratories, Inc., Knoxville, Tennessee; PCR, Inc., Gainesville, Florida; or Atlantic Microlab, Inc., Atlanta, Georgia. Iron analyses were obtained on a Perkin-Elmer 290B atomic absorption spectrophotometer. 19 n^-Anisole-n^-cyclopentadieny1 iron (I I) Hexafluorophosphate Ferrocene (10.67 g, 0.0575 mol), anisole (100 mL) , aluminum chloride powder (15.90 g, 0.119 mol), and aluminum powder (2.00 g, 0.07^1 mol) were added to the reaction vessel and heated to 1^0 C with stirring. At 110 C the solution turned black. However, no oil or solid coated the walls of the reaction flask as was normally observed. After 4 hours the mixture was cooled to 0°C and a 25^ aqueous methanol solution (130 ml) was added slowly to give a yellow aqueous layer. This was filtered and separated from the orange organic layer. The aqueous portion was washed twice with 30 mL portions of diethyl ether and then slowly added to a solution of ammonium hexafluorophosphate (3.0 g, 0.0126 mol) in 2 mL of water. The yellow precipitate which immediately formed was filtered and washed with water. This damp solid was dissolved in kO mL dichloromethane. After filtration the solution was added slowly to diethyl ether. A fine yellow powder was obtained which weighed 3.99 g (18.6%). m.p. 135-136°C (lit. m.p. 137. 5-138. 5°C) . ^'This reagent, used both as a solvent and reactant, was present in great excess with respect to the limiting reagent, ferrocene.

PAGE 68

61 NMR (acetone-d ) : cyclopentadlenyl , 5.15 (s,5); coordinated arene, 6 6.3^ (m,5); methyl, kM (s,3). IR (KBr): 3125 (s) , 2960 (w) , 1625 (w) , 1539 (s) , 1471 (s) , ]kk7 (s) , 1424 (s), 1388 (s), 1260 (s, broad), 1021 (s) , 995 (s) , 85O (vs, broad), -1 666 (m), 557 (s, broad), 490 (s) , 479 (s) , 462 (s) cm . 1 8 20 n^-Chlorobenzene-n^~cyclopentadienyl i ron(l l) Hexaf luorophosphate Ferrocene (10.25 g, 0.0551 mol), chlorobenzene (100 mL) , and aluminum chloride powder (23.84 g, 0.179 mol) were added to the reaction vessel and heated to 130°C with stirring. The solution turned black at 110°C. After 4 hours the solution was cooled to 0°C and 150 mL of ice water was slowly added. A green aqueous layer and an orange organic layer were obtained. The resulting mixture was filtered and the layers were separated. To the aqueous solution ascorbic acid was added (to reduce any ferricenium ion) and a yellow solution was obtained. The aqueous solution was refiltered, washed with diethyl ether (2 x 30 ml), and then slowly added to a solution of amnxsnium hexaf luorophosphate (5.00 g, 0.0321 nx3l) in 5 mL of water. The resulting yellow precipitate was washed with water to remove excess ammonium hexaf 1 uorophosphate and with diethyl ether to remove entrapped water and promote drying. The compound J was recrystal 1 ized from acetone and diethyl ether to give 5.91 g (28.3%) of a fine yellow powder, dec. 224-228°C*^ (no lit. m.p. reported). ^*This reagent was used both as sol vent and reactant and was present In large excess with respect to the limiting reagent, ferrocene. Aluminum powder was not added to this reaction in order to avoid the problem of dehalogenation of chlorobenzene and subsequent formation of T^S-benzene-n^-cyclopentadienyl iron(l I) cation along with the desired product . ^^•Product decomposed in this temperature range over 1.5 hours. The melting point for the PF,salt has not been previously reported.

PAGE 69

62 NMR (acetone-d ): cyclopentadienyl , 5-35 (s,5); coordinated arene, 6 6.^0-6.90 (m,5). IR (KBr): 3131 (s) , 3110 (s) , 150A (m) , lA45 (s) , ]k2k (s) , 1099 (s) , -1 8A8 (s, broad), 711 (s) , 555 (s), 510 (s) , 467 (s, broad) cm . 27 n^-Diphenylmethane-r)^-cyclopentadienyl i ron( I I ) Hexaf luorophosphate Ferrocene (19.45 g, 0.104 mol), diphenylmethane (34 mL, 0.183 mol), aluminum chloride powder (28.36 g, 0.212 mol), aluminum powder (3.10 g, 0.115 mol), and decahydron'aphthalene (95 ml) were added to the reaction flask and stirred at 145-150 °C for 4 hours. As the mixture was heated to the reaction temperature, a black oil formed on the sides of the vessel and the organic solution changed from orange to colorless. After 4 hours the reaction mixture was cooled to 0°C and 150 mL of 50% aqueous methanol was added slowly with concurrent loss of the black oil. The mixture was filtered and the aqueous layer was separated from the orange organic layer. The water layer was added slowly to an aqueous solution of ammonium hexaf luorophosphate (7.0 g, 0,044 mol). A yellow precipitate formed immediately and was collected by filtration and dried under a vacuum. The product was recrystal 1 ized from acetone and diethyl ether or dichloromethane and diethyl ether to give 17.9 g (38.61) of a fine yellow powder, m.p. 136-138°C (dec.) (no lit. m.p. reported). NMR (acetone-d,): cyclopentadienyl 5.22 (s,5), coordinated arene 6.47 (s,5), uncoordinated arene 7.37 (s,5), methylene 4.25 (s,2). 'If the product is not completely dry at this point, oiling will occur and good recrystal 1 ized product is almost impossible to obtain. If dichloromethane is used to dissolve the salt, any water remaining is easily removed by using a separatory funnel. The use of dichloromethane allows a more rapid work-up as it obviates the tedious step of completely drying out the product before recrystal 1 izat ion.

PAGE 70

63 IR (KBr): 3128 (m) , 1603 (w) , 1532 (w) , IA98 (m) . 1^68 (m) , 1^58 (m) , lif25 (s), 8i»5 (vs, broad), 7^0 (5), 710 (s) , 559 (s) , '97 (s) , 477 (s) cm . 27 n^-Fluorene-n^-cycl open tad ienyl i ron(l I) Hexaf luorophosphate Ferrocene (8.0 g, 0.0't3 mol), fluorene (10.0 g, O.O603 mol), aluminum chloride powder (11. 65 g, 0.0875 mol), aluminum powder (3.29 g, 0.122 a mol), and decahydronaphthalene {kO ml) were added to the reaction vessel and stirred at 155°C. At 120°C a black oil began forming and eventually completely coated the sides of the container. After 4 hours the stirring was stopped and the mixture was cooled to -78°C and 20 mL of 10^ aqueous methanol was added to the frozen mixture. The mixture was allowed to warm to O'C and 200 mL of water was added to complete the hydrolysis. The resulting orange organic layer and yellow aqueous layer were filtered and separated. The aqueous layer was added slowly to an aqueous solution of ammonium hexaf luorophosphate containing 3.2 g (0.0201 mol). The immediate yellow precipitate which formed was filtered, washed with water and dried under vacuum. The product was recrystal 1 ized from acetone and diethyl ether to give 6.3^+ g (3^.1%) of a yellow powdery product, dec. 163-165°C (no lit. m.p. reported). NMR (acetone-d ): cyclopentadienyl , 4.85 (s,5); coordinated arene, 6.35" 6 6.63 (m,2), 6.93-7.38 (m,2); uncoordinated arene, 7-38-7.82 (m,3) , 8.00.8.33 (m,l); exo CH, 4.23 (d,l ,J=23Hz) ; endo CH 4.50 (d, 1 ,J=23Hz) . IR (KBr): 3135 (w) , 2950 (broad, w) , 1612 (w) , 1437 (s) , 1422 (s) , -1 1388 (s), 840 (broad, vs) , 777 (s) , 730 (s) , 557 (s) , 496 (s) , 481 (s) cm 'Hexane was used as a solvent in one reaction. The reaction was run with a similar molar ratio of reactants in refluxing hexane. The product was worked up without cooling to -78°C and gave a 31.8^ yield.

PAGE 71

64 20 ri°-Hexamethylbenzene-ri -cyclopentadienyl i ron( I I ) Hexaf 1 uorophosphate Ferrocene (10.18 g, 0.05^7 mol), hexamethyl benzene (19.18 g, 0.118 mol), aluminum chloride powder (32.0 g, 0.2^0 mol), aluminum powder (5.00 g, 0.185 mol), and decahydronaphthalene (100 ml) were added to the reaco tion vessel and stirred at 150-160 C for 4 hours. At 115 C a brown viscous oil appeared on the walls of the container and remained until the reaction mixture was cooled to C and 100 mL of 20^ aqueous-methanol was slowly added. The mixture was filtered and the yellow aqueous layer was separated from the organic layer. The aqueous layer was then added to an aqueous solution of amnxsnium hexaf 1 uorophosphate (5.00 g, 0.0321 mol). The immediately-formed yellow precipitate was filtered, washed with water, and dried under vacuum. The product was recrystal 1 ized from acetone and diethyl ether to give 14.00 g (59.8^) of a yellow powdery product, dec. 283 C (complex originally reported as a BF,salt). 280 —° NMR (acetone-d,) : cyclopentadienyl , 4.74 (s,5); methyl, 2.58 (s,l8). o IR (KBr): 3014 (vw) , 2933 (vw) , 1448 (w) , I4l8 (s) , I382 (s) , IO76 (w) , -1 837 (vs, broad), 553 (s), 500 (w) , 461 (m) cm . 26 n^-Toluene-n "Cyclopentadienyl i ron( I I) Hexaf 1 uorophosphate Ferrocene (19-07 g, 0.103 mol), toluene (150 mL) , aluminum chloride powder (27.06 g, 0.203 mol), and aluminum powder (5.15 g, 0.191 mol) were added to the reaction vessel and stirred at 120 C. After 5 hours the reo action mixture was cooled to C and 100 mL of S% aqueous-methanol was slowly added. The resulting yellow aqueous and orange organic solutions were filtered and subsequently separated. The resulting solution was added to an aqueous solution of ammonium hexaf 1 uorophosphate (7.0 g, 0.0440 mol) a. This reagent was used as solvent and reactant and was present in excess,

PAGE 72

65 The immediate yellow precipitate which formed was filtered, washed with water, and dried under vacuum. The product was recrystal 1 ized from dichloromethane and diethyl ether to obtain 13-29 g (36.0?) of a fine yellow powder, dec. 286-289 °C^ (lit. m.p. 165 °C). 'H NMR (acetone-d ): cyclopentadienyl , 5.18 (s,5); coordinated arene, 6 6.40 (s,5); methyl, 2.58 (s,3). IR (KBr): 3122 (m) , 1530 (w) , 146A (s) , U21 (s) , 1393 (s) , 138? (s) , 843 (vs, broad), 552 (s), 495 (m) , 469 (s, broad) cm . 19 n^-Ani 1 ine-n^-cycl open tad ienyl i ron(l I) Hexaf 1 uorophosphate b Ferrocene (34.27 g, 0.184 mol), aniline (30.26 g, 0.163 nxal), aluminum chloride powder (94.00 g, 0.707 mol), aluminum powder (15.04 g, 0.557 mol), and decahydronaphthalene (l40 ml) were added to the reaction vessel and stirred at 200°C for 5 hours. During the heating a black oil coated the sides of the vessel and remained until the system was cooled to C and 200 mL of ]0% aqueous methanol was slowly added. The resulting orange aqueous layer, after filtering, was separated from the orange organic layer and washed twice with 30 mL portions of diethyl ether. The resulting solution was slowly added to a solution containing 10.0 g (0.0629 mol) of ammonium hexaf 1 uorophosphate. The immediate yelloworange precipitate which formed was filtered, washed with water and di"The melting point capillary was placed in the bath at 280 C. •The aniline was added last via an addition funnel very slowly to avoid a violent reaction with aluminum chloride. ^'Excess aluminum chloride (approximately four-fold excess with respect to the ferrocene) was used to insure that after the aniline was added and had reacted with the aluminum chloride there still remained sufficient aluminum chloride to react with the ferrocene.

PAGE 73

66 ethyl ether, and allowed to dry. The compound was recrystal 1 ized from acetone and diethyl ether to give 21.58 g (36.9^) of a powdery orange product, dec. 2'42°C (lit. dec. 250°C) . NMR (acetone-d ) : cyclopentadienyl , A. 98 (5,5); coordinated arene, 5.806 6.38 (m,5); NH , 5.57-5.80 (broad, s,2). IR (KBr): 3^99 (s) , 3^01 (s) , 32^8 (w) , 3135 (w) , 3096 (w) , I63A (s) , ISS'* (s), 1472 (s), 1421 (s), 1389 (s), 1305 (s) , 1162 (s) , 845 (vs, broad), 740 (s) , 556 (s), 475 (s), 461 (s) cm"'. Mass Spectrum, 150°C probe temperature (m/e, relative intensity): 372, 1.7; 345, 6.7; 307, 0.7; 28O, 5.8; 252, 21.0; 214, 0.4; 213, 2.0; 186, 50.4; 179, 100; 121, 30.6; 107, 34.6; 93, 32.6; 92, 15.0; 91, 39.0; 66, 62.5; 65, 53.2; 23.6, 64; 56, 17.9; 51, 8.2. n^-Carbazole-n^-cycl open tad ienyl iron (I l) Hexaf luorophosphate Carbazole (9-56 g, 0.0572 mol), ferrocene (5.23 g, 0.028 mol), aluminum chloride powder (7.70 g, 0.0579 mol), aluminum powder (2.19 g, 0.081 mol), and decahydronaphthalene (50 ml) were added to the reaction flask and stirred at 145°C for 4 hours. During the heating to reaction temperature the sides of the reaction vessel became covered with a black solid which prevented any observation of solution color. After 4 hours the system was cooled to 0°C and 60 mL of 16% aqueous-methanol solution was slowly added. The solid dissolved after the subsequent addition of *ln this particular reaction pure n^-ani 1 ine-n^-cyclopentadienyl i ron ( I I ) hexaf luorophosphate was obtained. However, in one reaction a mixture of iT-complexes of aniline and tetralin was obtained. These compounds were easily separated by column chromatography on alumina. After elution of the Tr-tetralin complex with dichloromethane, the ir-aniline complex was removed with acetone.

PAGE 74

67 100 mL of water and boiling the solution for 5 minutes. After the mixture was filtered, a yellow aqueous layer was separated from an orange organic layer. The aqueous layer was washed with 75 mL (3 x 25 ml) of diethyl ether and the resulting aqueous layer slowly poured into an aqueous solution containing 2.35 g (0.0144 mol) of ammonium hexaf luorophosphate. An immediate orange precipitate formed which was filtered, washed with water, and dried under vacuum. The product was recrystallized from dichlorome thane and diethyl ether to give 5-21 g (42.8^) of an orange powder, dec. 204-206 C. 'H NMR (acetone-d ): cyclopentadienyl , 4.62 (s,5); coordinated arene, 6.07" ^ a 6.42 (m,2); uncoordinated arene, 8.28-8.50 (m,l), 7.07-7-78 (m,5) ; NH, 10.38 (s,l, broad). IR (KBr): 3417 (s) , 3097 (m) , 1627 (s) , 1567 (s), 1498 (s) , 1443 (s) , 1376 (s), 1331 (s), 1249 (s) , 838 (s) , 759 (s) cm" . For C H .NFePF, calculated: C, 47.14^, H, 3.26%; N, 3.23%; Fe, 12.9%. 17 14 6 Found: C, 47.15%; H, 3.63%;N, 3.23%; Fe, 12.9%. Mass Spectrum, 150°C probe temperature (m/e, relative intensity): 288, 0.1; 186, 50.5; 168, 46.9; 167, 100; 166, 59-4; 139, 37.5; 121, 25.0; 107, 23.8; 85, 43.8; 66, 12.5; 65, 9.4; 56, 18. 8. n^-Diphenylamine-n^-cyclopentadienyl i ron(l I) Hexaf I uorophosphate Ferrocene (10.03 g, 0.0549 mol), diphenylamine (I8.3O g, O.IO8 mol), aluminum chloride powder (14.20 g, 0.107 mol), aluminum powder (4.34 g, O.I6O mol), and decahydronaphthalene (100 mL) were added to the reaction ^'overlapping multlplets of 2 coordinated and 3 undoordinated arene protons .

PAGE 75

68 vessel and stirred at 155"l80°C. During the course of the reaction a gray-green solid formation formed on the walls of the vessel. After 3.5 hours the mixture was cooled to 0°C and 80 mL of ]2% aqueous methanol was slowly added. The resulting mixture was filtered and the layers were separated. The orange, aqueous layer was added to a solution containing ammonium hexaf luorophosphate Ct.OO g, 0.0188 mol) in 10 mL of water. An orange precipitate formed immediately and was filtered, washed with water, and dried under vacuum. To increase the yield, the organic layer was washed with water (2 x 30 ml) and the washings were added to a solution of ammonium hexaf luorophosphate (l .00 g, 0.0063 mol) in water (5 mL) . The orange precipitate which formed was filtered, washed with water, and dried under vacuum. The two products were combined, dissolved in 60 mL of dichloromethane, and recrystal 1 ized by slow addition of the solution to 250 mL of diethyl ether. A yield of 1.37 g (5.8^) of orange needles, m.p. l80-l8l°C, was obtained after filtering and drying. NMR (acetone-d ) : cyclopentadienyl , ^.95 (s,5); coordinated arene, 5.856 6.35 (m,5); uncoordinated arene, 7.02-7.62 (m,5) ; NH, 7-87 (s,l, broad). IR (KBr): 3389 (s) , 3117 (w), 15^8 (s), U69 (m) , 1386 (s) , 839 (s, -1 broad), 557 (s) , ^67 (m) cm . Calculated for C H NFePF : C, ^6.92^; H, 3.7U; N, 3-22^, Fe, 12. 17 16 6 Found: C, 47. 16^;H ,3.88^;N, 3.16^; Fe, 12.2% Mass Spectrum, 150°C probe temperature (m/e, relative intensity): 290, >0.1; 186, k.8; 170, 13.1; 169, 100; 168, 75.0; 167, '»3.2; 121, 3.5; 83, 26.9; 66, 15-5; 65, 11.6; 51, 21.1 .

PAGE 76

69 n^~1 ,3>5"Tri isopropylbenzene-n^-cyclopentadienyl iron(l I) Hexaf luorophosphate Ferrocene (9.00 g, O.O^Sif mol), 1 ,3,5-tri isopropylbenzene (20.31 g, 0.0996 mol), aluminum chloride powder (23.00 g, 0.173 mol), aluminum powder (1.97 g, 0.07^^ mol), and decahydronaphthalene (75 mL) were added to the reaction vessel and stirred at 155 C. During the reaction a brown oil formed on the sides of the vessel. After 4.5 hours the reaction was stopped by cooling the mixture to C and slowly adding I50 mL of 20% aqueous methanol. A green aqueous layer and an orange organic layer were formed. The mixture was filtered and the aqueous layer was separated from the organic layer and slowly added to an aqueous solution of ammonium hexaf luorophosphate (5.00 g, 0.0321 mol). A yellow precipitate formed immediately and was filtered, washed with water and dried under vacuum. The product was recrystal 1 ized from acetone and diethyl 3 ether to give 10.41 g (45.8^) of a fine, yellow powder, dec. 264-265 C. 'H NMR (acetone-d,) ; cyclopentadienyl , 5.07 (s,5); coordinated arene, 6.27 (s,3); methyl, 1.42 (d,l8, J = 6.5 Hz); methine, 3.22 (m,3, J = 6.5 Hz). IR (KBr): 3125 (w, broad), 2967 (vs) , 2940 (s) , 2882 (m) , 1624 (w) , 1529 (s), 1466 (s), 1421 (s), 1384 (s) , 844 (vs, broad), 554 (s), 489 (w) , 452 (w) cm . Calculated for ^2Q^2<^^^^^f>' ^' 51-08^; H, 6.22^; Fe, 11.88^. Found: C, 51.16%; H, 6.26%; Fe, 11.6%. 'The sample was placed in the apparatus at 252 C and the yellow compound turned orange immediately. There was a slow loss of color, with concomitant oil vaporization until 264°C, at which time rapid decomposition occurred.

PAGE 77

70 . Mass Spectrum, \hoPC, probe temperature (m/e, relative intensity): 325, 0.5; 20i», 26.3; 190, 15.6; I89, 100; 186, 26.4; 161, 58. 1; 147, 12.0; 133, 14.1; 121, 15.1; 119, 15.2; 107, 25.7; 105, 25.9; 91, 36.7; 66. 11.1; 65, 13.3; 56, 12.0. n^-Thiophenol-n^"cyclopentadienyl i ron(l I) Hexaf luorophosphate n^-Chlorobenzene-n^-cycl open tad lenyl i ron(l I) hexaf luorophosphate (2.32 g, 0.00612 mol), sodium hydrogen sulfide (2.16 g, 0.047 mol), and acetonitrile (20 ml) were added to a dry 100 mL Schlenk tube against a countercurrent of nitrogen. Upon stirring an immediate reaction took place as evidenced by the formation of an intense red solution. After 20 minutes the solution was filtered and the blue-green residue was washed with four 30 mL portions of acetonitrile and filtered. The combined filtrates were taken to dryness under vacuum to give an intense red solid. An alumina column was prepared by dry packing the column and evacuating the resulting column to remove any entrapped oxygen. The product from the above reaction was dissolved in 100 mL acetone and then added to the column against a countercurrent of nitrogen. A yellow band eluted with acetone was discarded. An orange band remained on the column and was eluted with 30^ aqueous acetone. Acidification of this solution with concentrated hydrochloric acid resulted in a color change from red-orange to yellow. Excess ammonium hexaf luorophosphate was added and the yellow solution was taken to dryness under vacuum. The residue was recrystal 1 ized from acetone and diethyl ether and then dichloromethane and diethyl ether (both under nitrogen) to give 1.16 g (50.4^) of an olive green product, dec. 236238°C. 'H NMR (acetone-d ) : cyclopentadienyl , 5.20 (s,5); coordinated arene, 6 6.28-6.78 (m,5); SH 3.6-3.92 (s, broad, 1). IR (KBr): 3124 (m) , 2584 (w) , 1625 (w) , 1505 (w) , 1451 (m) , 1421 (s) ,

PAGE 78

71 1408 (m), 1385 (m), 1099 (m), 1011 (w) , 835 (vs, broad), 703 (w) , 552 -1 (s), 503 (m), A67 (s, broad ) cm . Calculated for C, H SFePF : C, 35.13?; H, 2.95%; Fe, ]k.S5%. 11 11 6 Found: C, 35.25%; H, 2.96%; Fe, 1^.41%. Mass spectrum, lAO^C probe temperature (m/e, relative intensity): 231, 0.2; 230, 1.5; 187, 9.7; I86, 72.4; 185, 10.9; 184, 10.3; 121, 39.8; 110, 93.4; 109, 36.4; 107, 100; 104, 39.9; 88, 27.5; 85, 34.2; 84, 19.6; 78, 11.3; 77, 21.3; 69, 25.1; 66, 76.6; 65, 49.2; 56, 22.5; 51, 21.2. n^-Phenol-n^-cyclopentadienyl i ron( I I ) Hexaf luorophosphate n^ -Ch lorobenzene-ri^ -eye 1 open tad ienyl i ron( I I ) hexaf luorophosphate (4.54g, 0.0120 mol) and sodium hydroxide (5.00 g, 0.125 mol) were added to a 100 mL round bottom flask that had been previously flushed with N . 2 After the addition of 50% aqueous acetone (8O mL) , the reaction mixture was stirred under nitrogen. During the reaction two layers formed, an orange aqueous layer and an intense brown organic layer. After 36 hours the organic layer was evaporated under vacuum without prior separation. With the evaporation of the organic layer substantial amounts of solid formed in the flask. The residual slurry was filtered to give an orange filtrate and a yellow solid which was washed with water. Acidication of the filtrate with concentrated hydrochloric acid produced a color change from orange to yellow and concomitant formation of a small amount of yellow solid. This mixture was then extracted with five 40 mL portions of dichloromethane. The fractions were combined and blown dry under N . ' 2 A brown, wet product was obtained which was dissolved in acetone, filtered, and precipitated by the addition of diethyl ether to give golden needles (2.24 g) . Further purification was achieved by dissolving the

PAGE 79

72 product in acetone and chromatographing It on an alumina column. A yellow band moved rapidly down the column with acetone and was discarded. An orange band remained at the top of the column. This was eluted with a 30% aqueous acetone solution and collected when it came off the column. The solution was acidified with concentrated HCl and then tal
PAGE 80

73 iodide (1.5 ml) was added. There was an initial formation of a small amount of yellow precipitate followed by a slower change in the color of the solution from red to yellow. The resulting solution was taken to dryness under vacuum to give an orange oil. The oil could not be crystallized by trituration with pentane, benzene, or diethyl ether. Furthermore, recrystal lizat ion using chloroform and pentane produced only an oil. The product was obtained by the addition of 0.1M aqueous silver nitrate to precipitate any silver iodide. There was an immediate formation of a white solid with a concomitant yellow solution. The solution was filtered and ammonium hexaf luorophosphate (0.50 g, 0.0031 mol) added to obtain an orange-yellow precipitate. This was extracted with dichloromethane (20 mL) and precipitated with pentane to obtain 0.38 g (29.7^) of an impure brown solid. Purification of the product was accomplished by dry packing the product on a 1 x 20 cm column of silica gel. The product was eluted as a broad yellow band with 250 mL of dichloromethane. This was blown dry under nitrogen to give orange needles. The orange needles were recrystal 1 ized from dichloromethane and pentane to give a flocculant orange powder, m.p. 153-156. 5°C. •H NMR (acetone-d ) : cyclopentadienyl , 5.11 (s,5); coordinated arene, 6 6.37-6.08 (m,3), 6.08-5.83 (m,2) ; methyl, 3.21 (s,6). IR (KBr); 3126 (m) , 2925 (w) , 1564 (s) , 1493 (w) , 1447 (m) , 1431 (m) , 1367 (m), 1235 (m), 1193 (m) , 1005 (w) , 83O (s, broad), 656 (m) , 556 (s) , 475 (m), 463 (m) cm' . For C H TeNPF, calculated: C, 40.3U; H, 4.13^; Fe, 14.5^. 13 16 6 Found: C, 40.23%; H, 4.1U; Fe, l4.7%.

PAGE 81

74 Mass Spectrum, llCC probe temperature (m/e, relative Intensity): 239, 15.2; 228, 0.2; 227, 1.8; l87, 13-5; l86, 100; ]8k, 10.9; 121, 99.4; 120, 91.9; 107, 56.5; 77, 17.5; 56, 8.4. n^-Thioanisole-Ti^-cyclopentadIenyl i ron ( I I ) Iodide To a dry, nitrogen filled Schlenk recrystal 1 Izat ion tube (l-5-n^cyclohexadienyl-6-thione)-ri^-cyclopentadienyl i ron( if) (0.51 g, 0.0022 mol) was added against a countercurrent of nitrogen. The complex was dissolved in chloroform (30 ml), filtered, and then reduced in volume under vacuum to 15 mL. To this solution methyl iodide (0.8 mL, 0.0084 mol) was added and stirred, producing a slow change in color from red to orange. After 30 minutes diethyl ether (80 ml) was added to produce an initial orange oil which slowly crystallized. After sitting for 1 hour, the yelloworange solid was filtered and dried under vacuum to give 0.58 g (70.8^), dec. 119-122°C. 'H NMR (chloroform-d) : cyclopentadienyl , 5.28 (s,5); coordinated arene, 6.65 (s,5); methyl, 2.81 (s,3). IR (KBr): 3073 (m) , 3035 (m) , I63O (w) , 1503 (m) , 1444 (s) , I4l8 (s) , 1387 (m), 1159 (m), 1146 (w), 1089 (s), 1006 (m) , 979 (w) , 854 (s) , -1 689 (m), 500 (m), 469 (s, broad) cm . Calculated for C H FeSI: C, 38.74^; H, 3-52^; Fe, 15.0^. Found: C, 38.65^; H, 3.42%; Fe, 15.2%. Mass Spectrum, 120"'C probe temperature (m/e, relative intensity): I87, 13.0; 186, 100; 124, 99.2; 123, 20.5; 121, 44.9; 109, 42.5; 91, 29.2; 65, 20.3; 56, 24.3.

PAGE 82

75 n^-Anisole-n^-cyclopentadienyl iron(l I) Iodide The title compound was prepared by treating the filtrate obtained from recrystall izing a sample of (l-5-n^-cyclohexadienyl-6-one) -n^-cyclopentadienyl irond I) f rom d ichloromethane and diethyl ether. Methyl iodide (1 mL, 0,017 mol) was added and the solution was stirred for 1 hour. The red solution slowly changed to a cloudy orange solution during this period. The solvent was then evaporated under a stream of nitrogen and the resulting orange residue was recrystal 1 ized from chloroform and diethyl ether, dec. 1A3-146°C. 'H NMR (chloroform-d) : cyclopentadienyl , 5.25 (s,5); coordinated arene, 6.50 (s,5); methyl, A.08 (s,3). n^-Carbazolyl-n ^-cyclopentadienyl i ron(l I) n^-Carbazole-n^-cyclopentadienyl iron(l I) hexaf luorophosphate (0.89 g, 0.00205 mol) and sodium amide {Q.Qk g, 0.0215 mol) were added to a dry, nitrogen filled 100 mL Schlenk tube. Ammonia (90 ml) was condensed into the system to produce an immediate reaction as seen by the formation of a dark red solution. After stirring for 3 hours the ammonia was evaporated and the resulting red solid was extracted with four 90 mL portions of benzene. These extracts were filtered and dried under vacuum to give 0.70 g (90^) of a red solid. The product was recrystal lized from benzene and diethyl ether at 0°C to give a fine red powder, dec. 176-178°C. 'H NMR (acetone-d ) : cyclopentadienyl, ^.13 (s,5); coordinated arene, 6.93-7.18 (m,2); 5.50-5.82 (m,2) ; uncoordinated arene, 8.07-8.27 (m,l), 7.27-7.80 (m,2), 6.65-6.93 (m,l). IR (KBr): I6l3 (m) , 15^4 (m) , 1A77 (m), 1^3^ (s) , I389 (s) , 1330 (s), 1287 (m), 123^ (s), 1007 (m) , 8kk (s) , 757 (s) , 559 (m) , i.28 (s) cm^

PAGE 83

76 Calculated for ^^j^^/^^^> ^3. 'it. Found: Fe, 19.8%. Mass Spectrum, 200°C probe temperature (m/e, relative intensity): 287, 0.1; 186, 17.9; 168, 13.4; 167, 100; 166, 22.4; 139, 11.1; 121, 8.9; 107, 12 1 U 56 13. 1; 83.5, 11.4. Parent peak exact mass calculated for C H N Fe, 287.0396; found, 287.0373. n ~Fluorenyl -n eye 1 open tad ienyl i ron( I I ) n^-Fluorene-n^-cyclopentadienyl i ron( M) hexaf luorophosphate (O.98 g, 0,00226 mol) and sodium hexamethyldis i lylamide (1.29 g, 0.00705 mol) were added to a dry, nitrogen filled 100 mL Schlenk tube. Dry diethyl ether (70 ml) was added and led to an immediate reaction as seen by the formation of a blue-green solution. The solution was stirred for 1.5 hours, then filtered and taken to dryness under vacuum. A deep green residue weighing 0.27 g (43%) was obtained. The compound was recrystal 1 ized from diethyl ether and pentane at -78°C, dec. 105-107°C. 'H NMR (acetone-d ): cyclopentadlenyl , 4.02 (s,5); coordinated arene, 6.75-7.13 (ni,2)^, 6.38-6.57 (m,l), 5.27-5.72 (m,2); uncoordinated arene, 8.08-8.33 (m,l), 7.13-7.68 (m,2); vinyl CH, 5.75 (s,l). IR (KBr): 1602 (s) , 1532 (s) , 1470 (s) , 1386 (s) , 1338 (s) , 1226 (s) , -1 1005 (m), 835 (s), 752 (s) . 472 (s) cm . Calculated for C,oH Fe: Fe, 19-5%. Found: Fe, 19.8%. a. overlapping multiplets of one coordinated and one uncoordinated arene ring proton.

PAGE 84

77 Mass Spectrum, I'tO'C probe temperature (m/e, relative intensity): 286, 1.0; 186, 69.8; 166, /»9. 7; 165, lOO; 121, 33.3. Parent peak exact mass calculated for C „ H,, ^^Fe, 286.0^*^4; found, 286.0^35. 18 IM (l-5"n-6Pi phenyl me thylenecyclohexadienyl ) -ri ^ -cy cl open tad ienyl i ron( I ! ) n^-Triphenylmethane-n^ -eye 1 open tad ienyl i ron(l l) hexaf 1 uorophosphate (0.5^^ g, 0.00103 mol) and sodium amide (0.40 g, 0.0106 mol) were added to a dry, N filled, 100 mL Schlenk tube. Ammonia (90 ml) was condensed 2 into the Schlenk tube to produce an immediate reaction as seen by the formation of an intense red solution. The solution was stirred for h hours with no further visible changes. The ammonia was then allowed to evaporate leaving a red solid. The solid was extracted with three 90 mL portions of diethyl ether, filtered, and then dried under vacuum to yield 0.3^ g (83%) of a brick-red product. The product was recrystallized at -78°C using ether and pentane to give a red powder, dec. 115117°C. 'H NMR (acetone-d ) : cyclopentadienyl , 4.55 (s,5); coordinated arene, 6 ^.12-4. 38 (m,2); 4.82-5.11 (m,2) , 5.9^-6.25 (m,l); uncoordinated arene, 6.88-7.46 (m,10). IR(KBr): 3058 (m) , 1588 (s) , 1423 (s) , 1270 (s) , 837 (s) , 758 (s) , 697 -1 (s), 598 (s), 468 (m) , 303 (s) cm . Calculated for C H Fe: Fe, 15.4%. 24 20 Found: Fe, 15.2% Mass Spectrum, 130°C probe temperature (m/e, relative intensity): 364, 93.0; 298, 23.3; 243, 60.4; 186, 21.5; 165, 100; 121, 39.5. Parent peak exact mass calculated for C H„„ Fe, 364.0910; found, 364.0920. 24 20

PAGE 85

78 (1 -5-Tf-6-Pheny1 iminocyclohexadienyl ) -n^~cyc1opentadieny1 i ron ( I I ) ri^D i phenyl am ine-n^ -eye 1 open tad I enyl i ron( I I ) hexaf luorophosphate (0.56 g, 0.00128 mol) and sodium amide (0.75 g, 0.0192 mol) were added to a dry, nitrogen filled, 100 mL Schlenk tube. Ammonia (90 mL) was condensed into the Schlenk tube to produce an immediate reaction as seen by the formation of an intense red solution. The solution was stirred for 3 hours with no further visible change and then the ammonia was allowed to evaporate leaving a red-orange solid. This was extracted with two 90 mL portions of benzene, filtered, and dried under vacuum. The resulting product was recrystal 1 ized from benzene and pentane at 0°C to give 0.31 g (85%) of a red powder, dec. 122-12't°C. 'H NMR (acetone-d ): cyclopentadienyl , k.kS (s,5); coordinated arene, 6 4.58-A.82 (m, 2-under cyclopentadienyl peak), 5.17-5-55 (m,2) , 5.57-5.83 (m,l); undoordinated arene, 6.72-7.00 (m,3); 7.05-7-^2 (m,2) . IR(KBr): 3066 (m) , 1596 (m) , 1562 (s) , ]k3k (s) , 1^75 (s) , ]k2] (m) , 1386 (s), 1217 (m), 81A (m), 76A (m) , 717 (m) , 667 (m) , ^98 (m) , kSk (m) , A66 (m) cm . Calculated for C H NFe: Fe, 19.3%. 17 15 Found: Fe, 19.3%. Mass Spectrum (no heat) (m/e, relative intensity): 289, 1.^; 186, 31.5; 169, 100; 168, 55.9; 167, 37.3; 121, 15.0. Parent peak mass calculated 12 1 lit 56 „ for C H N Fe, 289.0552; found, 289.0552. (1 -5~n -Cyc1ohexadienyl-6-one) -n -cyclopentadienyl i ron (I l) n -Chi orobenzene-n^ -eye 1 open tad ienyl i ron( I I ) hexaf luorophosphate (5.A1 g, 0.0135 mol) and potassium hydroxide (4.18 g, 0.07^5 mol) were

PAGE 86

79 added to a nitrogen-flushed 250 mL round bottom flask. To the reactants 120 mL of 50% aqueous acetone was added to produce a yellow-brown solution which slowly turned orange-brown over 1 hour. After the reaction mixture was stirred for l8 hours under nitrogen, the acetone was removed by evaporation under vacuum. The mixture was rapidly filtered in air to remove Insoluble matter. An orange solution was obtained which was pumped to dryness under vacuum at room temperature to give a damp brown solid. The solid was extracted with two 100 mL portions of dichloromethane, filtered, and then taken to dryness under vacuum. The resulting orange solid was recrystal 1 Ized from dichloromethane and ether to give 2.39 g (82.7^) of an orange powder, dec. 112-llA°C. 'H NMR (acetone-d ) : cyclopentadienyl , 4.57 (s,5), coordinated arene, 6 4.67-4.97 (m,2), 5.33-5.63 (m,3). IR (KBr): 3060 (w) , 1661 (m) , 1535 (s, broad), 1471 (s) , I4l7 (m) , 1387 (m), 1346 (s), 1142 (m) , 1112 (m) , 1047 (w) , 1005 (m) , 841 (s) , 689 (s), 542 (s), 476 (s) cm"V Calculated for C H FeO: Fe, 26.]%. 11 10 Found: Fe, 26.6^. Mass Spectrum, 110°C probe temperature (m/e, relative intensity): 214, 36.2; 186, 49.4; 121, 36.3; 94, 100; 66, 24.3; 65, 28.7; 56, 19.3; 51, 85. Parent peak exact mass calculated for C H Fe , 214.00800; 11 10 found, 214.00791 . ( 1 -5-n^ -Cyc 1 ohexad ienyl -6thi one) -n^ -cyclopentadienyl i ron ( I I) n -Chlorobenzene-n^-cyclopentadienyl i ron (II) hexaf luorophosphate (7.07 g, 0.0187 mol), sodium hydrogen sulfide (4.19 g, 0.0748 mol), and

PAGE 87

80 acetonitrile (70 ml) were added to a dry 100 mL Schlenk tube against a countercurrent of nitrogen. Upon stirring, an intense red solution developed. After 20 minutes the solution was filtered and the remaining blue-green solid was washed with ten 20 mL portions of acetonitrile. The combined filtrates were reduced in volume to approximately 150 mL and 100 g of alumina was added against a countercurrent of nitrogen. The mixture was then taken to dryness under vacuum. This alumina was then added, against a countercurrent of nitrogen, to a 5 x ^0 cm column previously packed with alumina dried under vacuum. A small yellow band was eluted with acetone and was discarded. The desired product was eluted as a broad red-orange band using 20^ aqueous acetone. The red solution was dried under vacuum to give a red-brown product. To remove the last traces of water it was necessary to heat to A5 C . This product was recrystallized from dry dichloromethane and dry diethyl ether. The resulting red-orange powder was dried under vacuum to give 2.38 g (55.3^) • Darkens at 125-128°C. Decomposes at 183°C. 'H NMR (acetone-d.) : cyclopentadienyl , 5.02 (s,5); coordinated arene, 6.00-6.37 (s, broad, 3), 6.37-6.8A (s, broad, 2). 'H NMR (chloroform-d) : cyclopentadienyl, A. 70 (s,5); coordinated arene, 5.37-6.00 (s, broad, 3), 6.00-6.3^ (s, broad, 2). IR (KBr): 3025 (w) , 1628 (w) , 1^87 (s) , l4l8 (s), 1399 (m) , 1383 (m) , 1082 (s), 8A2 (s), 711 (m), 6Al (m) , A63 (s) cm' . Calculated for C-.H^^FeS: Fe, 2^.3^. Found: Fe, lh.0%.

PAGE 88

w Mass Spectrum, 100°C probe temperature (m/e, relative intensity): 230, 3.5; 186 (C, H S), 85.1; 186 (C H Fe) , 100; I85 (C H S) , 56.0; \8k 12 10 10 10 12 9 (C H-S), 36.1; ]5h, 20.5; 152, 10.0; 121, 31. A; 110, 15.9; 77, 1^.2; 12 o 1 2 1 66, 52.1 ; 65, 31.5. Parent peak exact mass calculated for C H 56 32 '' '' Fe S, 229.9852O; found, 229.98663. (l~5~n^-6-|minocyclohexadienyl)-ri^-cyclopentadienyl I ron ( I I) a. Reaction in Dichloromethane n^-Ani 1 ine-ri^-cyclopentadienyl i ron(l 1) hexaf 1 uorophosphate (1.05 g, 0»00292 mol), sodium amide (l.3'» g, 0.03'»'t mol), and dichloromethane (90 mL) were added to a dry, 100 mL Schlenk tube against a countercurrent of nitrogen. Stirring this solution produced an intense red color. After 1.75 hours the solution was filtered and taken to dryness under vacuum. The residue was recrystal 1 ized from dichloromethane and ether to give 0.49 g (79.1^) of a red solid, dec. ]hO-]hS°C. 'H NMR (acetonitri le-d ): cyclopentadienyl , 4.53 (s,5); coordinated 3 arene, 5.30-5.63 (m,3) , 4.83-5.05(m,2) ; NH, not observed. IR (KBr): 3^58 (m) , 3^00 (m) , 3100 (m) , I63I (m) , 1551 (s) , 1469 (s), 1469 (m), 1342 (m) , 1149 (w) , IOO5 (w) , 838 (s, broad), 661 (m) , 557 (s) , -1 472 (m, broad) cm Mass Spectrum, SO'C probe temperature (m/e, relative intensity): 372, 0.1; 307, >0.1; 213, 4.5; 186, 35.7; 121, 28.8; 93, 100; 92, 12.1; 66, 89.6; 65, 50.5; 56, 16.3. Exact mass calculated for C ^H ^^Fe 372.02620; found, 372.02633. Exact mass calculated for ^^C 'h ^^Fe , 306.9871 ; found, 306.9894. Parent peak exact mass calculated for 12 1 ]L 56 Cj^ Hj^ '^N Fe, 213.02390; found, 213.02352.

PAGE 89

82 b. Reaction in Ammonia n -Ani 1 ine-n^-cyclopentadienyl iron(li) hexaf luorophosphate (2.0^ g, 0.00568 mol) and sodium amide {k.G^ g, 0.119 mol) were added to a dry, nitrogen filled 100 mL Schlenk tube against a countercurrent of nitrogen. Ammonia (90 mL) was condensed into the system to produce an initial intense red solution which turned black. After 3 hours the ammonia was evaporated to leave a tacky black solid. Extraction of this solid with dichlorome thane produced only a pale yellow solution. c. Reaction in Tetrahydrofuran n^-Ani 1 ine-ri^-cyclopentadienyl i ron( I I ) hexaf 1 uorophosphate (1.09 g, 0.00301 mol) and sodium amide (2.00 g, 0.0513 mol) were added to a dry, nitrogen filled, 100 mL. Schlenk tube against a countercurrent of nitrogen. Tetrahydrofuran (90 mL) was added to produce an intense red solution which then turned black in less than ten minutes. The sample was not worked up. d. Reaction in Acetonitrile Acetonitrile (80 mL) was added to a dry, nitrogen filled, 250 mL round bottom flask against a countercurrent of nitrogen. To the acetonitrile sodium metal (0.90 g, 0.039 mol) was slowly added to produce rapid bubbling and a white precipitate. After 10 minutes the bubbling stopped and a pale yellow solution was noted. n^-Ani 1 ine-n^-cyclopentadienyl iron(ll) hexaf luorophosphate (2.53 g, 0.00707 mol) was added to the mixture resulting in the formation of an intense red solution. After stirring for 3 hours, the solution was filtered and reduced in volume to 10 mL under vacuum. Diethyl ether (300 mL) was added to obtain a red-orange precipitate. This was filtered and dried under vacuum and was identical with the compound prepared in CH2CK.

PAGE 90

83 Reaction of n^-Diphenylmethane-ri^-cyclopentadienyl i ron ( I I ) Hexafluorophosphate with Sodium Amide n^-Diphenylmethane-n^-cyclopentadienyl Iron(l I) hexaf 1 uorophosphate (1.06 g, 0.00245 mol) and sodium amide (1.07 g, 0.027^ mol) were added to a dry, nitrogen-filled, 100 mL Schlenk tube against a countercurrent of nitrogen. The system was sealed and ammonia (90 mL) was condensed into the system to give an intense red solution. After stirring for 2 hours, the ammonia was evaporated and the resulting red residue was dried under vacuum. The mixture was then extracted with four 100 mL portions of diethyl ether. These were filtered, combined and taken to dryness to give 0.25 g (36^) of a red-brown powder. The product was recrystal 1 ized at -78°C from diethyl ether and pentane, dec. 66-68°C. 'H NMR (acetone-d ) : cyclopentadienyl , h.OO (s,2), 4.13 (m,3) ; arene, 6 7.27 (s,10); methlne, 3.83 (s,l). IR (KBr): 3021 (w) , 2903 (vw) , 1597 (m) , 1576 (m) , 1490 (m) , 1448 (m) , 1411 (m), 1383 (s), 1259 (m), 1105 (s), 998 (m) , 8o4 (s) , 698 (s) , 470 (m) -1 cm Regeneration of n^-Diphenylmethane-ri^-cyclopentadienyl i ron( I I ) Cation A sample of the deprotonated ir-d i phenyl methane complex was prepared as described previously. To the filtered red solution of the complex in benzene 4 M hydrochloric acid (4 mL) was added. There was an immediate reaction to give a yellow solution. This was filtered and the product was precipitated by the addition of ammonium hexaf 1 uorophosphate. The resulting yellow precipitate was filtered and dried. The 'H NMR spectrum of this compound was identical to that of the starting material.

PAGE 91

8k Attempted Reaction of Methyl Iodide with Certain tt -Are ne-ireye) open ta dieny] i ron( II) Sal ts a. n^-Ani 1 ine-n^-cyclopentadienyl i ron(l I) Hexaf luorophosphate n^-Ani 1 ine-ri^-cyclopentadienyl irond I) hexaf 1 uorophosphate (0.30 g, (0.0008A irol) and methyl iodide (0,46 g, 0.0032 mol) were stirred together for 30 minutes in dichloromethane (15 ml). The solvent was then removed under vacuum and 'H NMR spectral analysis of the resulting showed it to be exclusively the starting cation. b. n^-Phenol-n^-cyclopentadienyl irond I) Hexaf luorophosphate n^-Phenol-n^-cyclopentadienyl i rond I) hexaf luorophosphate (0.10 g, 0.00028 mol) was stirred with methyl iodide (2.0 g, O.OU mol) in dichloromethane (5 ml) for 'tO minutes. The solvent was then removed under vacuum and 'H NMR spectral analysis of the resulting residue showed it to be exclusively the starting cation. c. n^-Thiophenol-n^-cycl open tad ienyl iron (I I) Hexaf 1 uorophosphate n^-Thiophenol-n^-cyclopentadienyl irond I) Hexaf 1 uorophosphate (O.IO g, 0.00027 mol) was stirred with methyl iodide (2.0 g, 0.014 mol) in chloroform (5 mL) for kO minutes. The solvent was then removed under vacuum and 'H NMR spectral analysis of the resulting residue showed i t to be exclusively the starting cation. Preparation of n^-ani 1 inium-n^-cyclopentadienyl i ron( I I ) Cation a. From n^-Ani 1 ine-n^-cyclopentad ienyl i ron( II ) Hexaf luorophosphate n^-Anil ine-n^-cycl open tad ienyl irond I) hexaf 1 uorophosphate (approximately 0.05 g) was dissolved in acetone-d, (0.50 mL) and filtered into a 'H NMR tube. Acetic acid-d (0.2 ml) was added to the solution and the 'H NMR spectrum of the sample was obtained.

PAGE 92

85 'H NMR (acetone-d ) : cyclopentadienyl , 4.85 (s,5, broad); coordinated 6 arene, 5.85 (s,5, broad); N-H, S.'jO (s,2, broad). b. From (l -5-n^-6-lminocyc1ohexadienyl ) -ri^-cyclopentadienyl iron (I I) (l-5-n^-6-lminocyclohexadienyl)-n^-cyclopentadieny1 iron(l I) (approximately 0.10 g) was dissolved in acetone-d (0.50 ml) and filtered into a 'H NMR tube. Acetic acid-d (0.2 mL) was added to produce an immediate change from a red to an orange solution. A 'H NMR spectrum was obtained of this solution. 'H NMR (acetone-d ) : cyclopentadienyl, 4.85 (s,5, broad); coordinated 6 arene, 5.85 (s,5, broad); N-H, 9-90 (s,l, broad). Preparation of n^-Benzene-n^-cyclopentad ienyli ron( I I ) Hexaf luorophosphate in Ethylpyr id in ium Bromide-Aluminum Chloride Eutectic Ferrocene (10. g, 0.0538 mol), aluminum powder (2.00 g, 0.074 mol), and benzene (100 ml) were added to a 250 mL 3-neck round bottom flasl< equipped with a mechanical stirrer and a condenser. The flask had been previously charged with approximately 100 mL of the ethyl pyridinium a bromide-aluminum chloride eutectic. The brown solution was heated with stirring at 80°C and as the temperature approached 80°C, the solution became black. After 3 hours the system was cooled and then rapidly poured over 200 g of ice. A vigorous exothermic reaction occurred and was followed by the formation of an orange organic layer and a yellow aqueous *Koch, Miller, and Osteryoung reported that the preparation of the eutectic was accomplished by the addition of ethylpyr idinium bromide to aluminum chloride in a 1:2 mole ratio. For the preparation of the eutectic in this experiment ethyl pyridinium bromide (59 g, 0.31 mol) was carefully mixed with aluminum chloride (65 g, 0.49 mol) to produce a brown liquid. The mole ratio was considerably less than that reported by Koch et_ a_l_. it was found that addition of the full two equivalents of aluminum chloride resulted in the solidification of the eutectic.

PAGE 93

86 layer. After filtration of the mixture, the aqueous layer was separated from the organic layer and washed with two ^0 mL portions of diethyl ether. The aqueous solution was then slowly added to an aqueous solution of ammonium hexaf luorophosphate ('.O g, 0.025 mol ) . A yellow precipitate formed Immediately and was filtered and dried under vacuum. The product was recrystal 1 ized from acetone and diethyl ether to give \h.7 g of a pale yellow solid. 'H NMR showed the solid to be a mixture of the expected product and ethylpyridinium bromide. The mixture could not be separated by recrystal 1 izat ion from acetone, dichloromethane, or acetonitrile. Chromatographic separation of a column prepared with either alumina or silica gel was not possible either. Under no conditions could any discernable difference in R values be discerned. Reaction of Hydroiodic Acid with ri^~Anisole-n^-cyclopentadieny1 i ron ( ll) Hexaf 1 uorophosphate n^-Anisole-n^-cyclopentadienyl i ron( I I) hexaf luorophosphate (0.35 g, 0. COOS'* nx)l) and hydroiodic acid (20 mL, 57?) were added to a 50 mL nitrogen-flushed, round bottom flask. The mixture was heated to 150°C for kS minutes. it was then cooled and neutralized with solid sodium carbonate and the resulting yellow solid was collected. A 'H NMR spectrum showed the solid to contain only starting material. In a similar reaction in which the mixture was kept at 150°C for 16 hours decomposition occurred as evidenced by the formation of a brown solid. When the system was worked up in a fashion similar to that above, neither the starting material nor the desired Tr-phenol complex was .obtained. Attempted Synthesis of n^-Phenol-n^-cyclopentadienyl i ron( I!) Hexafluorophosphate Ferrocene (10.33 g, 0.0555 mol), phenol (50.5 g, 0,537 mol), aluminum

PAGE 94

87 chloride powder (15.17 g, 0.114 mol), and aluminum powder Cj.SS g, 0.168 mol) were added to the reaction flask against a countercurrent of nitrogen. A vigorous reaction occurred immediately upon addition of the regents to produce a green, grainy solution. Heating the mixture to 100°C caused continued bubbling and the solution became a black viscous oil. After 1 hour the reaction was cooled to 0°C and 100 mL of water was added. A black, oily organic layer and a colorless aqueous layer was obtained. No further workup was attempted. Attempted Synthesis of n^-Ani 1 ine-n^-cyclopentadienyl i ron( I I) Hexafluorophosphate Using Anilinium Sulfate Ferrocene (10.^2 g, O.O56 mol ) ,ani 1 inium sulfate (27.92 g, 0.109 mol), aluminum chloride powder (14. Al g, O.IO8 mol), aluminum powder (6.00 g, 0.222 mol), and decahydronaphthalene (120 ml) were added to a reaction flask against a countercurrent of nitrogen and stirred. The reaction mixture was rapidly heated to 155°C with the original pale green solution turning black at 130°C and forming a black oil on the sides of the reaction flask. After 1 hour at 155°C the black oil solidified and broke away from the sides of the flask to reveal that a white insoluble material was also present. After 3 hours the reaction was cooled to 0°C and 150 mL of ^5% aqueous-methanol solution was slowly added to produce an orange organic layer and a green aqueous layer. After filtration of the mixture, the aqueous layer was separated from the organic layer and washed with 50 mL of diethyl ether. The aqueous solution was then added to an aqueous solution of ammonium hexaf luorophosphate (4,0 g, 0.025 mol). The green precipitate which formed immediately was then filtered and dried under vacuum to give 1.28 g of a green powder. 'H NMR spectra showed the product to be the iron cyclopentadienyl complex of tetralin.

PAGE 95

56 Sodium Hexamethyldis i lylamide Sodium amide (11.3 g, 0.290 mol), hexamethyldisl iazane (hS.O g, 0.298 mol), and benzene (150 ml) were added to a dry three-neck 250 mL flask and stirred under nitrogen for 60 hours at 78°C, After 60 hours no ammonia was detected at the pressure release bubbler using damp red litmus paper. The hot solution was then rapidly filtered through filter paper and a yellow filtrate was collected. It was necessary to wash the solid remaining on the filter paper with hot benzene to obtain the maximum of product. The combined filtrates were evaporated under vacuum to give A3.0g(8l.O^) of a white powder. 57 Sodium Hydrogen Sulfide Sodium metal (12.61 g, 0.5^8 mol) and absolute ethanol (200 ml) were added against a countercurrent of nitrogen to a dry 3~neck 1 L round bottom flask equipped with a condenser and a gas dispersion tube. Addition of the ethanol produced vigorous bubbling from H evolution and refluxing ethanol. After 45 minutes the sodium had completely reacted to give a colorless solution. Hydrogen sulfide (26.0 g, 0.706 mol) was then slowly bubbled into the system. After 2 hours the hydrogen sulfide gas was stopped and 750 mL of dry diethyl ether was added. An immediate fine, white precipitate which formed was subsequently filtered under nitrogen and washed with diethyl ether to give 23.2 g (75. 4^) of sodium hydrogen sulfide.

PAGE 96

REFERENCES 1. T. J. Kealy and P. L. Pauson, Nature . 168 . 1039 (1951). 2. a. M. L. H. Green, "Organometal 1 ic Compounds," Vol. II, "The Transition Elements," 3rd Ed., Methuen and Co., Ltd., London, 1968. b. J. Organometal. Chem . . 100 (1975). c. R. F. Heck, "Organotrans i tion Metal Chemistry, A Mechanistic Approach," Academic Press, New York, 197^. 3. C. W. Bird, "Transition Metal Intermediates in Organic Synthesis," Logos Press, London, 1967. k. H. Alper, Ed., "Transition Metal Organometal 1 ics in Organic Synthesis," Vol. I, Academic Press, New York, 1976. 5. H. C. Longuet-Higgins and L. E. Orgel , J. Chem. Soc , 1959 (1969). 6. 0. L. Chapman, C.-C. Chang, J. Kale, M. E. Jung, J. A. Lowe, T. J. Barton, M. L. Tumey, J. Amer. Chem. Soc , 98 , 784^ (1976). 7. H. Sakurai, Y. Kamiyama, Y. Nadadaira, J. Amer. Chem. Soc , 98 , 7^53 (1976). 8. E. 0. Fischer and K. Ofele, Chem. Ber .. 90, 2532 (1957). 9. H. J. Dauben, Jr. and L. R. Honnen, J. Amer. Chem. Soc , 80 , 5570 (1958) 10. A. Steitwieser, Jr., and U. Mul ler-Westerhof f , J. Amer. Chem. Soc , 90, 7364 (1968). 11. G. Wilkinson, M. Rosenblum, M. C. Whiting, and R. B. Woodward, J_. Amer. Chem. Soc , 74, 2125 (1952). 12. H. H. Zeiss and W. Herwig, J. Amer. Chem. Soc , 78 ,5959 (1956). 13. a. T. H. Coffield, V. Sandel, and R. D. Closson, J. Amer. Chem. Soc , 79, 5826 (1957). b. T. H. Coffield and R. D. Closson, U. S. Patent 3130214 (1964) Chem. Abstr ., 6J_, 4397g (1964). 14. M. L. H. Green, L. Pratt and G. Wilkinson, J. Chem. Soc , 989 (I960). 15. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl . Akad . Nauk SSSR , 149 , 615 (1963). 16. D. Astruc and R. Dabard, J. Organometal. Chem ., Ill , 339 (1976).

PAGE 97

90 17. R. G. Sutherland, J. Organometal . Chem. Library , 3, 311 (1977). 18. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl . Akad . Nauk SSSR , 166 , 607 (1965). 19. A. N. Nesmeyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl . Akad . Nauk SSSR , 175 , 606 (1967). 20. I. U. Khand, P. L. Pauson, and W. E. Watts, J. Chem. Soc , C, 2261 (1968). 21. D. Jones, L. Pratt, and G. Wilkinson, J. Chem. Soc . '^58 (1962) 22. J. F. Helling and G. Cash, J. Organometal . Chem . , 73, CIO (1973). 23. E. 0. Fischer and H. Brunner, Z. Naturforsch . , B^. , ]S^, kOG (1961) 2^. A. N. Nesmeyanov, N. A. Vol'kenau, and E. i. Sirotkina, Izv. Akad . Nauk SSSR , Ser. Khim ., 1170 (1967). 25. E. 0. Fischer and W. Hafner, Z. Naturforsch ., 106 , 665 (1955). 26. D. Astruc and R. Dabard, C. R. Acad. Sci . Ser. C , 272, 1337 (1971). 27. C. C. Lee, R. G. Sutherland, and B. J. THompson, J. Chem. Soc , £. , 907 (1972). 28. I. U. Khand, P. L. Pauson, and W. E. Watts, J. Chem. Soc , £. , 2257 (1968). 29. V. R. Koch, L. L. Miller, and R. A. Osteryoung, J. Amer. Chem. Soc , 98, 5277 (1976). 30. J. F. Helling and W. A. Hendrickson, J. Organometal. Chem ., lAl , 99 (1977). 31. R. L. Burwell, Jr., Chem. Rev ., 5^, 615 (195^*). 32. D. E. Games, A. H. Jackson, L. A. P. Kane-Magui re, and K. Taylor, J. Organometal . Chem ., 88, 3^5 (1970). 33. J. W. Kang and P. M. Maitlis, J. Organometal . Chem . 30, 127 (1971). •34. A. Wu, E. R. Biehl, and P. C. Reeves, J. Chem. Soc , Perkin M , kks (1972). 35. W. S. Trahanovsky and R. J. Card, J. Amer. Chem. Soc , 9A, 2897 (1972). 36. G. Jaouen, A. Meyer, and G. Simmonneaux, J. Chem. Soc , £. ,813 (1975). 37. P. M. Treichel and J. W. Johnson, J. Organometal . Chem ., 88, 207 (1975) 38. C. White, S. J. THompson, and P. M. Maitlis, J. Chem. Soc , £. , 409 (1976).

PAGE 98

91 39. P. L. Pauson and J. A. Segal, J. Chen. Soc , D.. , 1677 (1975). kO. J. E. Sheets, W. Miller, and T. Klrsch, J. Organonietal . Chem ., 91 , 97 (1975). Al. J. E. Sheats, W. Miller, M. D. Rausch, S. A. Gardner, P. S, Andrews, and F. A. Higbie, J. Organometal . Chem ., 96 , 115 (1975). k2. R. Markby, H. W. Sternberg, and I. Wender, Chem. and Ind . (London), 1381 (1959). 43. J. W. Johnson and P. M. Treichel, J. Chem. Soc , £. , 688 (1976). kk. P. M. Treichel and J. W. Johnson, Inorg. Chem ., 16, 7^9 (1977). kS. J. W. Johnson and P. M. Treichel, J. Amer. Chem. Soc , 92» 1^27 (1977). kS. D. J. Cole-Hamilton, R. J. Young, and G. Wilkinson, J. Chem. Soc , D, 1995 (1976). ^7. W. S. Trahanovsky and R. A. Hall, J. Amer. Chem. Soc , 99, ^850 (1977). 48. C. White, S. J. Thompson, and P. M. Maitlis, J. Organometal. Chem ., 127 , 415 (1977). 49. A. N. Nesmeyanov, N. A. Vol'kenau, L. S. Shilovtseva, and N. A. Petrakova, J. Organometal . Chem ., 61 , 329 (1973). 50. a. J. W. Fitch, III and J. J. Lagowski , J. Organometal. Chem ., S_, 480 (1966). b. G. Hunter, E. 0. Fischer, and C. Elschenbroich, J. Organometal . Chem . , I, 330 (1965). c G. Hunter, E. 0. Fischer, R. D. Rischer, 0. L. Carter, A. T. McPhail, and G. A. Sim, J. Organometal . Chem ., 6_, 288 (I966). d. H. Kobayashi, M. Kobayashi, and Y. Kaizu, Bull . Chem. Soc . Jap ., 46, 3109 (1973). 51. R. R. Schrock and J. A. Osborn, Inorg. Chem ., % 2339 (1970). 52. a. R. Hoffmann and P. Hoffman, J. Amer. Chem. Soc , 98, 598 (1976). b. R. C. Haddon, Aust. J. Chem ., 30, I (1977) . 53. H. Werner, Angew. Chem ., _I6^, I (1977). 54. K. Nakamoto in "Characterization of Organometal 1 ic Compounds, Part I," M. Tsutsui, Ed., Interscience Publishers, New York, I969, P126. 55. M. Mathew and G. Palenik, Inorg. Chem . _!_!_, 2809 (1972). 56. C. P. Kriiger and H. Niederprum in "Inorganic Synthesis," Vl I I , H. F. Holtzclaw, Jr., Ed., McGraw-Hill, New York, I966, p. 15. 57. R. E. Eibeck in "Inorganic Synthesis," VI I , J. Kleinberg, Ed., McGraw-Hill, New York, 1963, p., 128.

PAGE 99

BIOGRAPHICAL SKETCH William A. Hendrickson was born on 2 July 1952 In Shreveport, Louisiana. In June 1970 he graduated from John F. Kennedy Senior High School in Bloomlngton, Minnesota. In May 197^ he received a Bachelor of Arts degree with majors in math and chemistry from Saint Olaf College In Northfield, Minnesota. In September 197^ he enrolled in the Graduate School of the University of Florida where he studied under a Graduate School Fellowship (197^-1975) and teaching ass istantships from both the Department of Chemistry and the Graduate School (1975-1978) toward the degree of Doctor of Philosophy. William A. Hendrickson married the former Lee Joan Davidson on 17 August 197A. 92

PAGE 100

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. C<^(^vL '-^ < JohrTT. Helling, Chairman Professor of Physical Science and Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^^^^1<^<^ /^^?»fcifc Merle A. Battiste Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ja'nies A. DeyrupProfessor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Robert C. S tourer Associate Professor of Chemi

PAGE 101

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Associate Professor of En This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1978 Dean, Graduate School

PAGE 102

UNIVERSITY OF FLORIDA 3 1262 08553 2744