Journul OJ Orgmomerollic ChtmLrr_v

Elsevicr Publishing Company Amsterdam 165

Printed in The Netherlmds

SHORT COMMUNICATION

Colored indicators for simple direct titration of magnesium and lithinm reagents

We have been engaged in a general study of the relative ability of a metal alkyl
(particularly lithium and magnesium reagents) to react in two rather different ways :
(i) to react covalently as a carbanionic Lewis base, e.g., to accept a proton from an
acid, such as an alcohol ; (ii) to undergo single electron transfer, i.e., reduce an electron
acceptor by charge transfer_ Interestingly enough, the latter process involves reaction
of the organometallic with a base (rather than us a base) to form an intermediate
charge transfer complex’. Quite likely the coordination intermediates so often
proposed for organometallic addition reactions can be regarded as charge transfer
complexes_ Although the general study is far from complete, one preliminary result is
reported here because of its practical utility. The complexes formed by some bases
with lithium and magnesium reagents are highly colored and sufficiently stable to
serve as indicators which allow the organometallic to be directly titrated with acid
(an alcohol) to a visible, well defined, stoichiometric end point.
Actually, both lithium and magnesium alkyls are coordinated by virtually any
base, including simple ethereal and ammoniacal Lewis bases, but to serve as practical
indicators we choose -n-type bases, specifically polycyclic aromatic bases which form
complexes with intense absorption in the visible spectrum. It is also necessary to
choose bases which undergo simple reversible coordination with the organometallic,
i.e., which do not react beyond formation of the charge transfer complex; aromatic
systems can undergo addition and reduction by metal alkyls2-‘. Two aromatic bases
which we find quite practical are IJO-phenanthroline (I) and 2,2’-biquinoline (II).

TABLE 1
RESULTS FP.OY DlREcT llTFcATIOh3 OF ORGANOMETALLIC REAGENTS WITH SEC-BUTYL ALCOHOL

Reagent Soloent Direct titration Independent

molarit analysis
molarity

Di-set-butylmagnesium C.sHlz 0.78 0.77b

Isobutylmethylmagnesium 0.71 O-72*
Phenylmagnesium chlcride &g 2.20= 2.13d
n-Butylmagnesium chloride 2.73 2.8’
n-Butyllithium C,=H ,‘4 ’ 1.621 1.659
see-Butyllithium C&I, 1.31 1.339
tert-Butyllithium CsH,, 1.54 l.52h

p Average of two or more titrations. * Comparison analysis used for magnesium dkyls in hydrocarbon
solutions involved hydrolysis of the reagent and use of the hydrolysate for titrimetric determinations of
the total base and the magnesium ion and for VPC analysis of the alkanes corresponding to the alkyd
groups; cf- ref. 8. ’ Average deviation of four analyses, 0.7 %. d Comparison analysis involved trimethyl-
silyiation of reagent and VPC of product as described in ref. 11. ’ Value furnished by commercial source of
sample analyzed, M and T.Chemicals, inc., Rahway, N. J. * Average deviation of five analyses, 0.7%.
g Comparison analysis involved oscillometry as described in ret 14. h Comparison analysis involved
oxidation of reagent with vanadium pentoxide ti described in ref. 13.

J. Organometal. Chenz., 9 (1967) 165-168

166 SHORT COMIMUNICATION

A milligram or so of (I) added to 10-20 ml of a hydrocarbon or ethyl ether solution

containing cu. 0.1-1.0 M dialkylmagnesrum reagent produces (instantly) a bright
violet solution. Titration of this solution with 1 M set-butyl alcohol in xylem muses
no significant dimunition in color, until two mole equivalents of the titrant are added
and the color fades out sharply. For example, a sample of di-set-butylmagnesium in
cyclohexane* which was found by indirect analyses to be 0.77 M, was found by our
direct titration to be 0.78 M. This result and those from titrations of other organo-
metallics are summarized in Table 1.
The sharpness of the end point with ethereal solutions of magnesium alkyls
appears to be better than with hydrocarbon solutions. In the hydrocarbon, precipita-
tion of magnesium alkoxide formed during titration may cause turbidity or viscosity
problems ; the alkoxide precipitates in particulate form and these problems are elimin-
ated by prior addition of a tertiary amine or ethyl ether. Both ethyl and butyl ether
have been used as solvents for direct titration of butylmagnesium chloride. Our
analysis of a sample of this Grignard reagent, available commercially and indicated
to be cu. 2.S M, showed it to be 2.73 M. Phenylmagnesium chloride was also satis-
factorily analyzed by the direct titration. Absolute accuracy is not really established
for any independent analyses of organomagnesium reagentsg-ii, but the direct
titration does show good precision; reproducibility is within ordinary volumetric
analysis error.
The real accuracy for this direct titration has been established with organo-
lithium reagents, because the accuracy is known for independent analyses we used on
the lithium compounds for comparison 12- 14. Addition of a few milligrams of (II)
and about 5 ml of cu. 1.5 M butyllithium in hexane to 20 ml of benzene produces a
yellow-green or chartreuse-colored solution. After titration with 1 M set-butyl
aIcoho1 in xylene, the solution is cIear and colorless; the disappearance of the green
color occurs sharply after addition of one mole equivalent of titrant. For example a
sample of butyllithium solution, available commercially and indicated to be 1.6 M
in lithium reagent, was found to be 1.62 M by this direct titration with see-butyl
alcohol. Our independent analyses of the same solution showed it to be 1.65 M.
Equally good results in titrating butyllithium were obtained by forming the
indicator with (I), which gives a rust-red complex. Hydrocarbon solutions of sec-
butyllithium and tert-butyllithium reagents were also satisfactorily analyzed using
(I) and direct titration with set-butyl alcohol. Because they react with the reagents,
ethers cannot be added to the titration mixtures with (I) or (II) and lithium reagents*.
However, there is no problem here with turbidity and precipitation in the hydrocar-
bon titration mixture because the lithium set-butoxide is quite soluble15_ If the reac-
tion mixture is protected with an atmosphere of nitrogen, it remains bright and clear

* Ethereal lithium reagents can be directly analyzed by an oscillometric te&niquet5_

J. Organometal. Chem, 9 (1967) 165-168

 SHORT COMMUNICATION 167

throughout titration. Reaction with n-butyl alcohol as titrant is also stoichiometric16,

but the prim-butoxide formed is insoluble.
Using (I) and butyllithium in benzene, we have examined the reversibility of
formation of the colored charge transfer complex which serves as our indicator_
After the end point, addition of more lithium reagent to the titration mixture im-
mediately regenerates the red color. The additional reagent still requires just one mole
equivalent of alcohol to come back to the end point, loss of the red color. Obviously
excess alkoxide, a problem in many organometallic analyses, does not interfere with
this direct titration_
It is our impression that our indicator is relatively stable to protonation, that
is, the acid (alcohol) used as titrant apparently does not attack the indicator (metal
alkyl/base complex) until all of the uncomplexed organometallic reagent is reacted.
One reason for this belief is that toward the end of a titration, as indicator and un-
complexed reagent approach equimolarity just before the end point, color does not
temporarily disappear and return after each increment of titrant. This premature dis-
aF;pearance (and return) would be expected if the indicator complex were not re-
latively stable toward chemical reaction involving protonation.
Toward chemical reaction involving charge transfer, however, our indicator
complex is not particularly stable. If acetone, instead of alcohol, is used as titrant to
consume the butyllithium, the red color of the complex disappears well before a
stoichiometric end point. This instability of the charge transfer complex is thought
to be a consequence of the particular susceptibility of acetone to reduction. This
premature disappearance of our indicator color with acetone as titrant cannot be a
consequence of side reactions with the butyllithium, for the true end point is observed
with acetone as titrant using oscillometry’4.
In conclusion we suggest that direct titration of other organometallics may
prove possible using appropriate charge transfer complexes as indicators*. Although
many techniques have been devised for analysis g-14*16 of Grignard reagents and of
lithium reagents particularly, it is our experience that no other technique even
approaches this direct titration in expediency and simplicity. Except for the in-
escapable necessity of protecting such reactive solutions from moisture, air, etc.,
magnesium and lithium reagents in hydrocarbons may now be titrated as simply as
sodium hydroxide in water.

Acknowledgement

The authors were assisted in this work by suggestions made by personnel in the
Ellestad Research Laboratory of Lithium Corporation of America and by a Research
Fellowship for one of us (SCW) granted by The Tennessee Eastman Corporation.

Department of Chemhry, SPENCER C. WATSON

University of Tennessee, JEROME F. EASTHAM
Knoxville, Tennessee, 37916 (U.S.A.)
* We have not made an extensive search for indicators, but we did eliminate several other aromatics
(including dimethyl- and diphenylphenanthroline, fluorene and triphenylmethane) as being less suitable
before studying bases (I) and (II) with lithium and magnesium alkyls. Hydrocarbon solutions of aluminum
and zinc alkyls we fmd also give good color with base (II), but we have not searched for a titrant to use on
these other organometallics; their solvolysis with see-butyl alcohol is not a rapid stoichiometric reaction.

J. Organometal. Chem., 9 (1967) .165-168

168 SHORT COMhfUNICATION

1 C. G. .%XE~-TAS A&P 3. F. Ehs~~~ru,J. Am. Chem. Sot., in press.

2 w. !%XENK AND R. OcHs, Ber., 49 (1916) 608.
3 &I.W. BROWNA?~DR. C. Jam, J. Chem. Phys., 37 (1962) 1571.
4 H. J. S. WINKLER, H. W~.KT_ERM R. BOLLMGER, Chem. Commzm., (1966) 70.
5 G. A. RUSSELL,E. G. JANZENAND E. T. SIXOM; $. Am. Chem. Sot., 86 (1964) 1807.
6 J. A. DIXQN AND D. H. FISIMM, J. Am_ Chem. Sot., 85 (1963) 1356.
7 H. P. THRONDSEN AND I-L ‘&is.%J. Organometuf.Chem., l(l963) 301.
8 C. W. KAXIENSKIAXD J. F. E~IHAM, J. Organometal. Chem., 8 (1967) .X2.
9 M. S. -CH AP*D 0. IUNWJTH, Grignard Reactions of Nonmetallic Substances,Prentice Hall,
New York, 1954, pp. 92-98.
IO R. D. HOLLAXDERAND M. ANI’EUNIS,Bufl. Sot. Chim. Beiges, 72 (1963) 77.
11 H. 0. HOUSEI\NDW. L. E&SIXSS, J. Organometal. Chem., 4 (1965) 95.
12 H. &_,,N, F. K. CARIIEDGEAND S. Y. SM, J. Organometul_ Chem., 1 (1963) 8.
I 3 P. F. COLLINS,C. W. KAMENSKI,D. L. E%AY A~XIR. B. EL~AD, Anol. Chem., 33 (I 961) 468.
14 S. C. WATSON AND 3. F. EASTUAM,Anal. Chem., 39 (1967) 171.
t 5 C. W. KA~IIENS~ AKD D. H. LEAKS,J. Org. Cftem., 30 (1965) 3498.
I6 I’?_L. E-x, A&. Chem., 36 (1964) 854.

Received Novexnber 22n4 1966

J. Organontetal. Chem, 9 (1967) X65-168