Inorganic Chemistry: The Peroxo Complexes of Titanium
Inorganic Chemistry: The Peroxo Complexes of Titanium
Inorganic Chemistry: The Peroxo Complexes of Titanium
CONTRIBUTION OF INORGANIC
FROM THE LABORATORY CHEMISTRY,
EIDGENOSSISCHE
TECHNISCHE ZURICH,SWITZERLAND
HOCHSCHULE,
The appearance of an intense orange color on addition itates a yellow solid (NH4)sTi(Os)Fb,10 which is the only
of HzOz to an acidic solution of Ti(1V) was noted for crystalline compound obtained so far. I n all of these
the first time just a century ag0.l A considerable -preparations the ratio of peroxo groups t o titanium is
number of papers describe the analytical application of 1: 1. Compounds with higher ratios can be obtained
this reaction for the detection or photometric determi- only from strongly alkaline solutions.8z11y12
nation of either titanium or hydrogen peroxide. I n The species produced in solution also contain one
spite of the analytical importance, however, remarkably peroxo group per metal, as has been shown unambigu-
little has been established positively about the com- ously by photometric s t u d i e ~ . ~ ~ ~The l ~ - color
l~ of
pounds produced. these solutions is orange below pH, 1, is yellow around
It seems t o have been impossible to isolate the prod- pH 3, becomes pale yellow when the p H is raised fur-
ucts in the form of well-defined, crystalline compounds. ther, and disappears in the alkaline region. The ab-
Amorphous sulfates of approximate stoichiometric sorption spectra indicate the presence of a definite
compositions Ti(02)S04.3H202 and KZSO4.Ti(O2)- species below p H 1, there being no further changes in
so4 3H2O3 have been obtained by precipitation with the spectra a t still higher acidities.11'13~14~17 Electro-
ethanol. Corresponding amorphous peroxotitanium migration and dialysis studies18 show it t o be a cationic
oxalates, malonates, and maleinates have also been complex, the spectrum of which is not influenced by the
d e ~ c r i b e d . ~Moderately acid solutions (pH >2) slowly presence of simple anions such as HSOk-, NOS-, Clod-,
deposit an amorphous precipitate of peroxotitanium C1-.I6 If this species is assumed to be mononuclear,
hydrate TiOa(HzO),, which can be obtained more the formula can be denoted by Ti(O2) (OH)n-2(4-fi)+.
quickly by adding a base. The existence of one peroxo When the p H is increased, this ion will lose protons from
group 0z2-per titanium has been established by analy- the water molecules, which complete the coordination
sis, but there was some discussion as to the number of of the metal ion, thus increasing n and causing the color
water molecules ( x ) , viz., x = 1 (Ti(02)(OH)2)6-7or
x = Z.8vg Adding ammonium fluoride and NH3precip- (10) A. Piccini, C. R. Acad. Sci., 97, 1064 (1883); Gam. Chim. Ital., 17,
479 (1887); Z . Anovg. A&. Chem., 10, 438 (1895).
* To whom correspondence should be addressed. (11) M. Mori, N.Shibata, E. Kyono, and S. Ito, BuZl. Chem. Soc. J a p . , 19,
(1) G. Schonn, Z . Anal. Chem., 9, 41 (1870). 904 (1956).
(2) C.C.Patel and M. Mohan, Nature (London), 186, 803 (1960). (12) P. Melikov and L. Pissarevski, Chem. B e y . , 81, 953 (1898).
(3) R. Schwarz, 2. Anorg. Allg. Chem., 210, 303 (1933). (13) A. K. Babko and N. V. Ul'ko, U k v . Khim. Zh., 27, 101 (1961).
(4) G. V.Jere and C . C. Patel, Nature (London), 194, 471 (1962). (14) A. K.Babko and A. I. Volkova, Zh. Obshch. K h i m . , 21, 1949 (1951).
(5) F. Fichter and A. Goldach Hele. Chim. Acta, 18, 1200 (1930). (15) E.Gastinger, Z . Anovg. AZlg. Chem., 2711, 331 (1954).
(6) G.V. Jere and C.C. Patel, 2. Anorg. Allg. Chem., 819, 175 (1962). (16) Y. Schiippiand W.D. Treadwell, Helo. Chim. Acta, 81,577 (1948).
(7) F. Riveng, Bull. SOC.Chim. Fr., (5)12, 283 (1945). (17) A. K. Babko, A. I. Volkova, and S. L. Lisichenok, Russ. J . I n o u g .
(8)R.Schwarz and H. Giese, Z . Anoug. Allg. Chem., 176, 209 (1928). Chem., 11, 260 (1966).
(9) M.Schenk, Helv. Chim. Acta, 19, 625 (1936). (18) K.F. Jahr, FIAT Reu. Ger. Sci., 8, 170 (1948).
2381
2382 Inorganic Chemistry, Vol. 9, No. 11, 1970 J. MUHLEBACH,
K. MULLER,AND G. SCHWARZENBACH
changes. However, the n value of the species existing galvanic cell I11 determined with two exactly identical
in solution below pH 1 is not known and different glass electrodes.
authors disagree widely, suggesting Ti(H202)4 + (n =
0)13 or Ti(Oz)2+(n = 2)l0>I9 for its composition. : ; ; 1 5';) KCl
(111) glass 1 solution A glass
1. The Complex of Strongly Acidic Solutions.- potential (El - Ez)
The value n denotes the number of hydrogen ions which (El - E2) is positive if solution A is more acidic than R
are liberated when the peroxo complex is produced from and vice versa, but its absolute value will be smaller than
the titanium-aquo ion and HzOzwhich may be added in 58 log { [ H ] A / [ H J R mV,
) because of the considerable
any excess potential a t the phase boundary of the two liquids of
Ti4+(aq) + HZOZ --f +
T ~ ( O Z ) ( O H ) , - ~ ~ - n~ H
++ (1) different acidity. When the hydrogen ion concentra-
However, the existence of the simple aquotitanium tions of A and R are identical, (El - Ez) will be small
cation Ti(OH.Jn4+is doubtful, and it certainly does not although not exactly zero, because of small differences
occur above pH -0. Titanium chloride and titanyl in the activity coefficients of the ion H + in the two so-
sulfate are hydrolyzed immediately in water, resulting lutions and a still existing junction potential of small
in turbid solutions which can be made homogeneous magnitude. The following potentials were measured
only by adding a large excess of mineral acid, thus c, M 0.1 0.2 0.3 0.4
( E 1 - Ez),
mV +14.5 $3.5 +0.7 -2.2
making the determination of the number of hydrogen
ions produced by adding H202 (process 1) impossible. Because the number n ( = [ H ] ~ / 0 . 1must
) be an integer,
However, the following procedure proved to be prac- the result is unequivocal: [HI* = 0.3 and n = 3.
ticable. When a t low temperature (-30') pure liquid This fixes the stoichiometry of the orange-red com-
titanium chloride is introduced in an approximately pound formed in very acidic solutions. Assuming a
eutectic mixture of H202 and H 2 0 (which does not mononuclear complex, its formula is Ti(02)0H+.
solidify above -40°), the droplets of TiC14, disintegrated 2. Neutralization and Hydrolysis.-The color of
by the stirrer, dissolve slowly yielding a deep red homo- solution A changes to yellow on dilution and disappears
geneous solution. Our standard solution A was ob- on neutralization. A titration with NaOH carried out
tained through dilution and had the following composi- a t a total metal concentration [Tilt = 5 X M
tion: total titanium, [Tilt = 0.1; total peroxide, after adding NaC104 in order to create an ionic strength
[HzOz]t = 1; chloride, [Cl-] = 0.4; acidity, [H+] = p = 0.1 yielded the curve of Figure 1. The degree of
O.ln = ?.
PH
The number of hydrogen ions produced by reaction 1
is a t least n = 1; thus solution A has a pH of 1 or below
and all the titanium exists in the form of the definite
species with the pH-independent absorption spectrum 3.00,
mentioned above. The value n, therefore, should be
8.0C
an integer which can be determined by measuring the
concentration [H+].
::v
7.00
The accurate potentiometric determination of the
hydrogen ion concentration of a strongly acidic solution
cannot be carried out like a conventional pH measure-
ment because of the large diffusion potentials occurring
a t the liquid-liquid boundary to the medium containing
the reference electrode. We have used the following
galvanic cells
(I) glass electrode [ solution A I solution R I AgCl, Ag
4 5
0.
M , [ H ~ O Z=
] 5 X
5 7 8
(a > 4).
Although i t is impossible to analyze the curve of
0 w-
Figure 1 quantitatively, it tells US that the value f i = 4
is reached a t pH 3. This means that the solution con-
tains a t this pH an uncharged complex Ti02(0H)2 or its 0.50-
condensation products [TiO20], 'aq. However, the f i
value can be raised above 4 by neutralizing the solutions
0.40-
further, thus forming anionic peroxotitanium species.
At pH 7, ri = 5 is reached, corresponding to the mono-
nuclear complex Ti(Oz)(OH)a-. As mentioned, these 0.30-
solutions are unstable, and the condensation, leading
eventually again to the precipitate TiOp.aq, is accom-
panied now by a pH rise and a decrease of n on standing ax!-
OS0
(compare eq 2).
3. Deprotonation in the Range 1 < pH < 3.-The , ,
OQ
deprotonation between p H 1 and 3 which increases a JUT) Mo 3800 LWO LBO LLW A
from 3 to 4 is accompanied by a color change and can be
Figure 2.-Spectra of standard solution A diluted t o [Ti] =
studied spectrophotometrically. The condensation to 10-8 M and [ H z ~ z=] M with pH varying from 1 to 3 and
polynuclear species is quite slow in this range and with- constant p = 0.1 made up with HCl and KC1.
out influence a t times not more than a few minutes.
Solution A was diluted 100 times to [Ti], = M
with buffers containing KC1 and HC1 (in a few cases
NaOH), so that all the mixtures (solutions B) had an
Only €1 can be determined directly with a solution
ionic strength of p = 0.1 and a definite pH. The ab-
containing all the Ti as T i 2 0 b 2 + (pH <1). The other
sorption spectra were taken immediately after preparing
four parameters (€11, €111, 81,82) must be treated as un-
the solutions using the Cary 14 spectrophotometer.
knowns. Using a computer program by Professor G.
From the results (Figure 2 ) it becomes apparent that
Anderegg, a fairly good fit with the spectra of Figure 2
the increase of FZ from 3 to 4 cannot be simply due to the
was obtained furnishing the acidity constants pK1 =
deprotonation of the mononuclear Ti(Oz)OH+ and to
2.53 and pK2 = 2.05. I t is noteworthy that pK1 is
the formation of Ti(O2) (0H)z. The absence of isosbes-
larger than pK2 and that the intermediate I1 therefore
tic points indicates that the solutions contain a t least
has only a small range of existence, its maximum con-
three absorbing species and we can assume either that
all of these are dinuclear (case 3.1) or that condensation
centration a t pH l/2(pK1 +pK2) being only 22% of
[Ti205]t = '/2 [Tilt. It is not often found that the loss
takes place only when the acidity is lowered (case 3.2).
of a proton gives rise to the production of a stronger
3.1. When already the definite species existing be-
acid than the starting material.
low pH l is dinuclear, its formula would be [Ti(O2)-
3.2. We consider it therefore as more likely that the
OH]z2+which can be written as Ti205+aq2+(I). On
definite peroxotitanium present below p H 1 is mono-
raising the p H a simple deprotonation of coordinated
nuclear and that condensation to a dinuclear complex
water produces I1 and 111.
(eq 6) takes place when the acidity is lowered.
TizO?+ +Tiz05(OH)+ TizOs(0H)z (3)
2Ti(Oz)OH+ ---2. Ti206(0H)+ Tiz06(OH)z (6)
I I1 I11
I I1 I11
The optical density [D= (l/d) log (10/1)] is given by
eq 5 containing the molar extinction coeacients €1, €11, In order to distinguish between (3) and (6) more ac-
e111 of the three colored species as well as the acidity
curate photometric measurements would have to be
constants pK1 and pK2 of the proton donators defined carried out on solutions containing different metal con-
by eq 4. centrations [Tilt.
4. Rapid Neutralization in the Range 3 < pH < 9.-
As demonstrated in section 2, the a value rises above 4
when the peroxotitanium solution is titrated with
(4) alkali metal hydroxide implying the formation of anionic
2384 Inorganic Chemistry, Vol. 9, No. 11, 1970 J. M~~HLEBACH,
K. MULLER,AND G. SCHWARZENBACH
CHARTI
species. If no further condensation takes place, these R. value can be raised well up to 5. The condensation
species must be dinuclear and are formed by deprotona- proved to be sufficiently slow over the whole pH range
tion of the water molecules coordinated to the metal to permit a study of the deprotonation of the dinuclear
atoms of the central group TizOj2+. The general TizOj.aq2+by rapid alkalization.
formula Ti20j(OH),(2-”)+ therefore is appropriate for The proton-transfer reactions between Ti20b2fand
these products. OH- must come to equilibrium in a few microseconds,
All of these complexes, however, condense to poly- so that there should be ample time to study these
nuclears and the precipitate of peroxotitanium hydrate equilibria by measuring the pH of the mixtures before
(TiOs.aq) is finally obtained. These processes may be the condensation reactions become significant This
formulated with (7) and (S), depending on whether x has been accomplished with a streaming devicez0which
is smaller or larger than 2 . allows the mixing of the acidic solution containing the
H20 1 peroxotitanium and the solution of NaOH, as well as
x < 2: Tiz06(OH),(2-*j+ +- (TizOj),(OH),(*n-u) +
9
+ the determination of the pH of the fresh mixture within
( ( ~ l g-
) x)H+ about 5 msec. The instrument allows a second pH
+2Ti03.aq + (2 - x ) H + (7) measurement to be made about 10 msec after mixing.
1 It was found that the p H does not change within this
x > 2:
9
+
TizOs(OH),(*-2)- ---f - ( T i ~ 0 6 ) , ( O H ) , ( ~ - 2 Q ) - short time, thus proving that we are dealing with a
(x - oJ/!z))OH- true preequilibrium which can only be due to the simple
--f 2Ti03 aq + (cc - 2)OH- (8) protolytic processes
Below pH 3, x is smaller than 2, the condensing com- Ti205fOH)2 ===+ T120~(OH)3-& T I ~ O , ( O H ) ~4
~ - etc. (9)
P K ~ pK4 PKS
plexes are cations or uncharged TizOj(OH)n, and the
primarily formed lower polynuclears of the still homo- However, in a rapidly streaming solution, the pH
geneous solutions must also be positively charged cannot be determined as accurately as in a solution at
(2 > y / q > x). In this range the reaction-which is rest. This must be taken into account when the re-
accompanied by a pH drop-is slow, which made the sults presented in Figure 3 are evaluated. One of the
optical investigation of the deprotonation up to x = two pressure vessels of the streaming apparatus was
2 possible (reaction 3 or 6, section 3 ) . filled with the standard solution A, the other with NaOH
Above pH 3, x is larger than 2, the condensing species in a concentration yielding a definite degree of neutral-
are anions or Ti205(OH)2, and the precursors of the ization “a” on mixing. In the course of the experi-
final precipitate must also be negatively charged poly- ment, both solutions were first diluted in dilution
mers (2 < y / q < x ) , The rate of condensation- chambers with a solvent containing the inert electrolyte
which is accompanied by a pH rise-is now much KC1 and some HnOz, the diluted solutions then streamed
faster and speeds up with increasing pH. into the mixing chamber, whence the mixture passed the
It is reasonable to assume that the units Tiz05(0H)2 glass electrode. From the pH, fi was calculated with
will condense by forming p-hydroxy bridges between eq 2 . The concentrations in the mixing chamber were
the metal atoms as illustrated by the formulas shown in [Tilt = 3.5 X M , [HzOz]t = 0.1 M , p = 0.1
Chart I. (KC1). The precision of the pH is about +0.05 unit
Coordinated water has been omitted from these giving an appreciable error in fi for pH values below
formulas. If each T i carries such a molecule of H20, 4. (Solutions containing HzOz are usually super-
the composition of the peroxytitanium hydrate is saturated with 0 2 . Thus small bubbles were often
TiOa. 1.5H20 which is about what has been found formed in the narrow tubings which influenced the
analytically.* hydrodynamic behavior of the streaming liquids un-
Although the rate of condensation is rather large in favorably.)
the pH range 3-9, it is by no means instantaneous. In spite of these inaccuracies, Figure 3 demonstrates
The solutions remain clear for quite a while after adding definitely that the peroxotitanium complex can be
the base and in an automatic titrator (Figure 1) the (20) G. Schwarzenbach, Pure A p p l . Chem., 5, 377 (1962).
THEPEROXO COMPLEXES O F TITANIUM Inorganic Chemistry, Vol. 9, No. 11, 1970 2385
deprotonated up to an a value of a t least 6, correspond- in Figure 3 was computed with eq 11, using the follow-
ing to the dinuclear complex of the formula Ti2- ing pK, values: pK3 = 4.1, pK1 = 5.6, pK5 = 7.2,
Ob(OH)64-. (At still higher ratios of NaOH per ti- pK6 = 8.5. The index number corresponds to the x
tanium, the mixtures became very unstable and the in the formula Ti~05(OH),‘2-2’+. The definition of
p H values uncertain. The results of these experi- pK1 is given with eq 4 and the others are defined analog-
ments are not presented.) ously
t
x-0
[s = 2(5 - 3) and x = 1-61 (11)
From the optical investigation of the very acidic
solutions data for the first two of these protonation
constants have been obtained: pK1 = 2.5, pKz =
2.1 (section 3). These data, however, are uncertain
because of the doubts expressed in section 3 as t o
whether the color changes presented in Figure 2 are due
to the simple deprotonation of Tiz06. aq2+ (reaction 3).
We consider it more likely that the complex existing
below pH 1 is mononuclear which however has no in-
fluence on function 11 a t pH values >3.
Equation 11 is obtained by starting from
Lo
6 6
Q = x[Ti~0dOH)~l [Ti~0dOH)~l
x-1
4 5
5 8
b *-n
- x -a(fi-s)
further deprotonation, yet it can also be explained by
the formation of new species containing more than one
Figure 3.-Rapid alkalization of peroxotitanium obtained with peroxo group per titanium. The occurrence of such
the flow apparatus. The pressure vessels contained standard complexes in alkaline solutions containing an excess of
solution A and NaOH (concentration 0.3 40.6 M), respectively. Hz02 has been proved.8p11t12
During the measurement these solutions are injected separately 5. Peroxotitanium Chelates.-The formula (H20),-
into a stream of the solvent containing the inert electrolyte (KC1) TiOsTi(OHz)y2+from which the series of complexes
and some H102. The diluted solutions finally are mixed in the
mixing chamber where the analytical concentrations were as Tiz05(OH).(2-2)+ can be derived by detaching protons
follows: [Tilt = 3.53 X M, [H~OS]= 0.1 M , p = 0.1 suggests the feasibility of a replacement of the coordin-
(KCl). The line has been computed with eq 11. ated water by the ligand atoms of chelating agents.
Peroxotitanium oxalates were indeed described as early
As shown in Figure 3, the pH rises almost linearly as 190721and again about half a century later.6 How-
with a between pH 3.5 and 9.2, corresponding to the ever, these amorphous solids have never been well
range 4 < a < 6. A mononuclear complex Ti(O2) (OH)z characterized. Sweetser and BrickeP observed a
(corresponding t o a = 4) would have only two buffer color change when ethylenediaminetetraacetic acid
regions in this a range. However, the linear rise of was added to the acidic orange solutions of peroxo-
p H with a can well be explained with the dinuclear titanium. The adduct contains Ti, HzO2, and EDTA
species Tiz05(0H),; while x increases from 2 to 6 in the in the ratio of 1: 1: l Zand
3 its formation is the basis of a
formation of T ~ z O ~ ( O H(%) ~=~ 6),
- we get four strongly complexometric titration procedure for the determina-
overlapping buffer regions. With four pK values, tion of Tisz4
more or less evenly spaced, the course of the experi-
mentally obtained neutralization curve is readily in- (21) A. Mazzucchelli, Gase. Chim. Ita$., 87 (II), 545 (1907); 40 (I), 666
(1910).
terpreted, thus supporting the evidence for dinuclear (22) P. B. Sweetser and C. E. Bricker, Anal. Chem., 26, 195 (1954).
peroxotitanium complexes. (23) S. Musha and K. Ogawa, N i p p o n Kagaku Zasshi, 78, 1686 (1957).
(24) G. Schwarzenbach and H. Flaschka, “Complexometric Titrations,”
The curve drawn through the experimental points Methuen, London, 1969, p 200.
2386 Inorganic Chemistry, VoZ. 9, No. 11,1970 J. MUHLEBACH,
K. MULLER,AND G. SCHWARZENBACH
Since the number of water molecules coordinated to D, of mixtures of peroxotitanium and ligand is plotted
each metal atom in the complex Ti2Ob2+is a t least y V S . the ratio x = [Ti]t/c, c being the sum of the total
= 3 (section 4),terdentate ligands should be especially concentrations of titanium and ligand ([Tilt [Lit +
well suited to chelate peroxotitanium. They should = c), which was kept constant in all the mixtures which
be oxygen donors mainly because of the pronounced all were of pH 3.
A character of doTi(1V). If no reaction would take place between peroxo-
5.1. The Job diagrams (Figures 4-6) demonstrate titanium and the chelating agent, a linear rise of the
optical density with x would be observed, that is, the
straight line D,. The difference (Dl - D z ) shows a
pronounced maximum a t x = 0.5 in a t least two of the
three figures which proves unequivocally the formation
of a complex containing Ti and L in the ratio of 1: 1.
os The molar extinction coefficients of unchelated and
chelated peroxotitanium are similar and the difference
(Dl - Dz) therefore is only a comparatively small
quantity. It is quite clear, however, and noteworthy
that it rises linearly to the maximum but does not fall
D -D2
Dt
0 1 2 1 4 5 @ 7 I S blacllil
IO 9 7 6 5 4 1 2 I O.&[apc]
[I
0.1
0.6
Figure 6.-Continuous-variation curves of peroxotitanium with
dipicolinic acid, X 355 nm. The acidity was kept constant at p H 3;
peroxotitanium concentration, [Tilt = 0 +. 10 X
concentration, [chellt = 10 X +0 M .
M ; ligand
'I
as
/a00
7.00
Loo -
~
/ - = - - - P 7.00 -
4.04 -
500 -
9.00 -
3.00 -
L
3 4 5
a
w) 9 8 7 6 5 L 3 2 1 O&[NlA] Figure 10.-Alkalimetric titration of standard solution A di-
luted to [Tilt = 5 X M , [Hz02] = 5 X l o w 2M ,and p = 0.1
Figure 9.-Continuous-variation curves of peroxotitanium with (NaC104) with 1 mol of the disodium salt of nitrilotriacetic acid
nitrilotriacetic acid, X 375 nm. The acidity was kept constant a t per Ti.
pH 3; peroxotitanium concentration, [Ti]t = 0 -.f 10 X M;
ligand concentration, [chel]t = 10 X + 0 M.
1: 1: 1.2s However, the charge of the complex species
An alkali metal free complex of peroxotitanium and is not positively known. We again started from solu-
dipicolinic acid has also been obtained, which seems to tion A, obtained from Tic&, and added 1 mol of the
be mononuclear. It forms orange-red rhombohedra dialkali metal salt of ethylenediaminetetraacetic acid.
of the composition [TiOZ(Dipic)] .4&0 and behaves as The alkalimetric titration curve shows a sharp jump
a very strong acid. from pH 6 to 8 after the addition of 4 mol of strong base
5.4 It was t o be expected t h a t peroxotitanium (the same amount needed to produce Tiz0j(OH)2)which
would also form a chelate with the quadridentate anion leads to the formulation
(NTAa-) of nitrilotriacetic acid. The curves obtained Ti02(0H)+ -b H2EDTAa- + 2Ti(Oz)EDTAZ- + H + (16)
in applying the continuous variation method are shown
From the neutralized solution the solid complex
in Figure 9 and demonstrate again the formation of a
potassium salt was obtained, which forms orange mono-
complex containing one molecule of the ligand per
clinic prisms of the composition Ka[Ti(Oz)EDTA].HzO.
titanium. The color change during complexation is
All our preparations were somewhat deficient in potas-
more pronounced with NTA than with terdentate
sium. The molecule of HzO is easily removed and does
ligands and the difference in optical density is quite
not seem to be coordinated.
appreciable a t X 375 nm. Furthermore (Dl - D z )
5.6. The infrared spectra of the crystallized com-
rises linearly up t o x = 0.5 and decreases as a straight
plexes are given in Figures 11-14 for the frequency range
line afterward in contrast to the curves of Figures
650-1 100 cm-l. According to GriffithZ8the vibrational
4-6. From both observations we can guess that the
modes of the peroxide grouping of the peroxo complexes
NTA chelate might be mononuclear.
of various metals fall between 800 and 1000 cm-I and
I n the alkalimetric titration of solution A, to which
also our new peroxotitanium compounds show strong
1 mol of the dialkali metal salt of nitrilotriacetic acid
bands in this region. For those which we consider t o
has been added, the p H jumps sharply between 4.5
be metal peroxide group bands the frequency is given in
and 9 (Figure 10) after the addition of 4 mol of OH-
the figures.
per metal. A complex with the charge 2 - per Ti must
6. Conclusions.-Our experiments show definitely
be the only species present in this pH range, and when
that peroxotitanium is present in very acidic solutions
we assume i t t o be mononuclear, its formation has t o be
as a cation with a single charge per metal. The com-
formulated with
plex is probably mononuclear TiOzOHf below p H 1,
Ti(02)0H+ + HNTA2- v Ti(Oz)(NTA)OH2- +H+ (15)
but condenses to a dinuclear species a t least between
This complex could also be isolated in the form of a pH 1 and 3, and the unit TizOsappears t o be especially
well-crystallized lemon yellow salt: KZ[Ti(Oz)(NTA)- stable. The first products obtained on alkalization are
O H ] - 3 H z 0 . All the three water molecules are lost the deprotonation products of TizO5.aqZfof the general
simultaneously and long before Oz is given off and i t formula Tiz05(OH),(2-Z'+. These species condense
seems that they are not coordinated to the metal. slowly to the insoluble peroxotitanium hydrate TiOs-
5.5 The formation of the EDTA complex of peroxo- (HzO),, x being between 1 and 2 . The dinuclear unit is
titanium in solution has been mentioned 15 years ago.22 present also in the peroxotitanium chelates formed with
It contains Ti, 02, and the chelating agent in the ratio (28) W. P. Griffith, J. Chem. Soc., 5248 (1964).
THEPEROXO COMPLEXES OF TITANIUM Inorganic Chemistry, Vol. 9, No. 11, 1970 2389
-7- -n
Figure 11.-Infrared spectrum of Kz[TizOa(Dipic)~]5Hz0 where Figure 13.-Infrared spectrum of Kz [Ti02 [OH)NTA] *3Hz0
the arrows indicate peroxide band frequencies at 927and 879 cm-l. where the arrows indicate peroxide band frequencies a t 948, 923,
and 872 cm-l.
weight
c
Formula 7 - K or Na--
Calcd Found
--Ti---
Calcd
I
TABLE
Found
0 2
Analyses, %
(peroxide) ---C
----- ---H-- -N-
-
Calcd Found Calcd Found Calcd Found Calcd Found
678.17 11.55 11.1 14.70 14.5 9.43 9.5 24.75 24.70 2.07 2.13 4.13 4.39
700.00 6.56 6.7 14.26 14.2 9.13 9.2 24.00 24.00 3.17 2.88 4.00 4.23
319.01 0.00 0.0 15.63 ... 10.3 10.2 26.35 26.42 3.47 3.22 4.39 4.58
419.23 18.66 18.6 11.90 12.0 7.64 7.4 17.18 17.37 3.13 2.65 3.34 3.58
466.25 16.78 1 4 . 3 10.72 11.0 6.87 6.8 25.75 26.5 3.03 3.3 6.03 6.3
further ligands (in the series OHz, -COO-, 3 N, OH-) 75". The last 2 HzO come off at 150-175" together with O2
the more difficult it becomes to reduce Ti(1V) to Ti(II1) from the peroxo group. Under vacuum (0.01 Torr) elementary
oxygen starts t o be given off a t 100".
and the higher apparently is the frequency of the intense
4. N a ~ [ T i ~ O ~ ( D i p i8c H
) ~~] .-The
0 orthorhombic orange-yel-
absorption band. low prisms of the sodium salt were obtained with the same pro-
cedure after neutralizing with NaOH instead of KOH. Below
Experimental Part 2070 relative humidity, the crystals lost water. The thermo-
1. Stock Solution.-A 37.94-g (0.2-mol) amount of freshly gram taken in dry I ~ of z 1 a t m reveals the loss of 4 HsO a t 7 5 " ,
distilled Tic14 was added drop by drop to 150 ml of hydrogen again of 4 HzO a t 125', and of 0 2 a t 150". Under vacuum 25y0
peroxide (free of any stabilizer) containing 2.3 mol of H 2 0 2which of the peroxo oxygen was given off a t 50-60" and the remaining
was kept at about -30'. The addition took from 2 to 3 hr and 0 2 above 100".
vigorous stirring was essential. The temperature should not rise 5. [Ti(Oz)(Dipic)].4HzO.-One mole of dipicolinic acid per
above -25' in order t o prevent evolution of 0 2 but should not T i was added to solution A and the acidity of the homogeneous
be too low either to prevent Tic14 from solidifying. Solid TiCla mixture was reduced by adding 4 mol of KaOH. The still
reacted very slowly and accumulated in the suspension which, strongly acidic orange solution was concentrated below 25" and
when later warmed up, sometimes caused a n almost explosive then placed over SOY0 H2SO4. Orange-red triclinic prisms were
decomposition (danger!). After the addition of TiC14, the formed within some days. When these were dissolved in water, a
homogeneous deeply red mixture was diluted to 500 ml with ice yellow-orange strongly acidic solution was obtained. Two HzO
water. This stock solution with [Tilt = 0.4 hl, [Cl-] = 1.6 M , were lost below 10Yo relative humidity. The thermogram
and [HzOZ]= 4 M was kept at - 10' and did not show any signs taken in dry KZof 1 atm reveals the loss of 2 H20 at 75'; the
of change after many months. remaining water is given off together with O 2a t 150".
2 . Solutions for the Potentiometric and Optical Studies.- 6 . Kz[Ti(Oz)OH(NTA)].3HzO.--The solution of 1.91 g of
The standard solution A was prepared by diluting the stock solu- nitrilotriacetic acid (=0.01 mol) was added to 100 ml of stand-
tion four times and kept unchanged a t 2' for about 2 weeks. ard solution A and the mixture was neutralized with 600 ml of
The solutions B for the spectrophotometric investigation between 0.1 M KOH up t o p H -6. The volume was now reduced con-
p H 1 and 3 (section 3) were obtained by adding to 10 ml of solu- siderably a t a temperature not exceeding 25' and the salts were
tion A the necessary quantities of KCl and HCl (or iYaOH) and precipitated with ethanol. The separation of the very soluble
diluting to 1000 ml. For taking the curve presented in Figure 1, peroxo complex from KCl was achieved by fractional crystal-
10 ml of 1 M h'aClO4 was added t o 5 ml of standard solution A, lization, using water and ethanol containing some HZOZas sol-
the mixture was diluted to 100 ml, and the titration was started vents. The lemon yellow crystals formed triclinic prisms which
immediately. were kept over SOYc HzSOa. Below 5YCrelative humidity all the
For taking the Job curves (Figures 4, 5, 6, 9) a solution of three molecules of water were lost. Heated in dry NZ ( p = 1)
peroxotitanium of [Tilt = M and solutions of the chelating the 3H20 were given up a t 80-100°, b u t there was no evolution
agents of [chellt = l o T 3 M had to be prepared, both of ionic of Oz UP to 180".
strength = 0.1. These solutions had t o be mixed in the ratios 7 . Kz[Ti02(EDTA)]GHzO.-The solution of 2.92 g of ethyl-
1 : 9 t o 9 : l to solutions of approximately p H 3. T o 10 ml of enediaminetetraacetic acid ( = 0.01 mol) was added to 100 ml of
standard solution A, 100 ml of 1 M KC1 was added as well as standard solution A and the mixture was neutralized to p H -6
30 ml of 0.1 M NaOH and diluted to 1000 ml. On the other with 600 ml of 0.1 M KOH. The volume mas now reduced
hand, diglycollic acid or iminodiacetic acid, dipicolinic acid, considerably at a temperature not exceeding 25' and the salts
and the monoalkali metal salt of nitrilotriacetic acid, respectively, were precipitated with ethanol. The separation of the very
were made up with KCl to n/i solutions without adding any soluble peroxo complex from KC1 was achieved by fractional
strong base. The preparation of the titanium solution and mix- crystallization, using water and ethanol containing some H20z
ing with the chelating agent took about 3 min and the optical as solvents. The orange-yellow crystals form monoclinic prisms
density was then determined without delay. and are kept over 80% HzS04. All our preparations were some-
3. Kz[TizOs(Dipic)2] .5HzO.-The solution of 1.67 g of dipico- what deficient in potassium which probably is due to a partial
linic acid (=0.01 mol) was added to 100 ml of standard solution protonation of a carboxylate group not coordinated. I n a n
A; the mixture was neutralized to pH 4.5 with 500 ml of 0.1 M atmosphere below 5 % relative humidity H20 is slowly lost.
KOH and concentrated under vacuum between 20 and 25' Heated in dry Xz ( p = 1) one molecule of HzO was given up a t
to a volume of about 50 ml. A slight turbidity which may ap- 90-110", and there was no evolution of OZup t o 170'.
pear can be removed by adding some more base. The solution 8. Analysis.-The alkali metal content was determined by
was now placed in a desiccator over anhydrous CaClz. The atomic absorption, T i gravimetrically and C, H , and N by com-
orange-yellow monoclinic plates formed within a few days were bustion (see Table I). I n the peroxo dipicolinates the peroxide
washed with very little cold water and dried over 80y0 H z S O ~ . oxygen could be obtained by titration with KMn04. Nitrilo-
If the drying was done below 30y0 relative humidity, the crystals triacetic acid and ethylenediaminetetraacetic acid, however, are
got turbid and lost water. The thermograms taken in dry NZ oxidized by permanganate, but a titration with Ce(1V) was pos-
of 1 a t m reveal the loss of 1 HzO at 50' and 2 further HzO at sible.