Kraus, Nelson and The values of the ordinates for the distribution data are purely relative for each of the functions plotted; the shapes of the curves are the feature of consequence. Although all the evidence relating to the extraction by ethers indicates that the extracted species contains two chlorines, the fact that the extraction curves do not correspond to the spectral data is not surprising; the distribution depends on many factors other than the equilibrium in the aqueous phase, and this equilibrium itself is shifted due to the presence of ether in the aqueous phase. For example, a decrease in the distribution ratio for isopropyl etherrwould be expected a t the highest acidities even if there were no decrease in the amount of the extractable species in the aqueous phase.*O The rapid rise of all curves for acidities up to 4 M indicates that the equilibrium in the aqueous phase is the most important factor in determining the distributions a t these conditions. The anion resin data corresponds sufliciently well with the spectral data to suggest that the species containing two chlorines is the form bound by the anion exchanger. (11) K . A . Kraus, F. Kelson and G. E. Moore, THISJ O U R N A L , 77, 3972 (1955) (10) D . E. Campbell. A. H. Laurene and H. M. Clark, ibid., 74, 6103 (1952).
The most glaring failure of correspondence is the apparent leveling off in the amount of the two-chlorine species in 1-2 M HC1. The most likely explanation is that these solutions contain another species of Mo(V1) that also has an absorption maximum in the 260-270 mp region, and that the amount of the two-chlorine species actually follows the dashed line indicated on the figure. Examination of the spectrum of Mo(V1) in HClOd solutions indicates that species with such absorption characteristics are present a t such acidities. The trends displayed in the spectra of these chloro complexes of Mo(V1) are consistent with the trends displayed by other elements. The resolved peaks with maxima a t 264 and 302 mp are those associated with the Mo-C1 bond, and the complex with the larger number of chlorines has the absorption a t the longer wave length. It is also of interest that the absorptions a t the maxima of the two resolved curves differ by only 3% (well within the errors inherent in the method). This relationship has been observed previously with the chlorocomplexes of Sb(V),I1 but there are not sufficient data from other elements to decide whether this is a general type of behavior. (11) H. hf Keumann, t b r d , 76, 2611 (19j4).
EVANSTON, ILLISOIS
. . ~
[ C O S l ' R I H U I ' I O S O F THE I > & P A R T Y E S T O F CIIEXISTRY, CLARK UNIVERSITY]
B Y E>inn >>ton >>100
>ion
>ion
. . ... ..
,
,
METALCHELATE COMPOUNDS AS CATALYSTS
June 20, 1957
3 033
TABLE I (continued) Metal
Fe( 111) Fe(II1) Fe(II1) Fe(II1) Fe(II1) Fe(II1) M O O , ( VI) MoO*(VI) MoO?(VI) hloOr(V1) MOO*(VI) MoO>(VI) MoO*(VI) MoO*(VI) MOO?(VI) ,I.lOOZ(VI) MOO?(TI) hloOz(V1) MoOg(T1) MOO>( VI) Moon( VI) MOOl(V1) MOO,(T~I) Mo02(VI) ZrO(I1) ZrO(I1) ZrO( 11) ZrO(I1) ZrO(I1) ZrO(I1) ZrO( 11) ZrO( 11) ZrO(11) ZrO( 11) ZrO(I1) ZrO(I1) Cr(II1) Ti(1L‘) Sn(1V) Sn ( 11.) Sn(I\-)
Concn., pmoles/l.
Ligand
iJH
2-HxG HIMDA1 HEDDA HEDTA PDS DNS
7.0 7.0 7.0 7.0 7.0 7.0 7.00 7.00 7.00 7.00 6.95 7.00 7.00 7.00 7.10 7.00 5.20 6.90 7.10
... HEN EDTl HEDTA HEDD.1 2-HxG HIM D.l XTh PDS PDS PDS 5-SSA DKS CIT ASP.\ DPA EN TU HEDDrl PDS 2-HxG H I MDA CIT EDTA EDTA EDTA EDTA HEDTA HEDTA HEDT.1 2-HxG PDS PDS DSS 5-SSA
7.00 7.20 7.00 7.15 7.00 7.00 7.00 7.55 7.45 7.50 6.40 7.40 7.00 8.40 6.00 7.00 7.05 7.0 7.0 7.0 7.0 7.0
Metal
1035 1005 985 1020 1010 950 1000 1000 1000 950 960 980 960 970 965 4950 4950 990 890 890 945 970 950 920 1000 995 1013 988 919 515 515 1005 510 990 990 990 900 2970 1510
Sat’d Sat’d
The definitions of the abbreviations used to identify the various ligands employed are given in Table 11. The corresponding catalytic activities of a number of Cu(I1) chelates in the hydrolysis of DFP are listed in Table 111. Discussion Effect of Ligand on Activity of Cu(I1) Chelate Compounds.-The 29 catalytically active Cu(I1) chelates listed in Table I may be classified into the following groups on the basis of the type of ligand involved: amino acids (7), phosphates (2), odiphenols (2), peptides ( 2 ) ,diamines (12), hydroxyethylsubstituted diamines ( 2 ) , and polyamines containing more than two basic nitrogen atoms (2). Of these, the diamines form the most active Cu(I1) chelates, and have been explored to the greatest extent. The chelates of amino acids, peptides and phenolic ligands have somewhat less activity. The least active compounds listed in Table I are the Cu(I1) chelates of the polyphosphates and of ligands with more than two donor groups per mole-
Ligand
Sarin
1035 1005 985 1020 1010 950 1000 1000 1000 950 960 980 960 970 965 4950 4950 990 890 890 945 970 950 920 1000 995 1013 988 919 515 515 1005 510 990 990 990 900 5940 3020 3100 3160
960 965 960 965 935 855 1000 1000 1000 890 900 940 900 900 910 975 960 890 940 890 900 910 900 900 845 850 850 838 846 865 890 850 870 840 840 850 915 503 504 497 504
hit
kl (min. -1)
(mi=.)
>>loo >>loo >>loo >>loo
... . .. .. ....... . ... ... . . . ......,, ..... ...
> 100 > 100
.. .....
75 70 105 -95 w75 -75 -100 -60
,
....., .. ..,..... ...., . .. ..,.,... ........
..,.....
7.5
9.2 X 4 . 6 x lo-’ 1 . 4 X 10-l
1.5 5.0 -75 -90 -85 -75 -70 -60 -85 110 78 7.0 7.0 7.0 5.0 2.5 2.2 2.3 21 64 65 Inactive 100 Very large L’ery large Very large
........ .. . . .. . .. ,.. ... .. . .. . . . . .. . . .. ..... ...... . ........ .. . .. . .. ..... ... ... . ... ..... ,
, ,
,
,
,
,
,
,
,
1.4 X 2.8 X 3.2 X 3.0 x 3.3 x 1.1 x 1.1x
10-l 10-l
10-1 10-l 10-2
10-2 10-9
........ .. . , ... . . .. ... . .
cule. Not listed in Table I is a wide variety of chelate compounds which were not of sufficiently high activity to be interesting. These compounds R
R A ‘’
A2\ I R‘
OR
-0
c u/OH*
+
\+ / P
+
‘R
OH -
Products
3034
XE-ASPX ASPl CIT
cs cu
DAB DXP DIES DIPY DMES DTS
DPA EDT.1 EN GS GGG HXSPA HEDDA4 HEDT.4 HEN 2-HEX HIMD.4 2-HxG STA PDS PYR 533-4 TEEN TXEX TRIES TS TU
COURTNEY, GUSTAFSON, WESTERB.KK, HYYTIAINEN, CHABEREK AND MARTELL TABLE I1 KEYTO LIGAND ABBREVIATIONS N-aminoethylaspartic acid aspartic acid citric acid cis-l,2-diarninocyclohexane cis-A4-1,2-diaminocyclohexene 1,4-diaminobutane 1,3-diaminopropane dieth ylenetriamine dipyridyl S,N’-dimethyleth ylenediamine 1,8-dihydroxynaphthalene-3,6-disodiumsulfonate pyridine-2,6-dicarboxplic acid ethylenediaminetetraacetic acid ethylenediamine glycylglycine glycylglycylglycine If-hydroxyethylaspartic acid S,N’-dihydroxyethyleth ylenediamine-diacetic acid N-hydroxyethylethylenediarniiietriaceticacid S-hydroxyethylethylenediamine S,N’-dihydroxyeth ylethylenediamine S-hydroxyethpliminodiacetic acid X,X-dihydroxyethylglycine nitrilotriacetic acid pyrocatechol-3,5-disodium sulfonate pyridoxamine 5-sulfosalicylic acid
Vol. 79
ent that the more electropositive the Cu(I1) atom is, the greater will be its effect on the acceptor activity of the phosphorus atom. Therefore, one would expect that an electron donor which carries considerable negative charge would tend to decrease the rate of the reaction outlined above. The alternate reaction mechanism may be considered analogous to the push-pull mechanism advanced by Swain8to explain the action of pyridone in the mutarotation of glucose. Thus the hydroxy form of the Cu(I1) chelate may be considered to act as an amphoteric substance in the catalytic reaction, which may be formulated as
/ \
F R’
OR
OH?
A
R ‘ R
R
\ /
/ \
H
OR
R R S,N,N‘,N’-tetraethylethylenediamine The copper chelate would therefore have two func?.;,N,N’,N’-tetrametliylethylenediamine tions: I , the acidic function of accepting the fluotriethylenetetramine ride ion: and 2, the basic function of donating the trans- 1,2-diaminocyclohexane hydroxide ion. It is seen that the metal ion must truns-~4-1,2-diaminocycli~he~ene be as electropositive as possible for maximum catalytic activity. The more positive the metal ion in TABLEI11 the normal 1 : 1 chelate, the greater will be its tendCATALYTIC EFFECTSOF Cu(I1) CHELATES O N D F P HYency to hydrolyze to give the monohydroxy chelate DROLYSIS compound illustrated above. Also, the more posiDFP = 4.14 X 10-3 3r; PH 7.00; p = 0.1; T = 25.3” tive the metal, the greater will be its residual afCatalyst Concn , pmoles/l. t 1 / 2 (min ) finity for attracting the fluoride ion from the phosTMES 2070 15 phate group. DIPY 2070 23 The results listed in Table I indicate that the reE:i 20iO 60 action rate decreases with increasing negative HEN 2070 96 charge of the Cu(I1) chelate. The following lig2-HEiY 2290 250 ands are arranged in the order of decreasing cataDIES 2070 3000 lytic activities of their Cu(I1) chelates in the hyin-.lude the Cu(I1) chelates of ligands which are drolysis of Sarin: H E N > HxG > HIMDA > quite negative, or form very stable chelates, or ASPA > EDTA. A similar series may be taken both, such as the Cu(I1) chelates of EDTA, N T 9 , from the data of Wagner-Jauregg, et a1.,2 for the catalysis of D F P hydrolysis by Cu(I1) chelates: citrate and tartrate. Influence of Charge.-The qualitative observa- o-PHEN, E N > glycine, alanine >> EDTA. Furtion mentioned above indicates that the charge of ther examples are the greater activity of the Cu(I1) the complex may have a great deal to do with the pyrophosphate chelate over that of the correspondactivity of a copper chelate as a hydrolytic catalyst. ing tripolyphosphate compound, and the higher acThere are two general mechanisms by which a Cu- tivity of the Cu(I1) HEDDA chelate over that of (11) chelate compound may act as a catalyst in the HEDTA. Influence of Number of Donors in Ligand.hydrolysis of Sarin and DFP. The first involves the combination of the metal chelate compound Inspection of the catalytic activities of Cu(1I) with the substance being hydrolyzed, followed by chelates of various ligands listed in Table I indiattack of this complex with the hydroxide ion. cates that the highest activity is obtained with Thus the nucleophilic attack of the phosphorus bidentate donors, and that additional donor groups atom by the hydroxide ion is greatly increased by per mole of ligand reduce the effectiveness of the the interaction of the metal ion with the groups at( 8 ) C C S n a i n n n d J F H r n a u J r , THISJ O V R Y ~ I 7 4 , 2534 tached to the phosphorus atom. It is appar- (1952)
METALCHELATE COMPOUNDS AS CATALYSTS
June 20 1957
302 5
DMEN, DIPY and o-PHEN. It is notable that all of these substances form soluble 1 : 1 Cu(I1) chelates which do not disproportionate in aqueous solution a t pH 7 and above, whereas the 1: 1 Cu(I1) chelate of ethylenediamine is not stable and is quickly converted to the 2: 1 chelate with the precipitation of copper hydroxide. It is seen, therefore, that N-substitution of a 1,2-diamine is helpful in obtaining a stable aqueous catalytic chelate of Cu(I1). The influence of N-substitution on the specific catalytic activity of Cu(I1)-ethylenediamine chelates cannot be determined unambiguously from the examples available since changes of structure also affect the stability of the metal chelate compound. n n The influence of stability on catalytic activity is A A discussed below. \ /A Influence of Stability.--Although Wagner-Jaur\ /OH2 egg, et a1.,2stated that stability of a Cu(I1) chelate / \ A OH2 A -4 is not a significant factor in determining its cataA OHz W lytic activity, the evidence obtained in this research Bidentate Terdentate Quadridentate seems to indicate otherwise. Tt’hile it is certain Cu(I1) chelates formed with bi-, ter-, and quadridentate that many factors affect the catalytic activity of ligands chelate compounds, the evidence in Table I seems to from the substitution of alkyl groups on the basic indicate that catalytic activity generally decreases nitrogen atoms are nearly compensated for by co- with increasing stability of the metal chelate comordination of the hydroxyl groups so that the sta- pound. This generalization is supported by a numbilities in the former series do not differ signifi- ber of comparisons taken from the results of Table cantly.) I. Since the catalytic activities of the H E N and 21. I n the series of 1: 1 Cu(I1) chelates formed by H E N chelates are about half of that of EN, it is N-substituted ethylenediamines, the chelate staseen that a bidentate chelate compound containing bility constants decrease in the order: EX > two water molecules bound directly to the metal ion DMEN > TMEN. The order of catalytic activis apparently the most favorable structure for pro- ity is seen to be exactly the reverse. ducing catalytic activity in Cu(I1) chelate com2 . Although chelate stability constants are not pounds. There remains the possibility of cataly- available for monobipyridyl-Cu(I1) and for monosis on systems containing monodentate ligands. For o-phenanthroline-Cu(II), the order of decreasing the systems investigated, however, 1: 1 Cu(I1) stability of these compounds and of the correspondcomplexes with monodentate reagents were not ing E N chelate may be deduced from the published sufficiently stable t o prevent precipitation a t pH 7. stabilities of the chelates of other metal ions with Steric Factors.-It has been shown previouslyg these ligands to be: E N > o-PHEN > DIPY. I t is that the stability of a 1: 1 Cu7II) chelate compound seen that the order of increasing activity is the frequently is controlled by the tendency to dispro- same: E N < o-PHENlO < DIPY. portionate to the 2 : 1 chelate and copper hydroxide, 3. Comparison of the Cu(I1) chelates of the four as well as by its own intrinsic stability constant. diaminocyclohexenes and cyclohexanes with that of Thus a 1: 1 chelate with a high stability constant ethylenediamine shows no appreciable difference in may be unstable in solution if the 2 : 1 chelate is also catalytic activity, although i t is known that the very stable. Since Cu(I1) chelates are square stability of the EN-Cu(I1) chelate is 2 to 3 log K planar, the substitution of alkyl groups in place of units lower than those formed from the cyclic dithe amino hydrogen atoms would help to prevent amines. A possible explanation of this anomaly is this disproportionation reaction by steric repulsion based on the fact that the two types of structures between the substituents on two different moles of indicated below have equivalent electronic interacligand, as is indicated by the formula shown below. tion (binding energy) between the Cu(I1) ion and This effect has been observed in the case of the N- the basic nitrogen donor groups
metal chelate as a catalyst. Thus the hydrolysis of Sarin in the presence of the 1: 1 Cu(I1)-ethylenediamine chelate is three times faster than it is in the presence of the corresponding chelate of diethylenetriamine. It is, of course, difficult to separate the simple structural or steric effect of the number of positions of the Cu(I1) ion which are coordinated (see below), from allied effects such as the enhanced stability which results from the increased number of donor positions on the ligand. In the series of Cu(I1) chelates of EN, H E N and 2-HEN, there is no additional stability resulting from the relatively weak coordination of the auxiliary hydroxyl groups, in contrast to analogous series EN, D I E N and TRIEN. (The steric factors resulting
pc/*
(
R
R
R
\ / A-
i\ c u i
(
R
\ / N
)
/
CH2
CHZ
I CH2 \CH(
R”R ’ R \R substituted ethylenediamines such as TMEN, (9) A. E. Martell, R . Gustafson and S. Chaberek, Jr., paper no 34, International Congress on Catalysis, Philadelphia, P a . , Sept. 10-14, 1956 (to be printed in a special volume of “Advances in Catalysis.” Academic Press, Inc , hTew York).
NHz
NHn
-\/ CH I CH
\NHz
\ / OH2 cu
/ \OHz
I
OHz
/ \ cu
CHz
\
“2
OHn
The observed difference in stability, therefore. would be merely an entropy effect which depends on the initial and final states of the ligands in the formation of the metalchelate. Hence, the environment, or electronic condition of the Cu(I1) ion, would be (10) o-Phenanthroline data observed from ref. 3.
3036
A. MARTELL, s. CHABEREK,R . COURTNEY, s. l
the same in both cases, and the similarity of catalytic activity is not surprising. In view of these correlations with stability, one would look for the least stable metal chelate compound in order to obtain maximum catalytic activity. There is a definite limit beyond which this concept cannot be pushed, however, which is defined by the stability necessary to keep the metal ion in solution. Thus the most active catalyst would be the aquo ion itself. Since the aquo ion cannot be obtained in sufficient concentration a t a pH where the hydroxide ion concentration is also sufficient to carry out the reaction a t a reasonable rate, it seems that an important function of the metal chelate compound is to maintain the metal ion a t a reasonably high concentration (and activity) in aqueous solution. Influence of Metal Ion on Catalytic Activity.Measurements of the catalytic activities of a number of stable aqueous metal chelates of the Fe(III), La(III), Cr(III), Ti(1V) and Sn(1V) ions are listed in Table I. Not listed are a number of additional measurements made on chelates of other common metal ions such as Zn(II), Cd(II), Co(II), Fe(II), Ni(I1) and Pb(I1). None of the metal chelates, listed and unlisted, showed interesting catalytic activity. The Zn(I1) ion was unique in that it was the only third row metal for which no active chelates were found. On the other hand, a number of metal chelate compounds of oxo- metal ions such as ZrO(IV), U02(VI) and MoOZ(V1) showed very high catalytic activity toward the hydrolysis of Sarin, the most active being the ethylenediaminetetraacetato-zircony1 ion (f1l2 = 2.2 rnin.), the dioxomolybdenurn(V1) chelate of 3,5-disulfopyrocatechol (-7.0 min.), and the dioxouranium(V1) chelate of 3,6disulfo-1,S-dihydroxynaphthalene(-3.5 min.), the numbers corresponding to the approximate half times of hydrolysis in the presence of a 0.0010 molar solution of the catalyst. The lack of information on the stabilities of these chelate compounds and the small number of examples of very active chelates make correlation of activity with properties of the ligand impossible. It is noted, however, that all of the active catalysts contain one or more oxo groups bound to the metal ion. Although these chelates are in some instances formed with ligands containing many donor groups, it is possible to state in all cases that oxo- or hydroxygroups are also present. I t seems probable that these basic groups must be involved in the observed catalysis by a push-pull mechanism analogous [COSTRIBUTION FROM
THE
I T E S ~AND ~ ~ H.~HYYTIAINEN . ~ ~ ~
kfARTELL,
s. CHABEREK, JR.,
79
to that already proposed for the Cu(I1) chelates. Although the uranyl 5-sulfosalicylate chelate has been reported by Kyland, et u I . , ~ to be catalytically active, studies of systems containing this chelate indicate that it is unstable in aqueous solution at pH 7. General Conclusions.-The catalytic activities of metal chelate compounds listed in Table I are seen to increase with increasing pH and increasing concentration of the chelate. The measured rate constants increase with an increase in hydroxide ion and metal chelate concentrations, in accordance with the mechanisms suggested above. The rate constants do not increase with the first power of the concentration of these substances, and possible side reactions or equilibrium with inactive species is indicated. The authors are now engaged in an extensive study of these effects and hope to report their findings in a subsequent publication. The list of rate constants of D F P hydrolysis in the presence of various metal chelates in Table I11 is restricted to some of the more active catalysts with interesting structures. The results for these compounds, together with the results reported by Wagner-Jauregg, et al., and by Ryland, et aLI3indicate that the generalizations made above for the hydrolysis of Sarin also apply to DFP, although the latter compound is more slowly hydroly-zed by all reagents investigated. As a result of the correlation in this paper of the catalytic activity of metal chelates with structure and other properties of the ligand, the following generalizations may be made concerning the factors which tend to increase catalysis of hydrolysis by a metal chelate compound: 1, the presence of oxo, hydroxy or aquo groups bound directly to the metal; 2, bidentate ligands in the case of Cu(I1) (does not hold for metals of higher valence) ; 3, minimum stability of the metal chelate compound which will still prevent dissociation to the metal ion; 4, high concentration (hence high solubility) of the metal chelate; 5 , high pH. The most active catalysts found in this investigation in the order of decreasing activity are For Sarin: Cu(I1)ThlEN > Cu(I1)DMEN > ZrO(I1) EDTA > Cu(I1) D.4P > UOz(I1) (D9S)z > Cu(I1) DIPY > hl00di'I) PDS > CU(II) ?S > CU(II) 2-HEX > Cu(I1) H E N > Cu(I1) DIEN' For D F P : Cu(I1) T M E K > Cu(I1) DIPY > Cu(I1) E N > Cu(I1) DA4P> ~ C ~ ( 1H 1E ) N > du(11) 2-HEN > Cu(IT\ 2H E S >>> Cu(I1) D I E S
The tetramethylethylenediamine-Cu(1I ) chelate is the most active catalyst found up to the present time. WORCESTEX, MASS.
CHEWIC.4L LABORATORIES, CLARK L S I V E R S I T Y ]
Hydrolytic Tendencies of Metal Chelate Compounds. I. BY A. E.
LTOl.
R.
Cu(I1) Chelates1
c. COURTNEY, s. LVESTERBACK AND H . HYYTIAINEN
RECEIVED MARCH 4, 1957 The hydrolysis of 1: 1 chelates of copper with thirty ligands ha< been studied potentiometrically with a view to determining relative tendencies t o form hydroxy complexes. M o s t of the bidentate chelate compounds investigated were found t o undergo hydrolysis in a limited pH range, with pK values of 7 . 5 ?c 0.3. The number of donor groups in the ligand was found to be more important than stability in determining hydrolytic tendencies. I n certain cases the existence of a highly stable 2 : 1 chelate resulted in disproportionation of the 1: 1 chelate with simultaneous precipitation of half of the metal in the form of its hydroxide. (1) T h i s paper reports work done under coutract with the Chem;cal Corps, U . S . Army, Washinzton 2 5 , D. C
