Dec. 20, 1958
CATALYTIC DECOMPOSITION OF HYDROGEN PEROXIDE
Dp from fundamentals have not as yet been very successful, but we do feel that in the absence of a fundamental theory i t is best to use the empirical theory as accurately as possible in terms of the matrices of Tanabe and Sugano with no additional approximations. The fine structure observed a t the 4A1g,4Egtwin peaks a t 24960 and 25275 ern.-' could be due to a variety of causes such as: (1) a resolving of the accidental degeneracy of these levels due to electrostatic interaction with their environments, ( 2 ) spin orbit interaction and (3) a crystal field that is not purely octahedral but has some distortion. The second possibility is unlikely for the following reason. The two main peaks are separated by about 300 cin-’. The fine structures of the 4G and 4D states of the free gaseous ion are about 10 and 50 cm.-I, respectively.l1 The spin orbit interaction responsible for this splitting, therefore, is too small to explain the effect. The third possibility is unlikely since if it were the case, splittings of comparable order of magnitude would be present in some of the other degenerate levels in the crystal field. The same argument could also be used against the second possibility. The first possibility then appears to be the most likely even though i t is a little difficult to predict an order of magnitude for the effect. The degeneracy predicted for the 4Algand 4Eglevels is purely accidental and is due essentially to our approxi-
[CONTRIBUTION E O .
1536 FROM
THE
6477
mations. A general interaction of cubic symmetry could remove this degeneracy. Another point worth mentioning is the occurrence of the very low peak in absorption a t about 26500 cm.-’. This peak is most likely caused by a transition from the ground state t o a state of spin 1/2. This spin change of 2 is doubly forbidden“ and would give rise to the relatively weak absorption which is observed. There is one such state which cuts across the Orgel diagram and has about the proper values of E a t the value of Dq for the other levels in the crystal field. This state is of *Tzgsymmetry and is designated as the 2F2(deb) state by Tanabe and Sugano; it has the *I state of the free ion as its parent. Unfortunately the value of E for this energy level is the root of a 10 x 10 secular equation, and we have not as yet calculated the values of E for this level as a function of Dq from our best values of B and C. Transitions between states of different multiplicity are “forbidden,”l’ consequently the probability of transitions from the ground state 6.S to states of fourfold multiplicity is small so the light absorption coefficients, E , would be expected to be small as observed, namely, 0.02. The doubly forbidden transition ‘% to the state *Tzgderived from 2Iwould be expected to give rise to an even smaller value of e as is observed, namely, 0.002. CAMBRIDGE, MASS.
DEPARTMENT OF CHEMISTRY, Y A L E
UNIVERSITY]
Further Studies on the Catalytic Decomposition of Hydrogen Peroxide by Triethylenetetramine-Fe (111) Complex and Related Substances BY RICHARD C.
JAR NAG IN^ AND JUI
H. WANG
RECEIVED JUNE 27, 1958 T h e rate-determining step in the catalytic decomposition of hydrogen peroxide by (TET&k)Fe(OH)*+ was determined by studying the hydrogen isotope-effect on the reaction rate. A similar study was also made on the enzyme catalase. The complex tris-(2-aminoethyl)-amine-Fe(III)was synthesized and studied for its similar catalytic action. T h e inhibition of (TETA)Fe(OH)*+-catalysis by cyanide was quantitatively investigated and correlated with some magnetic susceptibility data.
Introduction The catalytic decomposition of hydrogen peroxide by triethylenetetramine-Fe(II1) complex, (TETA)Fe(OH)*+, was studied previously.2 The mechanism of this catalysis may be represented by Fig. 1, in which the ferric ion is polarized in the strong ligand-field of the triethylenetetramine and forms octahedral complexes. In these complexes, four of the octahedral orbitals of Fe(II1) are used to form coordination bonds with the tetramine, the two remaining adjacent orbitals may be used to combine with hydroxide ions or hydroperoxide ions or both. The hydroperoxide ion, OOH-, is potentially capable of acting as a bidentate ligand and forming metal chelates such as compound I11 in Fig. 1. But as depicted in Fig. 1, the 0-0 bond in (1) National Science Foundation Predoctoral Fellow, 1957-1968. T h i s work was taken f r o m t h e dissertation of R. C. Jarnapin submitted to the G r a d u a t e School of Yale University in partial fulfillment of t h e requirements for the degree of Doctor of Philosophy, June, 1958. (2) J. H. Wang, THIS JOURNAL, 77, 4715 (1955).
compound I11 is too short to allow the maximum overlap of the bonding orbitals of Fe(II1) and those of the bidentate ligand. Consequently the 0-0 bond in compound I11 is under strain and rendered more reactive, because the energy consumed in breaking this 0-0 bond is partially compensated by the energy gained in forming the two stronger Fe-0 bonds. Accordingly it was found2 that (TETA)Fe(OH)2+ decomposes hydrogen peroxide with great efficiency; the measured activation energy was only 6.6 kcal./mole. The above mechanism is further supported by the observation that tetraethylenepentamine-Fe(II1) complex is practically inert for the decomposition of hydrogen peroxide as compared to (TETA)Fe(0H)2+. Kinetic measurements showed that the rate of oxygen liberation in the above catalysis is approximately proportional to the concentration of hydrogen peroxide up to [H20z] = 0.94 M . This shows that, a t least in the concentration range studied,
KICI~ARD C. JARNAGIN
Mi8
the transition from compound I1 to compound 111 cannot be the rate-determining step in the above catalytic cycle, for in that case a Michaelis-Menten type of saturation curve would have been observed. However, it was not possible to decide from the previous kinetic data whether the rate-determining step is the transition from compound I to compound I1 or from compound I11 back to compound I. Recently, with the help of doubly 018-labeled hydrogen peroxide as tracer, Jarnagin and SVang3 showed that the two 0-atoms in each O2 molecule liberated in the above Catalytic decomposition of hydrogen peroxide originate from the same H 2 0 2 molecule. This means that the last step in the above catalytic cycle, whereby I11 goes back to I with the liberation of Oa, involves hydride ion or hydrogen-atom transfer instead of oxygen-atom transfer. This observation suggests an investigation of the hyclrogen isotope effect on the rate of the present catalysis as a means for diagnosing the rate-determining step in the above cyclic mechanism. Hydrogen Isotope Effect Since the transition from compound I to conipound I1 involves the displacement of a complete hydroxide ion by a hydroperoxide ion, its rate should not be drastically reduced if deuterium is used to replace ordinary water as the solvent. On the other hand since the transition from I11 to I was shown to involve hydride ion or hydrogenatom transfer, its rate may be pronouncedly affected if deuterium oxide is used to replace ordinary water as the solvent. Preparation of SOlLItiOnS.-Two TET.A-Fe( 111) catalyst solutions, one in ordinary water and one in deuterium oxide, and two corresponding TETX-blank solutions were prepared. Tlie preparation and detailed composition of the catalyst solutions are given helow. : I conservative estimate indicated greater than 9.55; readily exchangeable deuterium in thc deuterium oxide sulutioris. Cataly,t si,ln. in ordinarv water
(a)
15 p l , 504;, H2SOIin IIaO
C a t a l y s t soln. in deuterium oxide (YS'",j (a) 15 p l . 50'3 H,SO,
99.55, D20
iti
Fen(SO4),, in HnO 2~111.99.5!;1 D20 75 nip. TET;\ (Matlie- ((1) t o mg. T E T X (Mathesoxlj son) (e) The mixture wab diluted (e) The mixture was dito 5.0 Inl. with H20 luted to 5.0 ml. with and adjusted to @H 99.5c0 D20 and ad10.0 with 50s: H&O, j'usted to pH 10.0 and 511Co EaOH in with 50yo H2S04and Ha0 50% KaOH in D20 The D20z solutimis were prepared by diluting a maximum of 900 p1. of 30c/;, H20, with 32.0 nil. of D20. The pH was adjusted t o 10.0 with 50y0KaOH in D10. Rate Measurements .-The rate measurements were made by the usual manometric procedure but reduced in scale. One ml. of catalyst or blank solution was mixed with 10.U ml. of peroxide solution in a Warburg-tl-pe apparatus of total volume of about 53 ml. Tlie pressure readings were taken a t 5 to 10 scc. intervals for ;L total prriod o f d x i u t 2 or 3 minutes. (c) ((1)
3
1111. H 2 0
I~C)
The data in Table 1 clearly show that the hydrogen isotope-effect on the rate of the present rcactioii is quite pronounced, corresponding to a factor of 2.2 in the rate constants. We may therefore conclude that the reaction of compound I11 in Fig. 1 (3) R . C. Jarnagin and J €1. LVanX, T ~ r r Js O U R X A I . , 80, 78li (1958).
A N D JUI
H. WANG
Lol.
so
TAULE I RELATIVE RATE5
IIlOi A \ U I I L C ) I31 ~ TETX-Fe(II1) CHELATE pH 10 0, teinp = 20°, [Fe(III)] = 1.94 X 10-6 il, [TETA4]= 9 1 X IO-* Jl,[H,Oz] or [D,O,] = 0 16'8 J l Kate f i x I-IcO (arbitrary) units)
H202
+
OF L)ECOMPOSITIOS OF
R a t e for U?O (arbitrary) units)
11.02
+
Rat? for H202 K a t e for ~ 2 0 2
2.3 2.2 1.9 2.3
4.40 4.73 5.38
10.1 1(J,2
10,2 10.5
+ H?O + 1120
4.56
-1v. 2 . 2
+k
0.17
with a second hydroperoxide ion to form compound I and 0 2 is the rate-determining step in the whole catalytic cycle. The hydrogen isotope effect on the rate of decomposition of hydrogen peroxide by the ferric enzyme catalase also was studied. The procedure is outlined beloy, the results are summarized in Table 11. TABLE I1 RELATIVE R A T E S O F 1)ECOMPOSITION O F
CATALASE AT 20"
ASD
1 1 2 0 2 AND Do02 B Y f 0.1
pH 7.0
[H:O?! or [I12021
(mole 1. -1)
0.185 ,184 .I8 ,1192 .091 ,089 . i 145 .045
Catalase
trary units)
trary units)
R a t e for DlOz DiO
11 6
6.5 8.4
5.4
1.8 1.4 2.3 2 .f) 2.4
4.8
:i.I)
11 9 11 6
89 13 14 6 5
1 1 2 3
10.6'
4.8 4.4
+
2.4 2.6 2.7 2.0 ,5r,:; L ) 2 0 alld 50% H20) 8.1 1. 3
Experiments with Catalase .--Two catalase solutions wcrc prepared from the crystalline enzyme obtained through Delta Chemical \Vorks. Catalase solution -1was prepared by dissolving 0.46 nig. of enzyme per nil. of cold phosphate buffer a t PH 7.0. .\fter standing in the cold for 36 hr., a filmy ppt. was thrown down. The ppt. was removed by centrifugation. C:ttalase solution B was prepared in a similar manner with 0.59mg. of enzyme per ml. of phosphate buffer in 99.556 DnO. The PH as measured with the glass electrode was 7.1 under conditions for which i t should have been 7.0 in ordinary water. Both solutions were stored under refrigeration. The procedure consisted of mixing 100 or 200 p l . of catalase solution with a peroxide solution made by adding 200 t o 300 pl. of 30% H202 to 10.0 ml. of a phosphate buffer prepared in either ordinary water or 99.57, D10. The rate of oxygen liberation was folloxed nianometrically as before.
The experimental uncertainty of the data in Table I1 is considerably larger than that in Table I. Nevertheless, it still seems rather certain that the hydrogen isotope effect on the rate of catalytic decoinposition of hydrogen peroxide by catalase is also quite pronou~~cc.tl.The explanation of this observation, as the tlcxtailed mechanism of catalase action, is unknown. Catalysis by Tris-(2-aminoethyl)-amine-Fe(111) Complex.-Tris-i"-aniinoethyl)-amine, (H&CH2CHnjn?;, forms a chelate compound with ferric ion. This chelate. which may be abbreviated as (TA%EA)-
CATALYTIC DECOMPOSITION OF HYDROGEN PEROXIDE
Dec. 20, 105s
Fe(OH)2+, has two readily exchangeable hydroxide ions attached to the ferric ion through two adjacent octahedral orbitals of the latter. Thus one may also expect (TAEA)Fe(OH)2+to catalyze the decomposition of hydrogen peroxide through a cyclic mechanism similar to that depicted in Fig. 1 for (TETA)Fe(OH)2+. Molecular models indicate that on the average the four coordinating groups in TAEA do not fit the octahedral orbitals of the ferric ion as well as the corresponding groups in TETA do. This could result in a weaker ligand-field and consequently a lower catalytic efficiency for (TAEA) Ge (0H) 2 +. Preparation of Tris-( Z-aminoethyl)-amine .-This 1% as prepared by a procedure similar to that of Ristenpart4 by the ammonolysis of X-(2-bromoethyl)-phthalimide a t 155 =IC l o ” , followed by the acid hydrolysis of the resulting triphthalimido compound. The glistening white needles of ( H3NCH?CH2)NC13 began to decompose a t 300’ without III II melting. Fig. 1.-The detailed mechanism of the catalytic decomposiT h e picrate formed from the trihydrochloride melted a t 228-300”. tion of hydrogen peroxide by (TETA)Fe(OH)2+. Rate Measurements.--llileasurement of the rate of the catalytic decomposition of hydrogen peroxide by TXE.4peroxide. This behavior parallels that of (TETA)Fe(II1) complex mas carried out by two different methods Fe(OH)Z+ catalysis and shows that for a mechanism with consistent results. I n the first method, 2.0-ml. samsimilar to that in Fig. 1, the second step cannot be ples of the decomposing hydrogen peroxide solution were the rate-determining one, a t least up to 1.0 M withdrawn a t convenient time intervals, blown into 1.5 F H2S04solution in order to quench the reaction and titrated H202, for in that case a saturation phenomenon for unreacted hydrogen peroxide with standard ceric sulfate should have been observed. The measured secondsolution, using o-phenanthroline-Fe( 11) as indicator. order rate constant a t 25’ for (TAEA)Fe(OH)2+ is In the second method, the catalytic decomposition of only slightly less than the value k = 1.2 X lo3 1. hydrogen peroxide 1% as follolved manometrically by means of a Warburg type of apparatus. mole-’ sec. for (TETA)Fe(OH)Z+. On the other The temperature of the reaction mixture was never ob- hand, a plot of In k vs. 1/T from the data in Table served to vary more than 0.2” during the course of the reacI11 yields an activation energy of 8.9 kcal. mole-’ tion. I n order t o correct for traces of Fe(II1) present as impurity which is considerably higher than the value 6.6 it1 the reaction system, a blank measurement with TXE.4 kcal. mole-l for (TETA)Fe(OH)z+-catalysis. and hydrogen peroxide but without added Fe(II1) was made Inhibition Studies for each temperature and hydrogen peroxide concentration studied. These blank rates were subtracted from the corThe possible inhibition of (TETA4)Fe(OH)2fresponding measured rates before the latter were used t o catalysis by fluoride, azide and cyanide, respeccompute the second-order rate constants.
The results of these measurements are summarized in Table 111. The second-order rate constant k is defined by the relation -d[H202]/dt = k [ H 2 0 2[T4EA-Fe(III)complex]. ] TABLE I11 SECOND-ORDER RATE COSSTANTS FOR THE CATAI,Y’?IC DECom=OsiTION OF H j O z B Y TSEA4-Fe(111) C O M P L E X PH 10.0, [TAEA] = 1.5 X mole/l.: [Fe(III)] = 8.9 X 10-7mole/l. [HzOzl (mole 1 . 3 )
Temp. (“C)
0.19 ,30 .41 .50 .75 .91 1 .00 0 439 ,446 ,434 ,436: 438 ,438
25 25 25 25 25 25 25 0.2 50 14 7 23 0 36 1 45 1
k X 10-3 (1. mole-’ see.-’)
0.62 .69
1 i
79 .85 .92 .17 .19 .:I8 .GO 1.3
2.2
Thus the catalytic deconiposition rate was found to be proportional to the concentration of hydrogen (4) E . Ristenpart, Ber., 29, ‘2526 (1896).
tively, was studied in this work, since these three ions were reported to inhibit catalase In a previous recent studyj2 it was reported that fluoride ion appears to activate (TETA)Fe(OH)z+ in its catalytic decomposition of hydrogen peroxide. A quantitative study on the effect of added fluoride was made in the present work. In order to get reproducible results, it was found necessary to mix the fluoride and (TETA)Fe(OH)z+solutions first and allow a sufficient time (ca. 5 minutes) to elapse before adding the hydrogen peroxide. The results are summarized in Table IV. The quantity “ratio of corrected rates” is defined as
Ratio of cor. rates = (rate with catalyst F-) - (blank rate (rate with catalyst) - (blank rate)
+
+ F-)
These data show that there was no activation of (TETL4)Fe(OH)2+-catalysis by fluoride. Indeed the data in Table IV do not rule out the possibility that fluoride may even have a small inhibitory effect. I t also was found in these measurements that the blank rates with fluoride frequently were as large or larger than the rate with catalyst in the (,j)R . 1,emherge and J. iT, I.egge, “Hematin Compounds a n d Bile Pigments,” Interscience Publishers, I n c . , New York, N. Y.,1949, p. -109. (6) J. B. Sumner a n d K . Myrback, “ T h e Enzymes,” Vol. 11, P a r t I . Academic Press. Inc.. New York. N. Y . . 1931. D. 422. (7) B. Chance, J . Bicl. C h e m . , 179, 1.99 ( l i 4 9 )
I~ICHAKD C. JARNAGIN
G4SO
We may also define the apparent forination constants hTt1 and IT3' as
These are iiot the true equilibrium constants, since [ E ] represents the total coiiceiitration of all the active species of the catalyst, e . ~ . ,(TETA4)Fe(TETA)Fe(OH)(OOH)+, etc. If we assume that both E1 and E12 are catalytically inactive, then the ratio of the inhibited rate R to the uninhibited rate Ro is given by
_____ 50 100 150 200 250 Total added cyanide mole l.-l). Fig. 2.-Inhibition of (TETX)Fe(OH)2+by cyanide.
0
-~
_L----.L---
0
absence of fluoride. This suggests that the apparent activation observed previously2 could be due to the failure to correct for the Fe(II1) which might be present in the sodium fluoride as an impurity. Similar studies were made on the effect of azide oii (TETA)Fe(OH)2+-catalysis. It was found that in the range PH 8 to 10, the azide ion has no marked effect on the rate of this catalytic reaction. TABLE I\' THEEFFECT OF FLUORIDE ON (TETA)Fe(OII)Z+ Temp. = 2'1.0 It 0.2", p H 10.0 i 0.2
0.41 .41 .3x .40
.25 ,082
7.3 7.3 13.1 7.3 7.3 7 .3
0.18 18
1.8 I
.x
3.4 2.(l 2.0 2 0
3 :3
.27 .27 27
0 .6
.8
.6 1.o
0.9 0 .0
On tlie other hand, cyanide ion was shown, as reported previously, to be a very strong inhibitor for (TETh)Fe(OH)2'. The data of the present cyanide-inhibition studies are summarized in Table V. TABLE 1. ISHIBITIOSOF ( T E T . ~ ) F c ( O H ) B~Y+ CYASIDE Temp. = 25.0 i 0.2', pH 9.0 i 0.1, [ F e ( I I l ) ] = 4.1 X 10-6 AI, [ T E T X ] = 9.1 X 10-3 ,lI,[ H 2 0 3 ] = 0.300 i 0.005 -11
Total a m t . of K C S a d d e d X 105 (for Wt . /I .)
IJ
1.OF) 1.82 3.63 7.28
10.9
Reaction r.,te (in arbitrar\Units. cor. for blank)
l'rital :imt. of K C S added X ,105 (for. wt:I.)
Reartion rate ( i n arbitrary u n r t s . cor. f o r blank)
20.3 i O . G 19.6 17.3 14.7
13, f i
6.52
15,s
6 . 04 5.112
11.:3
18.2 20.4 22.7
3.84 3.36
8.54
?'lie data iii Table 1- m:ty he interpreted in the following way Let [E] = concii. o f all cictii c species in tlle c:!talyst [ 11 = C O I I C of I ~ .free cyanide ion [ E l ] = concn. uf the moiiocyanide complex [ E l P ]= concii. o f the dicyanide complex.
The apparent forination constants, K' and K2', were evaluated from the experimental values of R!Ro by computing [I] for each of the inhibitioii measurements aiid graphically solving the resulting set of equations. In these computations we used Latimer's value of 4 X for the ionization coiistant of hydroq-anic acid.' For the present data, Kl' aiid Kz' were found to be 3.6 X l o 4 and 2.5 x l o 4 1. mole-', respectively. The measured rates R in Table are plotted *LIS. total cyanide concentration in Fig. 9 . The curve represents the values computed from the above equation with these experimental values of Kl' and K z ' , i.~.,R = Ro ' ( 1 -t 3.6 X 10'[1] 0 X los 111') Thc satisfactory rcyrcseiitation of the observed rates with the computed cur1.e in Fig. 2 supports our assumption that both E1 a r i d E12 are inactive. I t is known that the liydroliy, aquo aiitl halogen complexes of Fe(II1) exchange their ligands rapidly whereas the cyanide complexes do 11ot.~ It is possible that the substitution of one of the mobile ligands by a single cyanide ion n!ay produce a strong enough ligand-field to slow tlowii the rapid exchange by hydroperoxicle ioiis and heme upset the catalytic cycle. I n as much as a strong ligantl-field often leads to low-spin complexes,l i ' it see-i~iedadvisable to study the effect of ndc ! cyanide on the magnetic property of the presriiit catalyst.
+
Magnetic Measurements The moiar pra:iiagiietic susceptibilities of the TET,I-Fe(II1) conipleses were deteriiiined by nieaiis ol a conveiitioiiul Chuy apparatus with ;I coiiipeiisation tube. The illhomogeneous magnetic field was provided by a Varian 4-inch (V-4004) Electrol!iagrlet with .'-inch tapered pole caps. Tlie ~naxiiiiunifield strength in the j;S inch air gap was between 18 and 20 kilogmss. Tlie vessel teniperature was held a t 15 1 '. The calibration and checking oi the apparatus. the magnetic deterrninations a i d tlie coiiiputatioii of experimental data were carried out b y standard procedures. The results are sunriiarizctl i n Tables 1-1 and L7I1. I n these tables, lTT'ri-i~re.,entsthe difference i n weights of the saiiili!~i l l tlicl 1 ) i x tl :Lbsencc of thc.
*
L S J i\. 11. I..iiirn~~r. ' 0
(10) J . S. Griffith and 1 , . 11. Or:icl,
(1938).
L'iiiI
V:
