The Detection of a Type of Reaction Intermediate ... - ACS Publications

gave a white shining crystalline precipitate on shaking. After being left overnight for com- plete crystallization, the compound was filtered, washed ...
0 downloads 0 Views 545KB Size
3498

EVERLY B. FLEISCHER AND

to a solution of 1.10 g. of zirconium isopropoxide (0.00284 mole) in benzene (40 g.) gave a white shining crystalline precipitate on shaking. After being left overnight for complete crystallization, the compound was filtered, washed with benzene (15 cc.) and dried under reduced pressure (2 mm.) a t room temperature. Anal. Calcd. for Zr(OOC.C&)4: Zr, 15.86; (OOCC6H5), 84.14. Found: Zr, 15.62; (OOC.CsHs), 84.3. Reaction of Zirconium Isopropoxide with Excess Phenol. -.I solution of 2.07 g. of zirconium isopropoxide (0.0053 mole) dissolved in benzene (66 9 . ) and 4.0 g. of phenol (0.0426 mole) gave a clear solution on shaking. The mixture was refluxed in a bath a t 110'. The amount of isopropyl alcohol in the distillate collected in 2 hr. (about 30 ml.) was 1.52 g. (0.0253 mole). Pure benzene was distilled out under a high reflux ratio. The clear solution was left overnight for crystallization. A white shining mass crystallized out. The crystals were filtered, washed with benzene and then dried as usual. Anal. Calcd. for Z ~ . ( C F , H ~ O ) ~ C ~ H ~ O Zr, H : 16.25; (OCsHs), 83.75. Found: Zr, 16.3; (OC6HS),82.06. Reaction between Zirconium Isopropoxide with Excess Benzoic Acid and Phenol.-To a solution of 1.88 g. of phenol (0.0200 mole) and 2.44 g. of benzoic acid (0.0195 mole) in benzene (69 g.) was added 1.59 g. of zirconium isopropoxide (0.0011 mole), and a clear solution was obtained. The mixture was refluxed in a bath a t 110'. The amount of isopropyl alcohol in the distillate (about 30 ml.) was 1.20 g. (0.0200 mole). Remaining benzene was distilled out under a high reflux ratio. The reaction mixture was left for crystallization, when a white mass crystallized out. The crystals were filtered, washed with dry benzene and dried in the usual manner. Anal. Calcd. for Zr(OOC,C6H6)a:Zr, 15.86; (00C.ChH5), 84.14. Found: Zr, 18.01; (OOC.C&), 81.48.

[CONTRIBUTIOS S o . 157;

FROM r H E

JUI

H. WANG

VOl. 82

Reaction between Zirconium Tetrachloride and Salicylic Acid in Benzene.-The addition of 1.86 g. of salicylic acid (0.0135 mole) to a suspension of 0.7 g. of zirconium tetrachloride (0.00303 mole) in benzene (28 g.) gave a white precipitate on shaking. The reaction mixture was refluxed in a bath a t 110" for nearly 24 hr. till the evolution of hydrogcn chloride gas almost ceased. The product was filtered, washed with dry benzene and dried under reduced pressure (2 mm.) in a bath a t 40" for nearly 2 hr. A n d . Calcd. for (OOC C&aO)Zr ( O O C , C G H ~ O H )Z~r :, 18.18; ( H O C F , I [ , ~ ~ ~81.99. ), Found: Zr, 18.28; (HOCd&COO), 81.28. Reaction between Zirconium Isopropoxide and Excess Methyl Salicylate in Benzene.-To a solution of 1.50 g. of zirconium isopropoxide (0.0038 mole) in benzene (60 g.) was added 2.93 g. of methyl salicylate (0.0193 mole), when a clear solution was obtained. The mixture was refluxed in a bath a t 110'. About 40 ml. of the distillate were collected sloivly between 71-80'' in the course of about 3 hr. The remaining benzene was removed under a high reflux ratio and the reaction mixture (15 ml.) was left for crystallization. A shining white crystalline mass separated out. After decantation, the product was recrystallized from benzene (10 ml.). Now the crystals were separated by decantation and dried under reduced pressure (2 mm.) a t a bath (50') for nearly 2 hr. The solid has a sharp melting point of 187". Anal. Calcd. for Zr ( O C G H ~ C ~ ~ C HZr, ~ ) ~13.11. : Found: Zr, 12.93.

Acknowledgments.-The authors are grateful to the Council of Scientific and Industrial Research for a research grant and assistantship awarded to one of them (R.X.K.). Thanks are due to hlagnesium Electron Ltd., London, for zirconium tetrachloride.

STERLISC CHEMISTRY LABORATORY, Y A L E UKIVERSITY, S E I V HAVEN, CONNECTICUT ]

The Detection of a Type of Reaction Intermediate in the Combination of Metal Ions with Porphyrins BY EVERLY B. FLEISCHER' AND

JCI

H. WANG

RECEIVEDDECEMBER 7 , 1959

A reaction intermediate in the combination of metal ions with porphyrins is detected and studied. A structure with the metal sitting-atop the flat porphyrin molecule is proposed for this intermediate compound. Spectroscopic and equilibrium data of these sitting-atop type of complexes are presented.

The metalloporphyrins, which are known to be very stable metal complexes, have been extensively studied both experimentally2 and theoretically. It has been demonstrated4 t h a t many metal ions in the porphyrin complex, such as F e ( I I ) , Fe(III), Co(III), Ni(II), etc., do not undergo detectable exchange with similar ions in solution. In order to extract these transition metal ions from the metalloporphyrins, i t is necessary to use concentrated hydrochloric or even sulfuric acid a t high temperature. Conversely, even in basic aqueous solutions it is often necessary to heat the reaction mixture in order to introduce a transition metal ion into the porphyrin ring.? Based on their kinetic studies, Phillips and co-workers6 suggested t h a t the reaction (1) National Science Foundation Predoctoral Fellow, 1959-1460.

(2) For general references, see R. Lemberg and J . W. Legge, H e m a tin Compounds a n d Bile Pigments, Interscience Publishers, Inc., NeuYork, 1949. (3) Discussioizs Favadny Soc., No. 26, 8 1 (1958). (4) S. Ruben, M .K a m e n , B. Allen a n d P. Nahinsky, THISJ O U R N A L , 64, 2297 (1942). ( 5 ) B. Dempsey, M. B. Lowe a n d J . N . Phillips, paper presented a t

between zinc ion and solubilized porphyrin esters to form the corresponding metalloporphyrin takes place, as indicated in the reaction below, through a displacement rather than a dissociation type of mechanismP Zn++

+ RH2 + 2H20

RZn $. 2H3OT

where RH2 represents the porphyrin. M'e want to report some results which show not only that the reaction between metal ion and porphyrin to form metalloporphyrin takes place through a displacement mechanism but also that the ineta1 ion first combines with the porphyrin to form a reaction intermediate which subsequently decomposes to the normal metalloporphyrin and two hydrogen ions. Our work shows that this reaction intermediate has a structure depicted in Fig. 1, with the metal ion sitting atop the flat porphyrin molecule. The t h e Symposium on Hematin Enzymes, Canberra, Australia, Aug. 31Sept. 4, 1959. (6) F. Basolo a n d R. G. Pearson, "Mechanism of Inorganic Reactions," John tViley and Sons, Inc., New York, N . Y . , 1958, p. 91.

INTERMEDIATES IN PORPHYRIN AND METALIONREACTIONS

July 20, 1960

3499

Zn(II), etc., first form a new type of complex, with absorption spectra markedly different from t h a t of the metalloporphyrins, which upon prolonged standing a t 60 O change to the corresponding metalloporphyrins. 610

WAVELENGTH ( C # ) bo0 5FO 500

20L

/

to-

Fig. 1.-Proposed

sitting-atop type of structure for the reaction intermediate.

evidences for the existence of this sitting-atop type of complex are summarized in the following. Visible Spectra.-The spectra of protoporphyrin and its complexes are shown in Fig. 2 . If equivalent amounts of Fe(II1)-salt and the watersoluble protoporphyrin derivative I are mixed in aqueous solution a t room temperature, the visible CH=CH2

('HZ

n

5

Fig. 2.-Visible

spectra of protoporphyrin derivatives:

(A) protoporphyrin dimethyl ester; ( B ) hemin dimethyl ester; (C) Fe(II1)-SAPP; ( D ) dihydrochloride of protoporphyrin dimethyl ester.

Fig. 2C shows the visible spectrum of such a complex formed by treating ferric chloride with the dimethyl ester of protoporphyrin in chloroform solution. If alcohol or pyridine is added to this chloroform solution, the spectrum changes immediately back to that of protoporphyrin, 2A. This shows CH, CH* that the binding between Fe(II1) and protoporphy/ / CH, CH2 rin in the new complex is quite weak, as the proto\ \ porphyrin readily can be displaced by other ligands. /c=o /c=o We suggest that this new complex has a sitting-atop HS sH type of structure illustrated in Fig. 1. It is also of \ \ CH, CHZ interest to note t h a t the spectrum of the sitting-atop / / complex Fe(II1)-protoporphyrin dimethyl ester CH2 CH, (hereafter referred to as Fe(II1)-SAPP) bears a \ \ SH2 KHg I striking resemblance to t h a t of protoporphyrin diacid, RHb++, as shown in Fig. 2D. This observaspectrum of the mixture is essentially the same as tion suggests t h a t the observed absorption is probthat of I which is almost indistinguishable from ably due to electronic transitions in the protoporthat of the dimethyl ester of protoporphyrin shown phyrin rather than the metal ion part of the comin Fig. 2A. Upon prolonged heating above 50°, plex. Similarly if the acetone solutions of the the spectrum of the mixture gradually changes to sitting-atop complexes of Fe(II), Fe(III), Co(II), that of a hematin derivative as t h a t illustrated in Pt(IV), Zn(II), respectively, are diluted with Fig. 2B. Spectroscopically no intermediate com- water, the spectra immediately change to that of pound was detected in this process. We found that the dimethyl ester of protoporphyrin itself, Fig. 1-4, the rates of the successive steps in this reaction as water molecules rapidly displace the organic ligcould be affected markedly by varying the solvent and in the complex. composition, presumably due to the change in solThe visible spectra of the sitting-atop complexes vation of the ions. For example, in acetone solu- of a number of transition-element ions are summations some metal ions, Ni(II), Cu(II), Bi(III), rized in Table I. The visible spectrum of protoporHg(II), Cd(I1) react readily with dimethyl ester of phyrin diacid, RHd++, shows two absorption maxprotoporphyrin even a t room temperature t o form ima a t 553 and 596 mp, respectively. I t is also of the corresponding metalloporphyrins, whereas other interest to note t h a t the spectra of uranyl and vanmetal ions, Fe(II), Fe(III), Co(II), Pt(IV), Sn(II), adyl complexes are quite similar to the others, al-

EVERLY B. FLEISCHER AND JUI H. WANG

3500

though the structures of uranyl and vanadyl ions are quite different from those of the other ions. Infrared Spectra.-The infrared spectra of protoporphyrin and Fe(II1)-protoporphyrin complexes are shown in Fig. 3. (p).

WAVELENGTH

1001

3

4

5

6

I

I

1

I

7

8

9

10

spectrum of hemin dimethyl ester, Fig. 3A, one observes t h a t all the above N-H peaks disappear as expected. The infrared spectrum of Fe(II1)SXPP,Fig. 3C, shows t h a t the vibrational frequencies 3.05 and 6.15,~are shifted to 3.11 and 6.64 p , respectively, and t h a t the 9.06, peak in 3B apparently disappeared in the spectra of Fe(111)SXPP. This 9.06 I-( peak has been assigned to a bending vibration similar t o a wagging type except that the groups swing as a unit back and forth in the symmetry plane of the molecule.* The observed frequency shift of the 3.05 and 6 . 1 5 ~peaks is due t o the polarizing influence of the Fe-(111) on the N-atoms to which the former is bound. Composition of the Fe-SAPP Complex.-The cornposition of the complex Fe-SAPP was determined by Job's method of continuous variationsgas follows. I n order to study the reaction A

U c

: 01 u ,

40

I

,

,

, I

I

1

,

,

01

,

I

1

5000

+ nB

I

I

I

X , of B, but of constant total stoichiometric concentration, M , of A and B, were prepared. If CI, C2and C, represent the concentrations of A , B, and

I

AIB,t,respectively, in the equilibrium mixture, we have c1 + cs = M(l - X) (1)

I 8

c*+ nc3 =

MX CIGn = K C

1400

1200

IO00

If we impose the condition t h a t dc'

dX

Several absorption peaks in the infrared spectrum of protoporphyrin dimethyl ester, Fig. 3B, have been assigned to vibrational modes of the N-H group.'J In protoporphyrin dimethyl ester, Fig. TABLE I I'ISIBLE SPECTRA OF SoSITTISG-ATOP COMPLEXES OF rrRANSITION ELEMENT IONS WITH PROTOPORPHYRIN DIMETHYL ESTER IN ACETONE SOLUTIOS

-

Wave lengths of observed absorption maxima

536 538

~t(n7)

Sn(I1) Zn(I1) 538 \'(IV)b U(S'1)" 536 hIain absorption peak.

*

(md 568" 610 560" 603 562" 605 563" 602 G44 561" 600 64 1 563" 600 645 558" 600 567" 600 Vanadyl salt. Uranyl salt.

3 B , the 3 . 0 5 ~peak is due t o the N-H stretching vibration, the 6 . 1 5 ~peak due to N-H deformation and the 9.Mp peak due to the N-H rocking vibration. IVhen the two N-H groups were changed to X-D groups by isotopic-exchange with deuterium oxide, the 3.05p peak shifted t o 4 . 1 7 ~ . If one compares the spectrum of Fig. 3B with the infrared (7) C . (1933). (8) H. "Infrared Co., In..,

S. Vestling a n d 1. R. Downing, THISJ O U R N A L , 61, 3511 Fowler, N. Fuson and J. R. Dannl, Determinations o f Organic Structures," D. Van Xostrand N e w York. N. Y., 1949. pp. 45, 57. 11. Randall, R . G

=

0, then equa-

tions 1, 2 and 3 can be combined to give 7%

Fe(1l) Fe( 111) Cr( 111)

(2) (3)

I

1600

2000

Fig 3.-Infrareci spectra of protoporphyrin derivatives: (4)hematin dimethyl ester; ( B ) protoporphyrin dimethyl cster; ( C ) Fe(II1)-SAPP.

Metal ion

AB,,

solutions of varying stoichiometric mole-fraction, ,

C

20t

Vol. 82

=

--

K

1-x

(4)

Thc valuc of X can be determined experinicntally by plotting a quantity Y which is proportional to C3 L I T . the stoichiometric mole-fraction X and taking the point where Y is either a maximum or minimum. I n the present work the quantity Y used as a measure of C:, is defined as Y=D-Do

where D is the optical density of the equilibrium mixture, and Do is the optical density the solution would have if no complex were formed. D O can be computed from the experimentally determined molar extinction coefficients of ferric chloride and protoporphyrin dimethyl ester, respectively, in chloroform solutions. Since the molar extinction coefficients of ferric chloride a t the wave length used in the present work, 506 and 568 mp, are very much smaller than t h a t of protoporphyrin dimethyl ester, epp, and that of the complex, ecomplex i t can be shown that Y has a maximum value if ~~~~~l~~ > cpp, or a minimum value if ~~~~~l~~ < epp. Figure 4 shows the plot of Yvs. X for solutions of protoporphyrin dimethyl ester and ferric chloride in chloroform. Fig. 4(a), a t 568 mp where ecompiex > cpp, shows a maximum a t X = 0.5; Fig. 4(b), a t 506 nib where eComplex < epp, exhibits a minimum at X = 0.5. Thus n, in equation 1,is equal to unity, which shows that the molar ratio of Fe(IT1) to PP in the complex is 1: 1. (9) W. C. Vosburgh aud G. P. Couper. Tms J O U R N A L , 63, 437 (1941).

July 20, 1960

0.21 0.0

0.0

INTERMEDL\TES I N PORPHYRIN AND LIETAL I O N

y',

,

,

,

~

0.5

,

,

REACTIONS

CI

~

3501

CI

1.0

CI I ; ;

X .

-LIT-

__c

y"

'

/

H Fig. 5.-Proposed

dimeric structure of the sitting-atop type of complex.

total amount of PP per cc. but all in the complex form, then

Do-D Do-D, ~

is equal to the fraction of total

PP in the complex form in the equilibrium mixture. Pure PP itself was found by our freezing point depression measurements to exist in monomeric form in bromoform, CHBr3, solution. By analogy we may infer that pure PP also exists in monomeric Fig. 4.-Graphical determination of n. form in chloroform solutions. Since the total Equilibrium Studies.-Since it was known that stoichiometric concentrations of protoporphyrin ferric chloride exists in dimeric formlo FezCls, dimethyl ester and ferric chloride are known, the when dissolved in poorly coordinating solvents like concentrations of all the individual species in the chloroform, equilibrium measurements were carried equilibrium mixture can be calculated. Table I1 summarizes the result of measurements out to determine if the Fe(II1)-SbPP complex also exists as a dimer in solution. The formation con- in chloroform solutions containing 2.5y0 by volume stant of Fe(II1)-SXPP was so large in pure chloro- of methanol. Of the four equilibrium constants form solution that i t was necessary to add some defined above, it is seen that only K I ,which is equal 0.2 ilf-l, remains essentially constant for ethanol to the system to reduce the formation con- to 3.9 stant somewhat for convenient equilibrium meas- solutions of varying concentrations of the reactants. urements. These four equilibria may exist in the This shows that in 2.5% by volume of methanol in chloroform solution, ferric chloride is in monomeric mixture: form and reacts with PP to form FeC13-SAPP. Fee13 + P P Presumably in this complex the Fe(II1)-ion is oc[FeC13-S.4PP] tahedrally coordinated to two of the N-atoms of PP, FeC13-SAPP, KI = (I) [FeCla I [PP 1 three chloride ions and a methanol molecule. hleasurements also were made in chloroform solutions containing 3.0% by volume of methanol. In these solutions K I still remains essentially constant but has a smaller value, 2.0 f 0.2 M-l.

*

Fe2ClsSA(PP)z, KIII =

Fe?Clc-SA(PP)z, K r =

[FeL!Ie-SA(PP)2] -- (111) [FeC13]*[PP]2 [FenCl&A( PP),] [Fe?CIe][PPI'

(I")

For simplicity, the participation of ethanol molecules in the above reactions is omitted. The concentrations of various species in the equilibrium mixture were determined by a spectrophotometric method. If D represents the optical density of an equilibrium mixture of PP, ferric chloride and the sitting-atop complex, and D , represents that of a solution containing the same (10) For references t o t h e literature, see 1 ' .V. Sidgwick, "Chemical Elements and Their Compounds," Vol. 11, Oxford Univ. Press, 1950, p. 1355.

TABLE I1 EQUILIBRIUM MEASUREMENTS IN CHLOROFORM SOLUTIONS CONTAINING 2.5% B Y VOLUMEOF METHANOL ISittinz. atopcomplex]

[Flee

(mM)

PPI (mM)

2.95 3.72 4.33 4.89 5.47 6.35

8.95 8.18 7.57 7.01 6.43 5.55

[Free FeCla]

KI

KII

lo-*

(mM)

X 10-8

X

0.87 1.05 1.40 1.79 2.17 3.20

3.79 4.33 4.07 3.91 3.92 3.60

2.54 3.94 4.58 5.46 4.68 8.24

KIII

KIV

X 10-0

x

24.3 25.1 18.0 15.6 14.1 10.1

4.25 5.27 5.30 5.57 6.10 6.50

10-6

The results in 1.0% by volume methanol in chloroform solutions are summarized in Table 111. Here i t is seen that K I no longer remains constant, ~ to approach a fairly constant value but K I seems

3302

ARTHURL. SMITH

VOl. 8,"

Experimental

over the range of concentrations studied. This indicates that as we reduce the methanol concentration, the equilibrium gradually shifts from I to I V and t h a t in pure chloroform solution the sitting-atop complex probably exists in the form of a A probable structure of dimer, Fe2Cl&h(PP)2. this dimeric complex is depicted in Fig. 5 .

The dimethyl ester of protoporphyrin was prepared by the method of Grinstein," with the exception t h a t recrystallized samples of hemin, supplied by Fisher Scientific Company, recrystallized (97 yo),were used instead of fresh blood as the raw material. The 1n.p. of the dimethyl ester of protoporphyrin so prepared is 221 O . The anhydrous ferric chloride has a m .p . of 282". A11 visible spectra were taken on a Cary Recording Spectrophotometer. The infrared spectra were taken with a TABLE I11 Perkin-Elmer Recording Infrared Spectrophotometer with a matched pair of 0.1 mm. sodium chloride cells. DeuterEQUILIBRIUM MEASUREMENTS IN CHLOROFORM SOLUTIOVS ated chloroform, 9970 CDCIa, TWS used as solvent to dissolve COSTAINISG1 Oyc BY T'OLUME OF M E T H A ~ O L the Fe(III)-SA4PP for infrared studies. .%I1 the measure[Sitting-atop ment were made at room temperatures of 20 to 23". complex] [Free PPI [Free F e C h ] The ru',N'-bis-(2-aminoethyl) diamide I was prepared by (mM) (mM) (mM) KI X 10-3 K I V X 10-6 dissolving the dimethyl ester of protoporphyrin in excess of 42 5 128 5 36 3 41 0 37 ethylenediamine and heating the mixture a t 100" for several 1 02 30 0 139 6 62 2 15 hours under nitrogen atmosphere in a sealed tube. The 26 1 150 1 34 2 12 7 30 excess ethylenediamine was removed by vacuum evaporation. The purified product I was completelj- ioluble in On the other hand if more than lOYc by volume of water.

methanol is added to the chloroform solution, the PP in Fe(II1)-SXPP is completely displaced by methanol molecules, and the sitting-atop complex can no longer exist in solution as shown by the visible spectra.

[CONTRIBUTION FROM

THE E L E C T R O N

+

Acknowledgment.-This work was supported in part by a research grant (USPHS-RG-44S3) from the Division of Research Grants, U. S. Public Health Service. (11) If.Grinstein, J , B i d Chem., 167, 515 (1947).

TUBEDIVISIOX, RADIO CORPORATIOX

OP rhERIC.4,

LANCASTER, PESiVSYL\'ASIA]

Preparation of the Compound 7MgO~2Sn0,*As,05 BY ARTHVR L. SMITH RECEIVEDDECEMBER 9, 1959 ;Inalysis of a portion of the system Mg0-Sn02-As205 by means of X-ray diffraction discloses the existence of the compound 7Mg0.2Sn02.A%s206.The compound may be prepared by heating the respective oxides in requisite proportion at 1200" for 4 hr. T h e characteristic X-ray diffraction pattern of the compound is given

Introduction Compounds containing all three type oxides, h40, ?.IO2 and hIz05, are relatively rare. The present compound was discovered during an investigation t o improve the luminescence of the phosphor 6h4g0.A4szOj:Kin by coactivation with tin(1V) oxide. Anomalies in the luminescence data prompted a study of the ternary system. Because the M g O SnOz and MgO As205 edges have been defined through the work of previous investigators, only that portion of the ternary which was necessary t o identify the new conipound was studied. The work of Tanaka,' Berezhnov,2 Coffeen3 and Coughanour, et a1.,4 on the magnesium oxidetin(1V) oxide solid-state reaction may be summarized as follows: Two compounds exist, 2 hIgO.SnOY and Mg0.Sn04. There is complete agreement t h a t the compound 2 M g 0 ,SnOz is thermally stable t o a t least 1550". Although Coffeen claimed thermal stability of the compound MgO. Sn02 above l O O O " , Coughanour, in agreement with Tanaka, found it t o be unstable in the region

+

+

(1) Yasuo Tanaka, J . Ciiriii. .For. J n p u i ! , 61, 1023 (1940), B d i . Chein. Soc. J Q ~ I I P16, Z , 428 (1941) (C..4 , 36, 27i6 (1941). ibld., 41, 4393 (1947) ( 2 ) A . S . Berezhnov, C o m p l . T e i i d . n c a d . sci. C.R..S S . 5 3 , 17 (19-$0), ( C . A , , 41, 1535 (19471.) ( 3 ) TX'. W. Cotieen, J . Ain. C r i . S o c . , 36,207 (1953). GL) I.. W.Coughanour, R. S.I i o t h , S. M a r z u l l o a n d IC. E . S e n n e t t , J . R i s u a i c h . S a t / . Riir. Slaizdaudr, 64, 1-19 (1933).

800-1000", decomposing to 2MgO .SnOn and SnO?. The present work, all conducted in this system a t 1000" or above, failed t o detect the presence of hlg0,SnOz. This further confirms the work of Coughanour and Tanaka. The AIgO reactions are more complicated than those of ?rIgO+Sn02 because of the volatility of arsenic pentoxide. Guerinj and T r a r nicek, et aZ.,6 have noted the tendency of MgO+ AszOj compounds t o lose XszOa O2 during heating. Guerin states that hIgO ,Xss05loses arsenic oxide a t 500" t o give 2hIg0.Xs20j, which, in turn, loses arsenic oxide t o form 3iZIg.O .;ls205 a t 9OO". He also considered the orthoarsenate t o be unstable but did not realize that its deconiposition led t o a new and definite compound. Travnicek. et nl., found that a t temperatures above 1000" the loss of arsenic pentoxide from the orthoarsenate leads to the compound GMgO ,=ls20j. This compound can be prepared directly by firing the requisite proportions of the oxides and is stable t o at least 125O". The present investigation supports Travnicek's findings, but the decomposition of the orthoarsenate has been found t o be reasonably slow (about 10 t o 20';; decomposition a t 1200" for 4 hr.).

+

+

(,5) H e n r i Guerin, Coii!pl. r m c i . , 204, 1740 ( 1 0 : 3 i \ . (6) hl. Travnicek, F. Kroger, T h e P. J . Botden a n d 1'. &lm Physirn. 18, 4 1 , 33 (19.52).