476
Vol. 73
NOTES
DL-phenylalanines by a method analogous to that described in this comm~nication.~The use of acetyl-m-phenylalanine in preference to other acylDL-phenylalanines is based upon previous studies on the role of the acyl group in determining the stereochemical course of the reaction under discussions,6and the fact that acetyl-DL-phenylalanine is sufficiently soluble in water to permit the use of relatively concentrated solutions. $-Toluidine was chosen as the base because it is a solid, is available in a relatively pure state, and is relatively stable in contact with air. The authors wish to acknowledge the assistance of W. H. Schuller who checked the procedure given below. Experimental' Reagents;-Acetyl-DL-phenylalanine, lustrous plates, tr1.p. 152-154 , was prepared in yields of 93-96% from DLphenylalanine (Dow) by acetylation of 0.8 mole of the amino acid with 2.4 moles of acetic anhydride ando1.6 moles of sodium hydroxide a t a temperature below 10 . Thirty grams of finely ground papain (Wallerstein) was stirred with 150 ml. of water for 3 hours at bo, the suspension centrifuged for 15 minutes at 2000 r.p.m. and the supernatant solution reserved for use. p-Toluidine (Merck and Co., Inc.) was used without further purification. Acetyl-L-phenylalanine-p-toludide (I) .-To 2 1. of a 0.5 M acetic acid-O.5 M sodium acetate buffer containing 9 g. of L-cysteine hydrochloride was added 155.5 g. (0.75 molel of acetyl-DL-phenylalatiine, the suspension warmed t o 50 to effect complete solution, 80 g. (0.75 mole) of p-toluidine and the above enzyme solution added, the total volume brought t o 3 liters with the acetate buffer, and the clear solution, pH 4.6, incubated at 40 O for 7 days. The reaction mixture was maintained a t 5' for 2 hours prior t o the collection of the precipitate which was washed with 1 1. of water and air-dried to give 102-106 g. (92-95%) of I, m.p. 215217". This product 'IWS dissolved in 1.25 1. of hot 969;. ethanol, the hot solution filtered, the filtrate held at 5 " overnight, the precipitate collected, washed with 500 ml. of cold 96% ethanol, and air-dried t o give 86-89 g. of I, m.p. "19'; [aIz5Df35 =tl o ( 6 , 4 in pyridine). i l n d . Calcd. for ClgHd3zN2(296): C, 73.0; H, 6.8; N, 9.5. Found: C, 72.8; €1, 6.7; S,9.4. An additional 9-10 g. of I, m.p. 219", mas recovered from the mother liquor to give a total yield of recrystallized 1 of SS-SSl;,. L-Phenylalanine (111.--A suspension of 86 g. of recrystallized I in 1100 ml. of 20'3 hydrochloric acid was heated under reflux for 16 hours, the clear solution evaporated to dryness in z~ucuo,the residue dissolved in 300 ml. of water, again evaporated to dryness, the residue dissolved in 250 ml. of aqueous amnionia cautiously water and 450 ml. of added t o this solution. After the reaction mixture was held a t 5 " for 2 hours the precipitated p-toluidine was collected, washed with cold water, the filtrate and washings combined, the solution extracted with two 300-ml. portions of chloroform, the volume of the aqueous phase reduced t o cu. 500 ml. by boiling, the solution cooled to 5' (2 hours). the precipitate collected, washed successively xith 100 ml. of water and 40 ml. of 96% ethanol and air-dried to give 28 g. of 11,lustrous flat plates; [ a ] %-34 ~ j= 1" (c, 2 in water). Concentration of the mother liquor gave two additional crops of I1 of 6 and 4 g., respectively, or a total yield of 38 g. (8270). D-Phenylalanine (III) .-The filtrate remaining after the collection of I was evaporated in vacuo below 50 O to one-half of its original volume, acidified with 120 ml. of concentrated hydrochloric acid, stored a t 5' overnight, the crystalline precipitate collected, washed with 200 ml. of cold water and recrystallized from 2070 aqueous methanol to give 50-65 gb (77-84'%) of acetyl-D-phenylalanine (IV), m.p. 162-164 . [ a ] %-32 ~ * 1" (c, 2 inmethanol). Asuspen.;ion of 5O'g. of IV in 300 nil. of 20% hydrochloric acid was heated under reflux for 5 hours, the clear solution evaporated to dryness, the residue tlisiolvcd in 300 ml. of water, 100 ml. .~
( 5 ) E. L. Bennett and e. Niemann, THISJOURNAL, 72, 1800 (1950). (6) E. L. Bennett and C. Niemann, ibid., 72, 1798 (19.50). (7) All melting points reported are corrected.
of 28% aqueous ammonia added, the solution boiled t o remove excess ammonia, decolorized with 5 g. of Norite, the clear colorless filtrate stored a t 5' overnight, the crystalline precipitate collected, washed successively with 100 ml. of water and 50 ml. of 96y0 ethanol, and dried to give 27 g. of 111; [ a ] = D 4-34 * .lo (c, 1 in water). A second crop of 3 g. of 111 was obtained from the mother liquor t o give a total yield of 30 g. (77Yo). The average over-all yields of D and L-phenylalanine from DL-phenylalanine were 59 and 68%, respectively. CALIFORNIA INSTITUTE OF TECHNOLOGY PASADENA 4, CALIFORNIA RECEIVED AUGUST7, 1950
On the Interaction between Hexacyanatoferrate(II1) Ions and (a) Hexacyanatoferrate(I1) or (b) Iron(II1) Ionslap* BY JAMESA. IBERSAND NORMAN DAVIDSON'~
In view of the occurrence of non-additive light absorption ("interaction absorption") in hydrochloric acid solutions containing iron(I1) and (111),3 it was of interest t o determine whether the hexacyanatoferrate(I1) and (111) ions exhibited this same effect. The present note reports experiments showing that the answer to this question is in the negative. At the same time we have examined the well-known brown coloration of solutions containing iron(II1) ions and hexacyanatoferrate(II1) ions and have evaluated the equilibrium constant for the formation of the uncharged complex species, FeFe(CN)e, responsible for this color. Figure 1 shows the absorption spectra of 0.1 F Fe(CN)e-3 and 0.1 F Fe(CN)6-4solutions buffered a t a pH of 6.8 (where the ratio [ H F ~ ( C N ) O - ~ ] / [ F ~ (CN)o-*] is of the order? of 0.003). The data agree fairly well with those given by Kortum.6 A 1:l mixture of these two solutions had optical densities in the wave length range 270-350 mp which agreed to within 1% with those predicted by Beer's law. No interaction absorption was observed when the ionic strength of the solution was raised from 1.9 t o 4.4 by making i t 2.5 F in potassium chloride. It was thought that such a high ionic strength might diminish the electrostatic repulsion between the Fe(CN)6-4 and F ~ ( C N ) Cions. -~ It is noteworthy that, by virtue of the short light paths used, the absorption of a moderately concentrated solution of Fe(CN)ep4andFe(CN)6-3 has been measured in the wave length region close to the strong ultraviolet "electron-transfer" absorption bands of the components; this is the region where interaction absorption, when it occurs, is most n ~ a r k e d . ~One . ~ may conclude that there is little or no optical interaction absorption between Fe(CN)sW4 and F ~ ( C N ) Cin- ~aqueous solution. It has already been suggested3 that interaction absorption ( l a ) Research supported by the 0. N. R. under Contract N6onr-244. (lb) To whom inquiries concerning this article should be addressed. (2) For a table o l the extinction coefficients of solutionsof E F e ( C N h . RbFe(CN)e, Fe(Cl0dn and FeFe(CN)a order Document 2944 from American Documentation Institute, 1719 N Street, N. W., Washington 6 , D. C., remitting $0.50 for microfilm (images 1 inch high on standard 35 mm. motion picture film) or 50.50 for photocopies (6 X 8 inches) readable without optical aid. (3) H. MsConnell and h-.Davidson, TxIs JOURNAL, 72,5657 (1950). (4) 1. h l . Rolthoff arid Ti'. J. Tomsicelc, J . Phys. Chcm.. 39, 935 (1935).
(5) G . Kortiim, Z. ghysik. Chcm., B83, 254 (1939). ( 8 ) J. Whitney and N. Davidson, THIS JOURNAL, 71, 3809 (1949). (7) H . LMcConnell and N. Davidson, ibid., 72, 3168 (1950).
NOTES
Jan., 1951 between iron(I1) and (111) requires a bonding "bridge" between atoms of the two oxidation states; such a bridge is not possible for the case a t hand. Measurements of the variation of the optical density with concentration of Fe+3and Fe(CN)a-S were made a t the wave lengths 475, 500 and 525 mp for solutions a t a total ionic strength of 1.00 * 0.02 (moles/liter) and a t [H+] = 0.102 M . The data have been interpreted assuming the approximate validity of the mass action law a t the constant ionic strength used. Let a and b be the concentrations of free Fe+3and Fe(CN)6-S,and a0 and bo the formal (total) concentrations of these components. Let cij be the concentrations of the complexes [Fei+3(Fe(CN)6)j;3]3(i - j), eij their extinction coefficients, and Kij their formation constants, so that cij = Kijaibj. One set of results was obtained for a series of solutions in which a> b so that a G ao, and so that only the complexes [Fei+3(Fe(CN)6)1-3]3i- a are likely to have been present. Let D, be the optical density for a 1.0 cm. light path of a solution due t o the complexes present, ;.e., corrected for the light absorption by Fe+3and Fe(CN)6-3. Then D e = seilci, (ZeilKila'bo)/(l ZKi, a') and (&/D,) = (1 + ZKj, ~~)/(Bci,Ki, ai-') (1) Assuming that ellKll > ~ i ~ K -i ~ a(i~>~ 1) (;.e., that the contribution t o D,from the 1: 1 complex is greater than that from higher complexes), one expands the denominator of the right side of the above expression
+
(eiiKii)(abo/Dc) = 1
+ (Kii - :::E)+ a' a
--
X
477
4000
1000
300
u; 100
30
10 250
300
350
400
450
500
550
Vmd. Fig. 1.-Extinction coefficients (ecl = log,, (lo/I))of: 0.1006 F &Fe(CN)e, flH 6.8 (phosphate buffer; the absorption spectrum was the same in water and at (H+) = 0.50 M) ; 0.1043 F K,Fe(CN)e, PH 6.8 ( A , date of Kortum M FeFe(CN)s for the above two substances); 1.01 X (see text); 0.0141 F Fe(C10& in 0.5 M HC104. Data obtained with 1.00, 0.10 and 0.010 cm. cells.
It is important t o note that due to some decomposition the optical densities of the solutions used
Figure 2 exhibits plots of (&/De) vs. a for a series b X lo2. of solutions a t 26 * 2' containing Fe(C10a)3, 0 2 4 6 8 K8Fe(CN)6,NaC104 and HC104. The non-dependence of ab@, on bo justifies the assumption that 25 only complexes containing one Fe(CN)6-a are important for these solutions. The linear nature of 20 the plots demonstrates that a complex containing one Fe+3ion is formed (the points a t low a are es15 pecially convincing in this respect). The intercepts of the curves give values of 1/~11K11. How- 2 30 10 5 ever, as may be seen in equation ( 2 ) , the apparent x X ,--. linear nature of the plots for larger a does not prove 4 25 5 6 that there is no significant formation of 2 : 1 com\ plexes; it is possible that the positive and negative 2 v 2 s 20 terms in the coefficient of a 2 approximately cancel each other. The ratio of slope to intercept for a 15 curve is K1, - (e&&'ellKll). There is no systematic dependence of this ratio on wave length (Table 10 I) ; this suggests that the e21K21/e11K11term is small and that the entries in the second column of the 5 ~ P I I II1lI I I I1 I I I I I I I I I l l I t I I 1 I l l , 1 I I - J table are essentially K11. Values of ell calculated 0 2 4 6 8 on this basis are also given. Le
h
TABLE I COMPLEXING OF Fe(CN)s-8 BY EXCESS Fe+3 IllP.
A,
475
500 525
CllKll
x
10-8
17.4 11.9 8.5
KI1
-liter/moles (~zlK;l/~llKll~, 21.5 19.4
21.4
809 611 397
a x 102. Fig. 2.-A, plot of (abo/D,) vs. a for a0 > bo in a perchlorate medium: 0, bo = 1.01 X lo-* F; 0,bo = 5.05 X 10-4 F; A, bo = 3.36 X lo-' F. The points at a = 4.13 X 10-* M and a = 1.12 X M were obtained with 10 cm. cells. B, Plot of (&/De) YS. b for bo > ao in a nitrate medium: 0, a0 = 9.6 X 10-4 F; 0,QQ = 4.8 X 10-4 F.
NOTES
478
for the above measurements changed with time ; freshly prepared solutions were measured within 5 minutes of mixing and corrected to zero time using the following observed increases in D per 5-minute interval: 475 mp, 1%; 500 and 525 mp, 2.5%. Other small corrections were: (1) using 21 for K11, a rather than its first approximation, U Q , has been used in applying equation ( 2 ) ; ( 2 ) the values of a and of the contribution of FeII’ to the optical density have been corrected for the hydrolysiss of Fe+3to FeOH++. The extinction coefficients of the complex FeFe(CN)6,for 280-550 mp, given in Fig. 1 were calculated from the absorption spectrum of a solution that was 0.00705 F in Fe(ClO&, 0.0056 F i n K3Fe(CN)B and 0.50 &! in HC104. The concentration of FeFe(CN)a of 1.01 (+0.02) X M i n this solution, calculated from the optical densities a t 47.5, 500 and 525 r n p and the extinction coefficients of Table I, corresponds t o a K11 of 36 in 0.50 AI HC104. A second set of data was obtained for bo > uo where one should consider the complexes Fe(Fe(CN)6-.3)j3- 3j. These solutions a t 26 * 2’ and a t an ionic strength of 1.OO contained Fe(C104)3, K3Fe(CN)6,NaN03 and 0.102 ;lf HN03, since potassium perchlorate precipitated from solutions in which sodium perchlorate and perchloric acid were used to maintain the acidity and ionic strength. Figure 2 shows that (a&/D,) is a linear function of b; the parameters of the straight lines are given in Table 11. The apparent values of E&II and K11 TABLE I1 COMPIEXINGOF P e T 3 B Y E X C E SFe(CN)h-: ~ SOLUTION^ A , mfi
(Intercept) -1 X 10-1
Slope/ intercept
475 500 525
12 9 9 3 6 25
11 8 10 3 s 2
I
17 15 12 5
.ij
-76 3ii
+
=
1
(9) E. H. Swift, THISJOURNAL, 81, 2682 (1929). (10) E. 1%.Sw5ft. “ A System of Chemical Analysis,” Prentice-Iiall, Inc , S e w York, N. Y . , 1946, pp. 438-441. CONTRIBUTION S O . 1476 FROM OF CHEMISTRY THE GATESAND CRELLIN LABORATORIES
PASADENA, CALIF.
RECEIVED SEPTEMBER 30, 1950
Shifted Position of Infrared Absorption Bands Associated with the C=O Linkages in Polycyclic Compounds BY MARIELOGISEJOSIEN
AND
+ A‘V + b (Kll - 6 N - -)
BlZK12
+ bZ
EllILl
Accepting equation (3) and the values of sl1Kll of Table I, the intercepts recorded in Table I1 have been used to calculate N . From the ratio of slope to intercept one can then calculate K11 (neglecting the eUK~/el1Kllterm). The results for N agree as well as can be expected. The average deviation of the K11 values is greater than that of Table I, and the average (15) is 25% less than that of Table I. These discrepancies and the trend in the calculated values of K11 with wave length may be due t o contributions of the (el&z/tllK11) term. We feel that all of the results are sufficiently concordant t o indicate the essential validity of the interpretations given here. Experimental.-FejC10r)3~10H20was prepared by evaporation of a mixture of FeCt36Hz0with ex( 8 ) Using 10-1 lor t h e hydrolysis constant, E R a t > i n o w i t < h aiicl W. I1 Stockniayer, ’l‘ms JOTR ~ A T 64, , 3% (194:)
NELSONF u s o ~
Although a great deal of research has been niadc on the Raman and infrared band assigned to the C=O stretching vibration in long chain and single ring compounds, spectroscopic study of the C=O linkage in polycyclic compounds is still relatively unexplored. Among the studies which have been made in this latter field may be cited those of Dulou,’ Kohlrausch,2 Biquard,3 Lecomte,* Flett,6 Ti. N. Jones,6and Josien and Fuson.’ I n the course of an infrared spectroscopic study of some polycyclic hydrocarbons and their oxidation products the following values for the carbonyl group absorption bands have been obtained which we believe hare not yet been reported in the literature: Compound&
K
are smaller than those of Table I ; an attempt has been made to resolve this discrepancy by considering the formation of FeN03++,with formation constant N , so that (FeN03++)= iVu(N03-). Because 6bo = 1. The theionic strength was 1.0, (NO,-) equation that results is (ellKll)(aob/D,)
cess 60% HCIO4 and recrystallization. It was free of C1- and C103-. Iron(II1) was determined as recommended by Swift.g Ferro- and ferricyanide solutions were prepared from recrystallized K4Fe(CN)6*3&0 and GFe(CN)@and were determined as recommended by Swift.Io
IU A-ITKATE
I ’
0
VOl. 73
Pyreiiequinone-3,8 Pyretiequinone-3,10 -4cetylpyrene Chrysenequinone-1,2 Acenaphthenequinone 9-Fluorenone-1-carboxylic acid
Wave length (in microns) of band assigned t o t h e C=O stretching vibration
6.10
ti, 10 6.02 6.03 5.63 and 5.79 5.73 and 6.00
These spectral wave length values were obtained for crystalline samples prepared in paraffin oil paste form. A comparison of these results with other results given in the literature cited suggests the following comments: (a) We have assigned the two bands, 5-73 p and 6.00 p, of 9-fluorenone-lcarboxylic acid to the carboxylic acid C=O and the cyclic ketone C=O, respectively. The first assignment is in agreement with the usual band location for carboxylic acidsg; the second is without doubt located a t such a position because of the effect of conjugation. (b) From the appearance of two bands, 5.&3 p and 5.79 p, in acenaphthene(1) Dulou, These, 1933, Bordeaux, France. ( 2 ) Kohlrausch, 2.Elekfrochcm., 43, 282 (1937). (3) Biquard, Bull. soc. Chim. France, 7, 895 (1940); 8, 5.5, 725 (1041). (4) J. Lecomte, J . Phys., 6, 239 (1945). ( 5 ) Flett, J . C h e m Soc.. 1411 (1948). (ti) R. N. Jones, el u l . , THISJOURNAL, 70, 2024 ii948); 71, 2 4 1 (1949); 74, 813and 956 (1950). (7) h1. I>.Josien and N. Fuson, Com9t. r e d . , in press. (8) Tliese compounds were prepared b y Dr. St. Elmo Brady, Pro. fesaor of Chemistry, Fisk University, Nashville, Tennessee. (9) 11. &I.l
