366
.
NOTES
@
Antimony
0
FI uori ne
Fig. 2.- -Schematic chain structure of SbF5polymer.
n
w Antimony
0
Fluorine
Fig. 3.-Active end of SbF5 chain.
lines overlap and a septet of 1:2:3:4:3:2:1 intensity distribution results. Peak C is split into a triplet of 130-cycle separation. by the type B fluorines. The splitting of this peak by the type A fluorines is unresolvable. The loss of detail in the spectra obtained a t room temperature and a t ca. 80” indicates that the fluorine atoms exchange at these higher temperatures. At room temperature, the average lifetime for the exchange between fluorines of type A and B is calculated to be about 2 X sec. and a t 80°, the average lifetime for any type of fluorine is less than see. 11. The Model in Terms of Electrostatics and Polarization.-It is assumed that all the Sb-F bonds are principally ionic bonds, with little covalent character. The distance between neighboring Sb+5ions is more than twice the Sb-F distance in the SbFe- ion not only as a result of cou-
Vol. 62
lombic repulsion of the Sb +&ions,but also because of the decreased covalent character in the bonds between antimony ions and the “bridging” fluoride ions. A “bridging” fluoride cannot form two bonds each with as much covalent character as that of a singly bonded fluoride. Figure 3 represents a terminal group of an SbF5 chain. This may be considered as a basic group which is capable of reaction with an acidic group (e.g., an SbFs monomer or the fluoride-deficient end of another SbF6 chain). It is proposed that a type (D) fluoride ion is a less basic site than a type (E) fluoride ion as its electron cloud has been attracted toward the central Sb+5ion more strongly than the others by virtue of a dipole induced in the SbC5 ion. This induced dipole in the Sbf5 ion is the result of a virtual positive charge in the direction of fluoride ion (F) and is caused by (a) the neighboring Sbf6 ion in the chain (which is not shown in Fig. 3), and (b) the fact that fluoride ion (F) is farther from the terminal Sb+5than the other five fluorides. Therefore the attachment of the acid group always occurs a t one of the cis fluoride ions (E); never a t the trans fluoride ion (D). 111. The Model in Terms of Atomic Orbitals.Although the bonds in an SbF6 chain are principally ionic, their fractional covalent character is of some importance. Antimony has available only four bonding orbitalsg (5s5p3), and yet it must form bonds with six fluorine atoms. The four bonds to the non-bridging fluoride atoms are more covalent in character than the other two, hence the most stable configuration will involve maximum utilization of the four available orbitals by these four bonds. Each of the three principal axes of antimony has available one and one-third orbitals for bonding (s’/~P). If the bridging fluorides are trans to each other, the one and one-third orbitals in the direction of their axis are misspent on what are necessarily very ionic bonds. However, if the bridging fluorine atoms are cis t o each other, it is then possible to utilize a full orbital in each of the . bonds trans to the bridging fluorides, and a total of only two-thirds of an orbital is misspent on the very ionic bonds. In the cis configuration, the fluorine atoms trans to bridging fluorine atoms are the most strongly bound to the antimony and hence closer to the antimony. The bridging fluorine atoms are the most weakly bound and hence are farthest from the antimony. Acknowledgment.-The authors wish t o thank Prof. K. S. Pitzer for his helpful discussions of the bonding in antimony pentafluoride. (9) The 5 d orbitals are theoretically available for ap3dZ hybridiza. tions, hut it ia unlikely that they participate appreciably.
CALCIUM AMMONIUM PYROPHOSPHATES BYEARLH. BROWN, WALTER E. BROWN, JAMES R. LEHR, JAMESP. SMITHAND ALVAW. FRAZIER Division of Chemical Development, Tenneasee Valley Authority, Wilson Dam, Ala. Received October 1% 1967
One amorphous and two crystalline phosphates of interest as potential fertilizer materials were prepared by ammoniation of products of the hy-
C ’
NOTES
March, 1958
367
'
the above-mentioned unit cell becomes b-centered, and a smaller unit cell can be selected with the symmetry Czh-P21/m or Cz-P21 and $he dimensions a = 6.85, b = 11.51, c = 6.57 A. and 0 = 110.2"; cell content, Ca3(NH4)2(P207)2-6H~0. The smaller translational distances of the subcell, along with the symmetry elements of (% of the larger cell, generate the remaining symmetry elements of C&,, which suggests that C&, is the more probable of the two possible subcell space groups. For this to be true, NH4+ and P20,4- must be in special positions in the subcell. The Ca++ may be in either general or special positions, but in either event there would be a 1-in-4 deficiency of calcium to fulfill space-group requirements. This may account for the diffuse reflections. All the crystals suitable for single-crystal X-ray study had overgrowths of the other calcium ammonium pyrophosphate. A comparison of the relative intensities of the diffuse reflections of two crystals with different amounts of overgrowth indicated only a remote possibility that the overgrowths could have caused the diffuse reflections. Prevalence of the overgrowths suggests that the two pyrophosphates may be structurally related or a t TABLEI X-RAY PATTERNS (Cu KLYRADIATION;CAMERA DIAM. least have similar unit-cell dimensions in two directions. 14.32 CM.;WEDGE-SHAPED SAMPLE) Both crystalline salts have been synthesized also Ca(NH4hPzOr. HzO CadNHdz(PzO7)z. 6H10 from tetrasodium pyrophosphate, calcium chloride d , A. I d,A. I I d,A. I and ammonium salts. 7.23 VS 2.52 M 7.27 VW 2.39 W
drolytic degradation of vitreous calcium polymetaphosphate. Chromatograms of 5% acetic acid solutions of the crystals were identical with pyrophosphate chromatograms. Optical and X-ray properties are reported here. When concentrated NH40H is added t o the viscous, water-immiscible liquid product of the treatment of vitreous calcium polymetaphosphate with water,' a precipitate is formed. Initially amorphous, it crystallizes as spherulites when left in the mother liquor several days. Similar crystals are formed slowly when vitreous calcium polymetaphosphate is treated with concentrated NH40H. Average observed mole ratios CaO :NH3 :Pz06: HzO of 1.13:1.90:1:2.30 in the air-dried product suggest the formula Ca(NH4)2P207.H20. The crystals are thin, colorless monoclinic blades, tabular on (100) or (001). The refractive indexes are a = 1.520, 0 = 1.537, y = 1.540. The crystals are biaxial (-) with 2V = 40" (obsd.) or 46" (calcd.), y = b. Bx, lies in the a-c plane and is inclined to the plate by 68". Interplanar spacings and visually estimated intensities are shown in Table I.
5.51 4.86 4.23 3.86 3.56 3.39 2.98 2.90 2.85 2.75 2.68
WM S WM W VW MS M M VW
2.44 2.39 2.32 2.11 2.06 1.52 1.88 1.72 1.69 1.47 1.45
W VW
VW VW VW W VW W WM VW VW WM WM
6.35 5.73 5.56 4.95 4.20 3 27 3.19 3.11 3.07 2.84 2.74 2.70 2.59
W
S M MS WM M M W S W VW VW VW
2.23 2.18 2.13 2.09 1.88 1.85 1.80 1.73 1.64 1.56 1.50 1.42 1.41
VW VW VW W VW W W VW VW VW VW VW VW
Exposure of Ca3H2(Pz07)~.4H2Oi to concentrated "*OH for several hours yields another product, Average CaO :NH3 :P205 :H2O mole ratios of 1.51: 0.99: 1:3.52 in four preparations indicate the formula Cas(NH4)z(Pz0,)2.6H20. The crystals are monoclinic tablets or plates, tgbglar on (001). Principal forms are (201), (201), (012) and (001)-modifying forms, { 010) and { 1101. The crystals are biaxial (-) with 2V = 60" (obsd.) or 61" (calcd.), OAP = 010. N , A a is 27" in acute p. The refractive indexes are a! = 1.520, 0 = 1.528, y = 1.531. The powder pattern is shown in Table I. Lattice constants, as determined from b- and c-axis rotation and Weissenberg patterns, are a = 7.67, b = 11.51,C' = 11.00 b. and 0 = 92.5'. The systematic extinctions, hQ1 with h I odd and OkO with k odd, indicate the space group Cih-P21/n. With a unit-cell content of 2 [Ca, (NH,) 2 (PZO7) z.6H20], the calculated density is 2.08 g./cc.--exactly the density calculated from refractive indexes. Reflections hkl with h 1 odd are weak or absent and are elongated parallel to c*. When these are ignored,
+
+
(1) E. H. Brown, J. R. Lehr, J. P. Smith, W. E. Brown and A. W. . , Frazier, THIS JOURNAL, 61, 1665 (1957).
VIBRATIONAL SPECTRA OF DIMETHYL ETHER I N THE LOWER FREQUENCY REGION BYYO-ICHIRO MASHIKO AND KENNETH S. PITZER Contribution from the Deparfment of Chemistry, Uniuersify of California Berkeley, Calzfornia Received October 1 7 , 1967
There is some ambiguity in the assignments of the two torsional oscillations of methyl groups about the C-0 axes of the dimethyl ether molecule.'S2 It is the purpose of the present study to 100,
0
I
I
150
zoo
I
1
I
250
300
350
I
400 420
c m:'
Fig. la.
ascertain these frequencies by further measurements of the Raman and infrared spectra. Only Ananthakrishnan3 observed the line a t 160 cm.-l in the Raman effect and there is disagreement (1) K.S. Pitzer, J . Chem. P h y s . , 10, 605 (1942). (2) G . Heriberg, "Molecular Spectra and Molecular Structure. 11. Infrared and Raman Spectra of Polyatomic Molecules," D. Van Nostrand Co., New York, N. Y., 1945, p. 353. (3) R. Ananthakrishnan, Proc. I n d . Acad. Sci., A S , 285 (1937).
