J . Phys. Chem. 1987, 91, 671-674 TABLE 111: Deviations (A) of Atoms from the (111) Plane at O ( 3 )
Na( 1)
0.46 0.46
Ai?( 1)
Ag(5) O(2)
3.76 0.17
that this distinction is quite clear and that the cluster may accurately be viewed as a nearly linear (Ag3)0 molecule (Ag(4)Ag(5)-Ag(6) = 153"), each atom of which is coordinated to an Ag' ion at Ag(2), Ag(l), and Ag(3), respectively. See Figure 1. On the 3-fold axes of the unit cell near the centers of the six-ring planes, 7.4 Na+ ions at Na( 1) and 0.6 Ag+ ion at Ag( 1) are found. The single (Na,Ag)-O contact distance observed, 2.28 (1) A, must then be an appropriate average of the corresponding N a - 0 distance, 2.32 (1) A, in Na-AZ2 and Ag-0 distance, 2.25 (2) A, in Ag,,-A.1*2 The Na( l),Ag( 1) thermal ellipsoid is elongated normal to the six-ring plane, consistent with this averaging. It appeared from our previous work14 that hydrogen treatment at 350 "C had converted many Ag+ ions to Ago atoms by the reaction
Downloaded by STOCKHOLM UNIV on September 7, 2015 | http://pubs.acs.org Publication Date: January 1, 1987 | doi: 10.1021/j100287a035
2Ag+
+ H2
-
2Ag0
+ 2Ht
The approximate unit cell composition determined here, (Ag+)l 8(Na+)7 4(H+)z,sAgy,8-A, indicates that the following reaction has occurred. 2.8Ag'
-
+ 1.4H2
2.8H'
+ 1.8Ag0 (in the zeolite) + l.OAgo (out of the zeolite)
In summary, in the structure of Ag,,,Na7 4-A, dehydrated, treated with H2. and evacuated, all at 350 "C, 0.6 (Ag6)3+cluster
67 1
may be present per unit cell. This cluster may be viewed as a trisilver molecule in the large cavity of the zeolite, stabilized by coordination to three Ag+ ions (and the oxygens of one four-ring). On 3-fold axes near six-rings (Table III), 7.4 Na+ ions and 0.6 Ag+ ion are found. Also, 0.6 Ag' ion at Ag(2) and 0.6 at Ag(3) are associated with the oxygens of two different eight-rings. Altogether approximately 7.4 Na+ ions and 3.6 silver species are located per unit cell. The remaining 1.0 Ag+ ion per unit cell has been reduced and has migrated out of the zeolite framework to form small silver crystallites on the surface of the zeolite single crystal, accounting for its charcoal black color. The composition of the crystal could not have been Ag4Nas-A, as was intended, before treatment with hydrogen. This could have been due to our neglect of zeolitic water molecules in calculating the weight ratio of hydrated Na12-A and AglZ-A to be mixed in the preparation step. One reason for this neglect is that the water content of hydrated Ag12-A is not accurately known. It is also possible that the single crystal studied was not in proper equilibrium with the bulk sample. Acknowledgment. This work was supported by National Science Foundation (NSF-INT83-16234) and also by the Korean Science and Engineering Foundation. We are indebted to the University of Hawaii Computing Center. Registry No. Ag,6Na7,4Si12A112048, 105121-68-4. Supplementary Material Available: Observed and calculated structure factors for dehydrated Ag,,6Na7,,Si,zAl,20,s,Ag+-exchanged zeolite 4A treated with hydrogen at 350 "C (1 page). Ordering information is given on any current masthead page.
Crystal Structure of Fully Dehydrated, Partially Ag+-Exchanged Zeolite 4A, AS,.~N~,.,-A. Ag+ Ions Prefer 6-Ring Sites. One Ag' Ion I s Reduced Yang Kim Chemistry Department, Pusan National University, Pusan 607, Korea
and Karl Seff* Chemistry Department, University of Hawaii, Honolulu, Hawaii 96822 (Received: February 4, 198.5; In Final Form: August 13, 1986)
The structure of partially Ag+-exchanged zeolite 4A, Ag7.6Na4.4-A,vacuum dehydrated at 370 "C, has been determined by single-crystal X-ray diffraction techniques in the cubic space group Pm3m (a = 12.311 (1) A) at 24 (1) "C. The structure was refined to the final error indexes R1 = Rz (weighted) = 0.064 by using 266 independent reflections for which 1, > 30(1,). Three Nat ions occupy the 3 k i n g sites, and the remaining ions, 1.4 Na' and 6.6 Ag', fill the 8 6-ring sites; each Ag' ion is nearly in the [ 1111 plane of its 3 O(3) ligands, and each Na' ion is 0.9 from its corresponding plane, on the large-cavity side. One reduced silver atom per unit cell was found inside the sodalite unit. It presumably formed by the reduction of a Ag' ion by an oxide ion of a residual water molecule or of the zeolite framework. It may be present as a hexasilver cluster in one-sixth of the sodalite units or, most attractive among several alternatives, as an isolated Ag atom coordinated to four Ag' ions in each sodalite unit to give (Ag5)4+,symmetry 4mm.
Ag+ ions in zeolite A can be reduced by heating,l., by reaction with reducing agent^,^ or by the sorption of metal atoms.4 (Many reports of the reduction of Ag' by these methods in other zeolites can be found.) Recently the structures of dehydrated Ag,Nal,-x-A treated with H, at room temperature and at 330 "C were determined in an effort to learn more about the reduction of silver ions in partially Ag+-exchanged zeolite A.335 In the (1) Kim, Y.; Seff, K. J . A m . Chem. SOC.1977, 99, 7055-7057. (2) Kim, Y . ;Seff, K. J . Am. Chem. SOC.1978, 100, 69894997. (3) Kim, Y . ;Seff, K. Bull. Korean Chem. SOC.1984, 5, 135-140. (4) McCusker, L. B. Ph.D. Thesis, University of Hawaii, 1980.
0022-3654/87/2091-0671$01 S O / O
structure of dehydrated Ag6Na6-A396treated with 50 Torr of H2 at room temperature, 1.27 (Ag3)+ clusters and 0.7 (Ag3)2+clusters per unit cell were found in the large cavity. In the structure of Ag,,,Na,,,-A, vacuum dehydrated and treated with H2 at 350 "C, (Ag6)3+clusters of low symmetry were found in the large ~ a v i t y . ~ Gellens et al. proposed, from their X-ray powder diffraction studies ( 5 ) Kim, Y . ;Seff, K. Bull. Korean Chem. SOC.1985,6,202-206; J . Phys. Chem., preceding paper in this issue. (6) The nomenclature refers to the contents of the Pm3m unit cell. For example, Ag6Na6-A represents Ag6Na6Si,2Al,,0,,, exclusive of water if a hydrated crystal is considered.
0 1987 American Chemical Society
672 The Journal of Physical Chemistry, Vol. 91, No. 3, 1987 of variously treated dehydrated Agf-exchanged zeolite A, that linear Ag32+clusters have formed: each consists of an Ago atom, opposite a 4-ring in the sodalite cavity, bonded between two Ag+ cations located at the framework 6 - r i n g ~ . ~Hermerschmidt .~ and Haul8 verified by EPR studies the presence of octahedral Ag6"+ (n < 6) clusters in the sodalite cavity of dehydrated Ag+-exchanged zeolite A treated with hydrogen, and their results were duplicated by Grobet and Schoonheydt.' Preston and Morton identified Ag2+, Ag32+,and octahedral Ag6+ in zeolite A by EPR spectroscopy.10 The direct observation of the reversible redox couple Ag3*+ Ag30 in Ag+-exchanged zeolite A by FTIR was reported by Ozin et aLI1 Scholler et al. investigated the influence of monovalent cations in different positions in zeolite 4A on the diffusivity of trans-2butene.], They concluded that Ag+ ions preferentially occupy 6-ring centers and that Na+ ions prefer 8-ring sites.13 Similar results were obtained by Nitta et al.,I4 who studied the site selectivity of Ag+ ions in dehydrated Ag,Na,-A by calculating cation-lattice interaction energies and charge-transfer stabilization energies for Ag+ and Na+ cations in 6-ring and in &ring sites. This work was done to determine the cation distribution crystallographically, to see whether complete dehydration of Ag7,6Na4,4-Acould be achieved at 370 "C without generating Ag atoms, and to learn the structure for comparison with others of similar composition evacuated at other temperatures or treated with H,.
Downloaded by STOCKHOLM UNIV on September 7, 2015 | http://pubs.acs.org Publication Date: January 1, 1987 | doi: 10.1021/j100287a035
*
Experimental Section
Kim and Seff determined by a least-squares refinement of 15 intense reflections for which 19' < 28 < 24", is 12.311 (1) A. Data collection was done by methods described previously," except that only one unique region of reciprocal space was examined at a scan rate ( w ) of 2 deg m i d in 28. Standard deviations were assigned to individual reflections according to the formula, u(Z) = [w2(CT
+ B1+ B2) + (pZ)2]1/2
where o is the scan rate, C T is the total integrated count, B1and B2 are background counts, and the intensity Z = w(CT - B , - B2). A value of 0.0218 was assigned to the empirical parameter p to account for the reduced reliability of the more intense reflections. The intensities were corrected for Lorentz and polarization effects;]' the contribution of the monochromator crystal was calculated assuming it to be half-perfect and half-mosaic in character. An absorption correction (w = 2.3 mm-l) was judged to be negligible and was not applied.20 All 891 reflections for which 28 < 70' were examined by countermethods. Of these, only the 266 for which I L 3u(Z) were used for structure solution and refinement. Structure Determination
Full-matrix least-squares refinement was initiated by using the atomic parameters of the framework atoms ((Si,Al), 0(1), 0(2), and 0(3)), and of the cation positions at Ag(1) (6-ring) and Na(1) (8-ring) from the structures of AgI2-A'q2 and Na,,-A.,' Anisotropic refinement of framework atoms and Ag( 1) and isotropic refinement of Na( 1) converged to the error indexes.
First, the complete Ag+ exchange of a sample of zeolite 4A powder was accomplished by a static method. Two grams of R1 = clFo- ~ F c l ~ / =0.13 ~Fo zeolite 4A (Union Carbide, Lot 494107701 161) was allowed to exchange at 24 OC with a 7-fold excess of 0.1 N AgN03, and the R2 = [ Z W ( F ,- I F c 1 ) 2 / x ~ F , Z ] = 1 / 0.14 2 mixture was agitated periodically. Each day, the supernatant solution was decanted and a fresh aliquot of 0.1 N AgNO, was A subsequent difference Fourier synthesis revealed a large peak added. After 8 days, the zeolite, now Ag12-A,6 was filtered and 11.6 (6) e A-3 in height at (0.0, 0.0, 0.17). This is the previously dried. observed neutral silver atom p o s i t i ~ n , ~ , *and , ~ *was ~ ~ stable ~ in Samples of Ag,,-A and Na12-A were mixed in a 2:l mole ratio, least-squares refinement. Positional, occupancy, and anisotropic neglecting water content. To this mixture were added a few large thermal parameter refinement of the framework and of the Ag( l), single crystals of zeolite A which had been prepared by Charnell's Na( l), and Ag(2) positions converged with the error indexes, R1 method,ls with enough water to submerge all solid particles so = 0.064 and R2 = 0.060. that at equilibrium the large crystals would have the composition The occupancies of these ions per unit cell were refined to Ag( 1) Ag8Na4-A. After 3 days, the water was allowed to evaporate at = 6.83 (7), Na(1) = 2.75 ( 2 5 ) , and Ag(2) = 1.03 (1). In zeolite room temperature. A structures, 12 monopositive cations, or in this case 12 cations A single crystal, 80 km on an edge, was selected and lodged or Ag atoms, should be found per unit cell. Furthermore, the in a fine glass capillary. The clear, colorless, partially hydrated, 6-ring sites are usually favored, and all eight are usually filled, Ag+-exchanged crystal was dehydrated for 48 h at 370 "C and more than the 6.83 (7) found at Ag(1). When the thermal el1.5 X 10" Torr and was then sealed off in its capillary by torch. lipsoid at Ag( l ) became extremely elongated in subsequent reMicroscopic examination showed that the crystal was gray, neither finement, and because the expected stoichiometry would now black nor silver. require some Naf ions to be at 6-ring sites, the 3-fold-axis position Diffraction intensities were subsequently collected at 24 "C. was separated into Ag( 1) and Na(2) and refined with the conThe space group Pm3m (no systematic absences) was used straint that the sum of occupancies be 8.0. By this method, 6.6 ~ ~ ~ ~ ~Ag+ ~ ions at Ag( 1) and 1.4 Na' ions at Na(2) were found at 6-ring throughout this work for reasons discussed p r e v i ~ u s l y . A Syntex 4-circle computer-controlled P i diffractometer with a sites. To allow the occupancies to sum to 12, that at Na( 1) was graphite monochromator and a pulse-height analyzer was used fixed at 3, its maximum value, and that at Ag(2) at 1. throughout for preliminary experiments and for the collection of A final difference Fourier synthesis was featureless except for diffraction intensities. Molybdenum radiation ( K a I , X = 0.70930 one small peak, 1.3 (3) e k3in height at (0.34, 0.5, O S ) , which A; Kaz, X = 0.713 59 A) was used. The unit cell constant, as was unstable in least-squares refinement. The final refinement using anisotropic thermal parameters for all positions except Na(2), which was refined isotropically, con(7) Gellens, L. R.; Mortier, W. J.; Schoonheydt, R. A,; Uytterhoeven, J. B. J . Phys. Chem. 1981, 85, 2783-2788. (8) Hermerschmidt, D.; Haul, R. Ber. Bunsenges. Phys. Chem. 1980, 84, 902-907. (9) Grobet, P. J.; Schoonheydt, R. A. Surf. Sci. 1985, 156, 893-898. (10) Morton, J. R.; Preston, K. P. J . Magn. Reson., in press. (11) Ozin, G . A.; Baker, M . D.; Godber, J. J. J . Phys. Chem. 1984, 88, 4902-4904. (12) Scholler, R.; Notzel, P. Adsorpion of Hydrocarbons in Zeolites; Academy of Sciences of the GDR, Central Institute of Physical Chemistry, GDR: Berlin, 1979; Vol. 2, pp 126-136. (1 3) A discussion of the structure and of the nomenclature used for zeolite A is available. (a) Seff, K. Acc. Chem. Res. 1976, 9, 121-128. (b) Broussard, L.; Shoemaker, D. P. J . A m . Chem. Soc. 1960, 82, 1041-1051. (14) Nitta, M.; Ogawa, K.; Aomura, K. Zeolites 1981, I , 30-34. (15) Charnell, J. F. J . Cryst. Growth 197, 8, 291-294. (16) Seff, K.; Mellum, M. D. J . Phys. Chem. 1984,88, 3560-3563.
(17) Kim, y . ;Seff, K. J . A m . Chem. SOC.1978, 100, 175-180. (18) Peterson, S. W.; Levy, H. A. Acta Crystallogr. 1957, 10, 70-76. (19) Principal computer programs used in this study. Data reduction: Ottersen, T. COMPARE, University of Hawaii, 1973. Ottersen, T. LP-76, University of Hawaii, 1976. Full-matrix least-squares: Gantzel, P. K.; Sparks, R. A.; Trueblood, K. N. UCLA L S ~ American , Crystallographic Association Program Library (old) No. 317 (modified). Fourier program: Hubbard, C. R.; Quicksall, C. 0.;Jacobson, R. A. ALFF, Ames Laboratory Fast Fourier, Iowa State University, 1971. Johnson, C. K. ORTEP, Report ORNL-3794, Oak Ridge National Laboratory, Oak Ridge, TN, 1965. (20) Kim, Y.;Seff, K. J . Phys. Chem. 1978.82, 925-929. (21) Subramanian, V.; Seff, K. J . Phys. Chem. 1978, 81, 2249-2251. (22) Kim, Y . ;Seff, K. J . Phys. Chem. 1978.82, 921-924. (23) Kim, Y . ;Seff, K. J . Phys. Chem. 1978, 82, 1307-1311.
The Journal of Physical Chemistry, Vol. 91, No. 3, 1987 673
Partially Ag-Exchanged Zeolite 4A TABLE I: Positional, Therma1,O and Occubancy Parameters X
0 0 0 1118 (6)
0 1900 (3)
0 2296 (104)
Y 1841 (5) 2188 (13) 2954 (9) 1118 (6) 4314 (31) 1900 (3)
0 2296 (104)
811or Biao
Z
3711 (3)
5000 2954 (9) 3390 (9) 4314 (31) 1900 (3) 1715 (16) 2296 (104)
26 (4) 71 (17) 44 (14) 29 (6) 168 (74) 65 (3) 45 (10) 10 (l0)b
822
27 41 26 29 105 65 45
812
833
(4)
(15) (8) (6) (39) (3) (10)
14 (4) 31 (14) 26 (8) 47 (10) 105 (39) 65 (3) 37 (18)
0 0 0
813
0 0 0
occupancy factore
823
0 (51 ., 0 33 (20)
17 (14)
lO(10)
0 88 (5) 0
0 88 (5) 0
10 (10) -20 (93) 88 (5)
0
24d 12 12 24 3 6.6 (1) 1 1.4 (1)
OPositional and anisotropic thermal parameters are given X104. Numbers in parentheses are the esd’s in units of least significant digit given for the corresponding parameter. The anisotropic temperature factor is exp[-(Pl1h2 &k2 83J2 P12hk &hl 823kl)]. bIsotropic thermal parameters in units of A2. COccupancy factors are given as the number of atoms or ions per unit cell. “Occupancy for (Si) = 12 and occupancy for (AI) = 12
+
+
+
+
+
TABLE II: Selected Interatomic Distances (A) and Angles (deg).
Downloaded by STOCKHOLM UNIV on September 7, 2015 | http://pubs.acs.org Publication Date: January 1, 1987 | doi: 10.1021/j100287a035
(Si,Al)-O( 1) (Si,AI)-0(2) (%,AI)-O( 3)
O(l)-(Si,Al)-O(2) O(l)-(Si,Al)-O(3) 0(2)-(Si,A1)-0(3) 0(3)-(Si,A1)-0(3)’ (Si,Al)-O(l)-(Si,Al)’ (Si,Al)-0(2)-(Si,Al)’ (Si,Al)-0(3)-(Si,Al)’
1.644 (6) Ag(l)-0(3) 1.656 (5) Ag(2)-0(3) 1.686 (8) N a ( l ) - O ( l ) Na(1)-0(2) ~a(2)-0(3) Ag(I)-Ag(2) Ag(2)-Ag(2)’ 109.1 (7) 113.3 (5) 107.7 (4) 109.4 (6) 149.8 (1 1) 158.4 (8) 143.8 (7)
Ag(1)-Ag(2)-Ag(1)’ Ag(1)-Ag(2)-Ag(1)” 0(3)-Ag(l)-0(3)’ 0(3)-Ag(2)-0(3)’ 0(1)-Na(2)-0(3) 0(3)-Na(2)-0(3)’
2.283 (9) 2.84 (1) 2.74 (4) 2.37 (4) 2.46 (11) 3.316 (4) 2.99 (2) 89.7 (6) 172.1 (6) 119.9 (4) 61.1 (3) 63 (1) 107 (4)
“Numbers in parentheses are estimated standard deviations in the units of the least significant digit given for the corresponding value.
verged at R1 = R2 = 0.064. The goodness of fit, [Xw(Fo IFc1)2/(m- s ) ] ’ / ~ ,is 2.69 where m (266) is the number of observations, and s (34) is the number of variables in least squares. All shifts in the final cycles of refinement were less than 20% of their corresponding esd’s except that of the isotropic thermal parameter of Na(2) which shifted 36%. The full-matrix least-squares programI9 used in all structure determinations minimized X W ( A I F ~ )the ~ ; weight (w)of an observation was the reciprocal square of u, its standard deviation. Na+, and (Si,Al)l 75+ Atomic scattering factors for Ag+, Ago, 0-, were sed.^^,^^ The function describing (&AI)’ 75+ is the mean of the Sio, Si4+, AlO,and A13+ functions. All scattering factors were modified to account for the real component cf’) of the anomalous dispersion c ~ r r e c t i o n . Final ~ ~ ~ positional, ~~ thermal, and occupancy parameters are presented in Table I; bond lengths and angles are given in Table 11.
Discussion The 6.6 Ag+ ions at Ag( 1) are at the same position as was found in dehydrated Ag12-A.’~2Each Ag’ ion lies on a 3-fold axis and is nearly at the center of a 6-ring (Table I11 and Figure 1). Each is trigonally coordinated at 2.28 (1) A to three O(3) framework oxygens. As compared to the sum of the Ag+ and 02-radii, 2.58 A,28these bonds are quite short and therefore quite The three Na+ ions at Na( 1) are associated with 8-ringoxygens. These positions are located in the plane of the 8-rings but not at their centers so that favorable approaches to framework oxygens may be made. Each N a ( 1) ion approaches one O(2) oxygen at 2.37 A and two 0(1)oxygens at 2.74 A. At Na(2), 1.4 Na+ ions occupy a 3-fold-axis site near 6-rings and are recessed 0.9 A into the large cavity from the [ 11 I] plane (24) Doyle, P. A.; Turner, P. S. Acta Crystallogr., Sect. A 1968, 24, 390-397. (25) International Tables for X-Ray Crsytallography; Kynoch: Birmingham, England, 1974; Vol. IV, pp 73-87. (26) Cromer, D. T. Acta Crsytallogr. 1965, 18, 17-23. (27) Reference 25, pp 149-150. (28) Handbook of Chemistry and Physics, 63rd ed.; Chemical Rubber Co.: Cleveland, OH, 1982/1983; p F170.
Figure 1. Stereoview of a large cavity of dehydrated Ag7.6Na44-Ashown using ellipsoids of 20% probability. The Ag+ and Na’ ions are distributed within their equipoints of partial occupancy in a plausible manner. This arrangement is seen in about 60% of the unit cells (ca. 40% if Ag6 clusters coordinated to eight Ag+ ions have formed in one-sixth of the sodalite units. See ref 2 for a stereoview of Ag, clusters). The remainder have six ions at Ag(1) and two at Na(2), not seven and one, respectively, as shown here.
TABLE III: Deviations of Atoms (A) from the [ l l l ] Plane at O(3)“ O(3) 1)
0.20 0.06
Na(2)
0.90 -2.78
‘A negative deviation indicates that the atom lies on the same side of the plane as the origin.
at O(3) (Table 111). Each Na(2) ion is tri onally coordinated to three O(3) framework oxygens at 2.46 . For comparison, the sum of the conventional Na+ and 02-radii is 2.29 A.28 The fractional occupancies observed at Ag( 1) and Na(2) indicate the existence of at least two types of unit cells with regard to the 6-rings. For example, 60% of the unit cells may have seven Ag+ ions at Ag(1) and one Na+ ion at Na(2), and the remaining 40% would have six Ag+ and two Na+ at these positions, respectively. More extreme compositional variations are possible but are considered less likely unless Ag6 clusters have formed.Iv2 Sodalite units with hexasilver clusters would be expected to have eight Ag’ ions at Ag( l), leaving other sodalite units to have seven (ca 40%) or six (ca 60%) Ag+ ions, complemented by one or two Na’ ions, respectively, as above. The Ag(2) position is very similar to that of the neutral silver atom in the structure of dehydrated partially decomposed Ag,2-A.1~2The distance between Ag(2) and the nearest oxygen at O(3) is quite long, 2.84 (2) A, as compared to the corresponding ion-to-ion distances observed, and about the same as was seen before between a neutral silver atom and framework oxygens.’.2 The shortest available Ag(2)-Ag(2) distance, 2.98 (2) A, may b e considered slightly longer t h a n t h e Ag-Ag bond in silver metal, 2.89 A.29 One may conclude that an average of 1 reduced silver atom per unit cell has formed inside the sodalite unit, presumably by the reaction a Ag+ ion, the 12th ion, the ion in excess of the 11 of which can be accommodated at the 8 6-ring and 3 8-ring sites, with half of an oxygen atom (oxide ion) of the framework or of a residual water molecule. This neutral silver atom was also
1
(29) Reference 28, p F176.
J . Phys. Chem 1987, 91, 674-685
674
Downloaded by STOCKHOLM UNIV on September 7, 2015 | http://pubs.acs.org Publication Date: January 1, 1987 | doi: 10.1021/j100287a035
Figure 2. Possible isolated Ag atom coordinated to four Ag' ions to give (Ag5)4', symmetry 4mm. Each Ag' ion coordinates to the three O(3) oxygens of a 6-ring.
seen in other structures such as T16,5Ag5,5-Adehydrated at 440 0C23and Ag12-A treated first with H 2 and then with 02,both at 330 "C2* It may be that a neutral hexasilver cluster has formed in one-sixth of the sodalite units.'s2 It is also possible that an isolated atom exists in each sodalite unit, coordinated to at least 2.6 Ag' ions. This is the number of the original 6.6 Ag+ ions remaining if the maximum number of Ag+ ions per sodalite unit (4) position themselves far from the Ago atom. It appears far more reasonable that the maximum number of Ag+ ions would position themselves near the Ago atom to give (AgJ4+, symmetry 4mm (see Figure 2). Ag, and ( A E ~ ) ~then, ', seem to be the two most reasonable interpretations of the crystallographic result. Other clusters involving metal-metal bonds such as (Ags)6+ (the result of two (AgJ4' units sharing two Ag+ ions) or (Ag12)8' (the result of four (Ag5)4' units sharing all eight Ag+ ions) remain possible but seem less likely and too numerous to discuss. The cation which has been reduced is the one which is most easily reduced, Ag', and the cation site which is no longer occupied is the least favorable, the one opposite a 4-ring in the large cavity.
This position is the least satisfactory because the approach of the Ag' ion to framework oxygens has been the longest (possibly virtual) and the most one-sided in previous ~ t u d i e s . ' ~ ~ ~ ~ ~ The ionic radius2*of Ag', 1.26 A, is much larger than that of Na', 0.97 A. From a consideration of ionic radii only, one would expect the larger Ag+ ions to associate with the larger rings, the 8-rings. However, Ag+ ions are all found in 6-ring sites as predicted by the calculations of Nitta et ai.'4 and as demonstrated less directly by the experiments of Scholler et a1.I2 AgI2-A dehydrated at 350 O C has three water molecules in each sodalite unit, together with three Ag+ ions and no reduced silver atoms.30 Ag,,-A dehydrated at temperatures L 400 "C for about 2 days has no residual water and contains reduced silver atoms.',2 In this work, we observe 370 "C is a sufficiently high temperature to have brought about the complete dehydration and the partial reduction of Ag+ in Ag,,6Na4,4-A. Perhaps the formation of the silver atom and the loss of the last water molecule(s) occur together as a concerted process. A comparison of this structure with that of dehydrated (at 350 "c)Ag6Na6-A treated with H? and with dehydrated (at 350 " c ) Ag4,6Na7,4-Atreated with H25indicates that all of the reduced silver atoms and 8-ring Ag' ions in those structures were Ag' ions in 6-rings before H2 was added, except for one silver atom per unit cell which may have formed during dehydration. Some 6-ring Ag+ ions moved to %ring sites to coordinate to silver atoms or clusters. Acknowledgment. This work was supported by the National Science Foundation (NSF-INT-83- 16234) and by the Korean Science and Engineering Foundation. Supplementary Material Available: Table of calculated and observed structure factors (2 pages). Ordering information is given on any current masthead page. (30) Kim, Y . ;Seff, K. J . Phys. Chem. 1978, 82, 1071-1077
Meaning and Structure of Amphiphilic Phases: Inferences from Video-Enhanced Microscopy and Cryotransmlssion Electron Microscopy D. D. Miller, J. R. Bellare, D. F. Evans,* Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis. Minnesota 55455
Y. Talmon, Department of Chemical Engineering, Technion-Israel Institute of Technology, Haifa 32000, Israel
and B. W. Ninham Department of Applied Mathematics, Research School of Physical Sciences, Institute of Advanced Studies, Australian National University, Canberra A . C .T., 2601 Australia (Received: June 9, 1986; In Final Form: September 8, 1986) This paper attempts to come to grips with a major issue confronting association colloid science. It does so by illustrating some surprising features of aggregates of simple amphiphiles as revealed by two powerful complementary tools, video-enhanced microscopy (VEM) and cryotransmission electron microscopy (cryo-TEM), both of which allow direct visualization. The natures of these aggregates challenge existing theories and show up limitations of some other noninvasive, though indirect, techniques. The problem of the meaning of amphiphilic phases and their microstructure is discussed and the necessity for a different descriptive language emphasized. 1. Introduction
I . I . Problem. No one schooled in classical thermodynamics has any doubts on the existence of phases. The world of equi-
0022-3654/87/2091-0674$01.50/0
librium homogeneous and heterogeneous substances is made up of phases-in the thermodynamic limit. It is a matter of definition. For example, the phase diagrams of (a) a nonionic single-chained surfactant (CI2EO6)and (b) a double-chained cationic (dido0 1987 American Chemical Society