Thermal Stability of the Stilbite-Type Framework: Crystal Structure of the Dehydrated Sodium/Ammonium Exchange Form

American Mineralogist, Volume 6E, pages 414419, 1983

Thermal stability of the stilbite-type framework: crystal structure of the dehydratedsodium/ammonium exchange form

Wrlrnreo J. Monrren Katholieke Universiteit Le uven Centrum voor Oppervlaktescheikundeen Colloildale Scheikunde de Croylaan 42 8-3030Leuven (Heverlee),Belgium

Abstract

O-T bridges between them occurs. The cations are located at the six-ring site and the flat eight-ring site, and residual water moleculesat the boat-shapedeight-ring. The framework Oistortionsare cation-induced and readily affect the cuboid polyhedron of the framework which is formed by joining adjacent structural units'

Introduction protons, but also cations such as K+ and Rb* with a sufficiently low ionic potential (Passaglia, 1980) may Stilbite-type minerals are zeolites with a two-dimen- "stabilize" the stilbite-type framework. The K and Rb sional interconnected channel system. Ten-membered forms do not undergo appreciable contraction and their ringsof (Si, Al) Oatetrahedra (free diametet4.1 x 6.24, destruction occurs only at fusion, i.e., at about 10fi)"C. Meier and Olson, 1978)Iimit the pore size of the largest As the dehydrated Ca-form exhibits a smaller fraction channels. These intersect with smaller ch-annelswith of broken T-O-T bonds (Alberti et al', 1978)than the eight-ring apertures(free diameter 2.7 x 5.7A, Meier and dehydratedNa-form (Alberti et al.,1978), the number of Olson, 1978). The inner surface should therefore be cations might be an important parameter. A further readily accessibleto small molecules and these minerals reduction of the number of cations is attempted here in could have a potential use as molecular sieves or cata- the study of a dehydrated Na-H-stilbite form. It will be lysts. This would require an "activation" at elevated shown that the framework-contraction and the frame- temperatures to remove the adsorbed water molecules' work-destruction steps can be separated. Unfortunately, in the presenceof exchangeablecations, and the framework is de- stilbite contracts on heating Methods and results stroyed (Passaglia,1980). Contraction and destruction are related to the attractive force of the extra-framework Crystals of natural stilbite (STI) from the Faroe Islands cations. When a crystal of an isostructural phase, Na- (Denmark; Virginia Polytechnic Institute and State Uni- barrerite is heatedto 250'C (Alberti and Vezzalini,1978), versity Samplenumber B179)were exchangedin a lN 1: I the framework is very much distorted and T-O-T bonds NaCl:NHa Cl solution for five months. The exchange are broken. A similar behavior, but with a much smaller solution was renewed three times during this period' fraction of broken T-G-T bonds, occurs in dehydrated Electron microprobe analyseswere made on crystals of stellerite, the Ca variety of stilbite (Alberti et al., 1978). the same batch. (The analysis was made at the Depart- Jacobs et al., (1979)characterized the hydrogen form ment of GeophysicalSciences, The University of Chicago as the most stable. This was confirmed by a structural under the following conditions: solid-state detector, study of the mineral after ammonium exchange and Reed-Ware data reduction, beam current:I00 nA, no dehydration (Pearceet al.,1980). Only very minor differ- corrections for water or Na loss, totalwt'Vo between 80.2 enceswere found between the framework parametersof and 82.3). Calibration factors for the different elements the hydrated form (Slaughter, 1970;Galli, 1971) and those were determined using Anzoglass for Ca, Si, Na and Al' of the dehydrated hydrogen form. By a reduction of the and asbestos microcline for K. Taken strictly at face cation-framework interactions, which become progres- value, when calculatedto 72 oxygens with sufficient NHn sively pronounced as the water ligands are removed, a addedto balancethe charges,the unit cell of Na-NHaSTI collapse of the framework can be prevented. Not only contains before dehydration: 0003-0M)083/0304-04l 4$02. 00 MORTIER: STILBITE FRAMEWORK: DEHYDRATED SODIUMIAMMONIUM FORM 415

Table l. Positional parameters,population factors and anisotropic thermal parametersfor d-Na-H-STI

Aton Equi- Popula- x y z 'l Bzz A Btz -13 -23 point tion I r(l) 8j 7.r3(3) o.99309(10) o. t9272(7) o.25505(t2' 327(tt) l40(4) 4s4(r5) lo(5) 272(to) t2(6) r (2) 8j 7. l e(3) 0,261l7(lo) o.30682(7) o,2s557(ll) 235(lo) r39(4) 430(15) l (s) 177(to) -l I (5) r (3) 8j8 0.2oooo(lo) o. o8843(6) 0.499s4( | 2) 465(f0) 98(3) 769(t5) -17(5) 389(lo) - 2(6) r (4) 8j8 o. l l36s(9) o.3l 659(6) o,5oo45(tl) 334(9> r l8(3) 609(13) -r2(s) 3o7(e) - 5(6) r (s) 494 o 0.2soo7(9) o 348(13) l9o(5) 497(t82' o 248(13) o o(r) 8j8 o.9804(3) o. t987(2) o. rors(3) loo(3) 27(t) roo(4) -r8(2) 80(3) -r4(2) o(2) 8j8 o. l 203(2) o,3o2r(2) o. lo5o(3) 42(2' 26(r) 95(4) - 4(r) l8(3) t2(2) o(3) 8j8 0. 06l 7(3) o.2668(2) o.35s7 (3) r I r (4) 27(r) 108(s) -r3(2) 73(4' -2t(2) o(4) 8j8 0.0693(3) 0.l r89(2) o.35oo(3) 83(3) 25(r) 100(4) r2(2) 4e(3) t5(2) 0(5) 8j8 o,2927(3) o.2525(2' o.3s3s(3) 80(3) 26(r) 106(s) 8(2) 36(3) 2r(2) 0(6) 8j8 0.2812(3) o.3790(2) 0.3s21(3) 88(3) 26(t) loo(4) - 3(2) 54(3) -r1(2) o(7) 8j8 o.3513(3) 0.3l 23(2) o. 2os4(3) 73(3) sl (2) l4o(s) -r2(2) 7o(4) -23(2) o(8) 8j8 0,3 182 (3) o.l l03(2) o.4996(3) 68(3) re(l) 143(5) - 6(r) 7l(3) - 2(2' o(9) 4i4 o.l87r(4) o o.so28(5) 83(5) rs(l) t44(7) o 76(5) o o(lo) 4h4 o o.3460(2) v2 s7(4' 25(t) t32(7' o 65(4) 0

Na(l) 8j l.4s(4) o.2323(r2) o.2780(7) -0.0339( 1 4) 6. I (4) ro.tr Na(2) 2d t.74(4' o tl2 tl2 518(28) 65(5) 7os(38) o 361(27' o (3) Na 4i. 2,78(4' o.3860(7 ) t/2 o.2702(8) 239(ll) 6s(3) 395(r7) o r98(ll) o

r(l)' 8j 0.62(3) 0.0960( I 2) 0.2o39(8) 0.2918(14) 2.O(4)'.tr'. r(2)' 8j o.s3(3) 0.1936(13) o.2955(8) 0.2954(r5) 1.2(4)t"'r'

ll estinated standard errors i.n parentheges refer to the la€t disit: :i:r anisotropic ther@l par@etere of T(l) to T(5) x lO5, otherslx iO4; t3r.:3 themal par&eters of Na(l), T(l)t and T(2)' isotro;ic : D/i2.

(NHa)o76Na2 a5K6.15Caa s6Ale asSi26.52072. xH2O density at each of the T(l) and T(2) sites was split into separatepeaks, T(l) and T(l)', andT(2)and T(2)' respec- This compares with the original sample composition tively, and their population factors were allowed to vary (Pearceet al.,1980) of independently.The T(l) + T(l)' and T(2) + T(2)' popula- Na15K6 1Caa.1Ale 6Si26.aO72 . xH2O tions sum to7.75 and7.72respectively, instead of 8. This deficiency is significant and might be an indication of partial A singlecrystal of NaNHaSTI0.15 x 0.3 x 0.3 mm3in dealumination, as it frequenfly occurs during heat treatment was size selectedand mounted in a glass capillary. After of NHa-exchangedzeolites (McDaniel and Maher, 1976). : evacuation of the capillary at room temperature (P The residual electron density- at the completion of the x 0.013 10-6bar), the temperaturewas raisedto 383 K refinementdid not exceed|elA'wilh the exceptionof pgaks over a 3 hour period and maintained for 18 hours, near the centersof Na(2) and 2elA) and Na(l) (1.6elA). A subsequentlyraised to 603K over a2Vzhour period and previous anisotropicrefinement of Na(l), with its position after another 2 Vzhour period at 603 K, sealed off and fixed at the center of the six-ring, had not converged and cooled to room temperaturefor data collection. yielded very elongatedthermal ellipsoids.While a position As derived from Weissenbergphotographs, the system- displacedfrom the centerconverged ifrefined isotropically, atic absenceswere consistentwith the spacegroup C2lm. the residual electron density at the center of the peak might graphite-monochromatized Using MoKa radiation, 5158 also be an indication of a mixed occupancy. This latter reflections were measuredon a Syntex P21single crystal could also be true for the eight-ring site Na(2). Residual : diffractometer in the omega scan mode up to sin0/)t water molecules,K3* or Ca2* ions or even Al3* extracted yielding 0.65, 2521independent reflections ofwhich 1793 from the frameworkcould also be locatedat thesesites (see were larger than 3o1after averagingand were considered discussion).Although the T(l) and T(2) sites are split, no as observed. The unit cell parameterswere obtained by indicationwas found for split oxygen sites(see also discus- least-squaresrefinement of 25 intense reflections having sion). 20 : - yieldin : : = 21" 33', g a 13.57l(4), b 18.264(2),6 Atomic scatteringfactors (International Tables X- : for 11.323(4)Aand B 126.96(2). Data reduction was per- ray Crystallography,vol. lil, 1968)for Na+, Ot-, Al2* formed using the X-ray 76 system (Stewart et al., 1976). and Si3+were used.Anomalous scattering corrections were The intensities were corected for Lorentz-polarization applied to all atoms (Cromer and Liberman, 1970). efects, but no absorption corrections were applied The positional parameters, population factors and the (pv'oxa = 5.9 cm-t). With the starting parametersfrom anisotropic thermal parametersare given in Table I and the (Pearce d-H-STI et al.,1980), the structure was refined interatomicdistances and bond anglesin Table 2.1 using the'program NUCLS,a modified version of onrus (Busing et al., 1962)to final agreement factors for the observedreflections of n : : >llr"l-l4llDlr,l 0.055and I To receive a copy of the observed and calculated structure : p(wllFj - : wR |F"lDn*nlzl|D 0.oq3.Electron densi- factors, order Document AM-E3-218from the Business Ofrce ty due to non-framework electrons was located by ditrer- Mineralogical Society of America, 80fi) Florida Avenue, N.W., ence Fourier methodsand assignedto Na+. The electron Washington,D.C. 20009.Please remit $1.00for the microfiche. 416 MORTIER: STILBITE FRAMEWORK: DEHYDRATED SODIUMIAMMONIUM FORM

Fig. l. Stereodrawings of the skeletalmodels showingthe 4-connectednets formed by the tetrahedralatoms for (a) d-Na-H-STI and (b) d-Na-STI.

Discussion and that they are characterizedby the complete absence of any framework contraction causedby a rotation of the The framework 4-4-l secondarybuilding units (Meier and Olson, 1978) As clearly shown by Passaglia(1980), contraction and as is evident for d-Na-STI in Figure lb. destruction of the stilbite-type framework are related to Polyhedral building units (PU, not secondary building the interaction with the exchangeablecations. The frame- unit sincethis requires the framework to be entirely made work is indeed stableas the hydrogen form(Peatce et al., up of one type only) can be defined as in Figure 2. These 1980)or in the presenceof cations with a sufficiently low PU's consist of two four-rings joined at two opposite ionic potential such as K* or Rb* (Passaglia,1980)' The corners (T(3) and T(4)) through two O-T(5FO bridges' structural parameters of d-H-STI (Pearce et al., 1980) Each of these T(5) atoms is then shared by two neigh- indicated that the framework was very similar to the bouring PU's to form "chains" (see Fig. I showing hydrated natural form, and no destruction of the frame- skeletal models consisting of 8 PU's each). These work could be detected. For the collapsed phases (Bar- "chains" are joined laterally by oxygen bridges at three rerite-B,i.e., d-Na-STI, Alberti andYezzalini,1978; and of the four T-atoms in the four-rings of the PU's to form a stellerite-B,i.e., d-Ca-STI, Alberti et al., 1978)contrac- "dense silicate layer". Theselayers are further stackedin tion of the framework was accompaniedby breaking of a parallel way and linked at the fourth T-atom of the PU T-O-T bonds. This suggeststhat introduction of consid- four-rings. erable stress into the framework causesdamage. It is most convenient to quantify the distortion of the An instructive way of illustrating the framework, is framework using vectors describing the relative orienta- with skeletal models which show only the tetrahedral tions and cross distances of the polyhedral units. Three atoms (T : Si, Al). A comparison of d-Na-H-STI vectors across the PU's will be considered(see Fig. (present work) with d-Na-STI (Alberti and Vezzalini, 2);Yt, a vector in the directionof the chains;V2, between 1978) can be seen in Figures la and 1b.. A visual two corners of opposite four-rings more or less perpen- inspection of d-Na-H-STI, and d-H-STI (figure not dicular to the densesilicate layer, i.e., betweenthe top shown) indicates that these structures are very similar and bottom nodesby which the layers are connected;and MORTIER: STILBITE FRAMEWORK: DEHYDRATED SODIUMIAMMONIUM FORM 417

from 2 x 3.004,2 x 3.054 and2 x 3.27Ato 4 x 2.$A and 2 x 2.634 for d-Na-STI and to 4 x 2.584 and 2 x 2.614 for d-Ca-STI respectively. In this way, favorable coordination distances for cations at site 3 can be obtained. Site 3 is the highest occupied site both for d-Na- STI and for d{a-STI. Becauseof the absenceof any zig- zag in the chains and because of the excessively long coordination distances at site 3, it is most probable that this site in d-Na-H-STI is occupied only by residual water molecules. Consideration of similar electrostatic interactions sug- Fig. 2. Schematicdrawing of the polyhedralunit indicatingthe gests that the site 2 cations may be responsible for a vectorsVl, V2and V3. rotation of the PU's. A shorteningof the O(10)-O(10) distances should cause a decreaseof the angle between the V2 vectors and of that between two V3 vectors. Site 2 V3, betweenthe two other cornersof the four-rings, but in is unoccupiedfor d-Na-STI and only 10%occupied in d- the dense silicate layer. The following angles between Ca-STI. A direct comparison of the anglesbetween two thesevectors can then be taken into account: vectors V2 and two vectors V3 for the collapsed d-Na- (l) the angle between V2, and V3 in the PU must STI and d-Ca-STI frameworks with the open d-H-STI expressthe deformation of the PU; (2) the angle between and d-Na-H-STI samples is however not possible be- two adjacent Vl vectors quantifies the zig-zagformation causeof the large diferen ce in zig-zagof the chainswhich of the "chains", and (3) the anglesbetween two V2 and also affects these angles. Despite the 5" decreasein zig- (4) two V3 vectors are related to the relative tilt of the zag from d-H-STI to d-Na-H-STI, and the concomitant PU's acrossand in the densesilicate layer respectively. increase in the angle between the two V2 vectors, the The lengthsofthese vectorsand the anglesbetween them slightly smaller angle between the two V3 vectors could are given in Table 3. support a partial occupancy of site 2 by cations. The order in which the zig-zagin the chains varies, as The secondmost important site in d-Ca-STI is the six- expressedby the angles between the V1 vectors of two ring site (site 1). This site is unoccupied in d-Na-STI and adjacentPU's is: d-Na-H-STI, d-H-STI, d-Ca-STI and d-H-STI. Because of the short site l-framework dis- d-Na-STL It is remarkable that the behavior of the tances, there is also no doubt that, for the present partially exchanged sample (d-Na-H-STI), is different structure, this site is occupied by small cationic species from that of the fully Ca and Na exchangedsamples, in such as Na+, Ca2+ or Al3+ extracted from the frame- that the zig-zagin the chains decreasesfor d-Na-H-STI, work. The dimensions of the six-ring are already small, but increasesfor d-Na-STI and d-Ca-STI with respect to and it may be anticipated that the cations at this site do d-H-STI. This could be an indication of the existenceof not significantly contribute to the framework distortions. severalcation-framework interaction centers,all affecting It is therefore obvious that the cations at site 3 are the framework in a distinct way. principally responsiblefor the contraction of the frame- In the present structure, three different extra-framework in d-Ca-STI and d-Na-STI. work sites were found. Only one of these has sufficiently short coordination distancesto the framework oxygen^s, Table 3. Stilbite-framework analysis f.e., site (1) with Na(l)-O(7) distancesof 2.24 and2.27A.

The closestapproach from the center ofthe flat eight-ring d-Na-tt-STIr. d-g-STI*x d-ca-sTlf d-lla-sttit to the frameworkoxygens (2 oxygensO(10) is 2.814, i.e., about 0.46A in excess of the sum of the efective ionic I. Length of the vectors (A) prewitt, radii of Na+ and 02- lShannon and 1969).The {r [r(s)-r(s)l e.o4s 9.Ot3 8.701 8.402 Na(3)-O(a) distancesof 3.00A and Na(3)-O(6) distances {2 t1(3)-r(3)l 6.049 6,O17 5.979 5.861 of 3.054 make it very unlikely that Na-ions are located at [3 [r(ll-r101 4'428 4,467 4.506 4.680 this site at all. 2, Anglea in the PU (") The interaction of these sites with the framework will (t, be different for each site as can be derived from the \ztY"s 69.6 71.10 positioning of the cations in the framework, shown in 3, Angles betweetr two adjacent PUr6 (') Figure 3. The electrostatic interaction of exchangeable (z)[r/[r I 80.O t74.9 t45,9 t45. I cations in site 3 with the framework oxygens may be (t) XztYz 154.2 l5l.8 | 34.9 I 38.5 responsiblefor the zig-zag of the "chains" found in d- (+)[r/[r I 13,4 I t4.4 I t4.8 124,7 Na-STI (Alberti and Vezzalini, 1978)and in d-Ca-STI :l tlta pearce + present atructure; et el., lg8o; st"l1e.it"-B, Alberti et (Alberti et al., 1978). This interaction occurs together rr aL,, 1978i Barrerite-B, Alberti and vezzalini, 1976. with a shortening of the cation coordination distances 418 MORTIER : STILBITE FRAMEW ORK : DEHYDRATED SODIUMIAM M ON IUM FORM

Fig. 3. Stereo drawing of the d-Na-H-STI framework and cation positions. Thermal ellipsoids shown at 50% probability.

The differencesin length of the vectors Y1, V2 and V3 lite EAB to a sodalite type structure was proposed by between the d-H-STI and d-Na-H-STI structures are Meier and Groner (1981).The defects at the T(1)' and less than 0.044. The values for these two structures are T(2)' positions could be an indication of a similar process. very diferent from those of d-Na-STI and d-Ca-STI. The mechanism of breaking the link between two The dimensions of Yr and V2 decrease,and those of V3 adjacent"chains" ofPU's and the decreaseofthe zig-zag increasein the order d-Na-H-STI. d-H-STI. d-Ca-STI in the chains, is clearly different from the observationsfor and d-Na-STI. This is the sequencein which the zig-zag d-Na-STI (Alberti and Vezzalini, 1978) and d-Ca-STI of the chains increases.This may also be related to the (Alberti et al., 1978). In all cases however, the cuboid broken bonds in the framework. Concomitant with the polyhedron in the framework formed by two four-rings of contraction of the framework, oxygen bridges betweenT- two adjacent PU's which are joined at three of the four atoms of the four-rings of the PU's were broken to an cornersby oxygenbridges, constitutes the weakestlink, extent of lUVofor stellerite-B (Alberti et al., 1978)and to and is invariably affected. Broken bonds were not ob- about 50Vofor barrerite-B (Alberti andYezzalini, 1978). served for the heulandite-type framework after collapse The destruction mechanism is obviously diferent in the (Alberti, 1973)although it containsthe samePU's. Chains case of the present structure. Instead of a contraction of formed by identical PU's are also joined laterally, but the framework with respect to d-H-STI and natural through a sequenceof one four-ring and two five-rings, stilbite, the reverse is observed. From the splitting of the instead of through a sequenceof two four-rings and one T(1) and T(2) tetrahedral sites, and their population six-ring. Apparently, this arrangementhas considerably factors, it can be estimatedthat about lUVoof the T-O(7) more flexibility. bonds linking two adjacent PU's are broken. However, The channel apertures have changed only slightly as no electron density deficiency at the O(7) oxygens could compared to d-H-STI. The dimensions of the eight-ring be detected. If the positions of the framework oxygens at site 2 are 8.52A tO(9)-O(9)land 5.62A tO(10)-O(10) are not affected, this would suggestthat at these sites, an vs. 8.664 and 5.49A for d-Na-H-STI and d-H-STI inverted tetrahedron is formed (see Fig. 3), but with one respectively. The O(9!-O(9) distance across the boat- oxygenmissing. The O-T-O anglesat the T(l)' and T(2)' shapedeight-ring at site 3 is 5.03A vs. 4.93A for d-H-STI, tetrahedra are: O(3)-T(l)'-O(3):108'; O(3)-T(1)'- and the O(9)-O(9) distances across the ten-membered O(4):106'; O(1trT(t)'-O(4):100';O(2)-T(2)'-O(5): 106'; ring 9.16A and9.27Avs. 9.07A and932A. 0(5)-T(2)'-0(6): 104' and O(2)-T(2)'-0(6):98". The breaking of a T-O bond is most likely to occur when T is Al. Cations The weakening of the Al-O bond will certainly be en- The localization of the cations has already been partly hanced by protonation of the Al-O-Si bond during de- discussedin associationwith the distortion of the frame- composition of NHa* to H" and NH3 in the dehydration work. It is highly improbable that cations were located at step. Since T(l) and T(2) will not be simultaneously site 3. Residualwater moleculeswill certainly be present occupiedby Al, only one side of the T-O-T bridge should in the structure, considering the relatively low dehydra- be broken. Al3* may then behavemerely as an exchange- tion temperatureof 603 K. About 2.8 H2O molecules (or able cation and can be located at an exchange site (the OH-) can therefore be located at site 3. A total residual populationdeficiency ofT(l) + T(1)'and T(2) + T(2)'is electrondensity (Na++K++Ca2* as shownby the chem- an indication of this), or at a three-coordinated site ical analysisand Al3* from the deficientT(l) + T(1)' and formed by three of the four oxygens of the original T(2) + T(2)' sites) of 3.36 Na+ equivalentsshould be tetrahedron. It should be emphasizedhere, that in this locatedat the other two exchangesites. At sitesI + 2, 3.2 case,neither the =Si-O(7FH bond nor the (=gl-6-;r4tr+ Na+ equivalents were located, certainly within the limit group carries any residual charge. An inversion-type of the chemical analysis. Both sites probably have a mechanismfor the topotactic transformation of the zeo- mixed occupancy. MORTI ER : STILBITE FRAMEW ORK: DEHYDRATED SODI UM IAMMONIUM FORM 4r9

Acknowledgments C. H. Macgillavryand G. D. Rieck, Eds., p. 201.The Kynoch Press, Birmingham, England. The author thanks the Belgian "Nationaal Fonds voor We- Jacobs,P. A., tenschappeliikOnderzoek" for a researchposition as "Bevoegd- Uytterhoeven,J. B., Beyer, H. K. and Kiss, A. (1979)Preparation and properties verklaard Navorser". Financial assistance from the Belgian of hydrogenform of stilbite, Government (Dienst voor programmatie van het Wetenschaps- heulanditeand clinoptilolite zeolites. Journal ofthe Chemical Society,Faraday beleid), the "Fonds voor Kollektief FundamenteelOnderzoek", TransactionsI, 75, 883-891. McDaniel, (1976) and the Research Fund of the K.U. Leuven are gratefully C. V. and Maher, P. K. Zeolite stability and ultrastablezeolites. In acknowledged.The author is also grateful for Dr. J. J. Pluth for J. A. Rabo, Ed., Zeolite Chemistry and Catalysis,A. S. Monograph p.2E5. the microprobe anlaysis, Prof. G. V. Gibbs for the gift of the C. l7l, Meier, W. M., andGroner M. (1981)Zeolite samples, and Prof. G. S. D. King for the use of the P21 structuretype EAB: crystal structure and mechanismfor the topotactic transforma- difiractometer and the X-Ray system of computer programsand for helpful discussions. tion of the Na, TMA form. Journal of Solid State Chemistry, 37,204-218. Meier, W. M., and OlsonD. H. (1978)Atlas of ZeoliteStructure References Types. Juris Druck * Verlag A. G., Ziirich, Switzerland. Alberti, A. (1973) The structure type of heulandite-B (heat Passaglia,E. (1980) The heat behaviour of cation-exchanged collapsed phase). Tschermaks Mineralogische und Petrogra- zeolites with the stilbite framework. Tschermaks Mineralo- phische Mitteilungen, 19, 173-184. gische und PetrographischeMitteilungen, 2T, 67J E. Alberti, A., Rinaldi, R., and Yezzzlini, G. (1978)Dynamics of Pearce,J. R., Mortier, W. J., King, G. S. D., Pluth,J. J., Steele, dehydration in stilbite-type structures; stellerite phase B. I. M. and Smith J. V., (1980)Srabilization of the stilbite-type Physicsand Chemistryof Minerals,2,365-375. framework: Crystal structure of the dehydrated NHa-ex- Alberti, A. andYezzalini, C. (1978)Crystal structures of heat- changedform. In L. V. C. Rees, Ed., Proceedingsof the collapsed phases of barrerite. In L. B. Sand and F. A. "Fifth International Conferenceon Zeolites" , p. 261. Naples, Mumpton, Eds., Natural Zeolites, p. E5-98. PergamonPress, Heyden,London. New Jersey. Shannon,R. D. and Prewitt, C. T. (1969)Effective ionic radii in Busing,W. R., Martin, K. O., and Levy, H. A.,(1962),ORFLS, oxides and fluorides. Acta Crystallographica,B25, 925-946. a Fortran crystallographic least-squaresrefi nement program. Slaughter, M. (1970) Crystal structure of stilbite. American U. S. National Technical Information Service, ORNL-TM- Mineralogist, 55, 387-397. 305. Stewart,J. M., Machin, P. A., Dickinson,G. W., Ammon, H. Cromer, D. T., and Liberman, D. (1970)Relativistic calculations L., Heck, H., and Flack, H. (1976)Eds., "The X-ray Sys- of anomalousscattering factors for X-rays. Journal of Chemi- tem", Technical report TR-446, The University of Maryland, cal Physics,53, 189l-1898. U.S.A. Galli, E. (1971)Refinement of the crystal structure of stilbite. Acta Crystallographica,827, 833-841. Manuscriptreceived, November 4, 1981; International Tables for X-ray Crystallography. Vol. III (1968), acceptedfor publication, November 2, 1982.