Download - 27Al/31P solid-state n.m.r. structural investigations of AlPO4-5 molecular sieve

ELSEVIER

*7Al/31P solid-state n.m.r. structural investigations of AlPO,-5 molecular sieve

C.A. Fyfe, ICC. Wang-Moon, and Y. Huang Department of Chemistry, University of British Columbia, Vancouver, BC, Canada

27Alf1P solid-state n.m.r. connectivity experiments were used to investigate the structure of the molecular sieve AlPO,-5 hydrated with D,O. Hydration with D,O resulted in an increased efficiency of coherence transfer from octahedral 27AI to 31P compared with hydration with H,O. The experiments prove directly that both the tetrahedral aluminum and the octahedral aluminum formed upon hydration are connected to phosphorus in the framework. The two-dimensional 27AI + 31P transferred-echo double-resonance and cross-polarization experiments show a downfield offset of the octahedral 27AI correlation from the tetrahedral 27AI correlation, consistent with the growth of a downfield shoulder on the 31P resonance as water is incorporated into the lattice.

Keywords: AIPO,-5; *‘AV3’P solid-state n.m.r. connectivity experiments; framework octahedral aluminum

INTRODUCTION

Molecular sieve systems are open framework structures containing channels of molecular dimensions. Zeolites are an important class of aluminosilicate materials that are used widely as molecular sieves because of the se- lectivity of their channel systems of organic molecules.‘,’

toward the adsorption When combined with their

acidic properties in the acid form, this gives them unique catalytic properties. During the 1980s there was an explosive growth in research on the related alumi- nophosphate (AlPO,) molecular sieves,” which have the general formula

[ WQJ ,W,) ,I . n&O (1)

Unlike most zeolites, the AlPO, molecular sieves are ordered, with an exact alternation of AlO, and PO, tetrahedra. The Al/P ratio, therefore, is always unity, and the system is neutral with no charge-balancing ex- traframework ions present. As a result, the AIPO,s can- not act as acid catalysts, although they have molecular sieving properties similar to those of zeolites.

AlPO,-5 is a large-pore molecular sieve with a unidi- meonsional pore system of diameter approximately 8 A and only one inequivalent tetrahedral site in its dehydrated framework.4 AlPO,-5 was the first novel framework of the AlPO,-based families whose structure was determined, although the more recentlv discovered zeolite SSZ-24 has the same topology.c” AlPO,-5 has a high water adsorption capacity (up to 28% of its dry weight), resulting in the conversion of up to 40% of the Al to octahedral coordination, as evi-

Address reprint requests to Dr. Fyfe at the Dept. of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC V6T IZI Canada. Received 27 March 1995; revised 14 July 1995; accepted 4 Au- gust 1995

Zeolites 16:50-55, 1996 0 Elsevier Science Inc. 1996 655 Avenue of the Americas, New York, NY 10010

dented by 27Al MAS n.m.r.6 This structural transforma- tion is readily reversible by dehydration over P,O,, which removes the octahedral 27Al resonance. In addition, the X-ray diffraction (XRD) pattern of hydrated AlPO,-5 shows that the framework structure has not collapsed. These observations suggest that the octahedral Al observed is not extralattice but is formed by hydration of a portion of the tetrahedral Al framework sites. This is in contrast to other systems, such as VPI-5,’ where hydration (and tetrahedral to octahedral conversion) occurs but at specific sites in the framework.

High resolution solid-state n.m.r. spectroscopy8 has emerged as a powerful complementary technique to XRD for structural investigations of molecular sieves. ‘,” Although XRD is sensitive to long range order and periodicity, solid-state n.m.r. can be used to determine short and medium range order (i.e., local environments). Recently, we have extended the po- tential of the solid-state n.m.r. to the mapping of 2”Si/2gSi connectivities in zeolites.‘” More recently, we have introduced double-resonance n.m.r. connectivity experiments involving 31p in VpI5 and ~Po4-8~~a~~~~~S~~~1~~~l~t~1~

and 11B/23Na/2gSi in borosilicate glasses.‘” Specifically, we have used cross-polarization (CP), transferred-echo double-resonance (TEDOR), rotational-echo double- resonance (REDOR), and dipolar-dephasing (DD) difference experiments to select connectivities in these systems via the heteronuclear dipolar interactions between the pairs of nuclei. In the present work we have applied these techniques to an investigation of AlPO,5 to try to obtain direct evidence that the framework re- mains intact during the hydration process.

EXPERIMENTAL

The sample of AlPO,5 was provided by the Union Car-

0144-2449/96/$15.00 SSDI 0144-2449(95)00099-R

bide Corporation. It was exchanged with D,O by dehy- drating at 70°C overnight, adding an excess of D,O, exchanging with D,O in a sealed vial for 5 h, removing the excess D,O by absorption on filter paper, and fi- nally packing and sealing the sample in a lo-mm rotor for n.m.r. analysis.

The n.m.r. experiments were performed using a Bruker MSL 400 spectrometer modified to include a third radiofrequency channel, with resonance frequen- cies of 104.264 and 161.977 MHz for “Al and 31P, respectively. The *‘Al and 31P chemical shifts are refer- enced to Al(NO,), and 85% H,PO,, respectively. A home-built double-tuned probe that incorporated a lo- mm supersonic spinner supplied by Doty Scientific Inc., with an internal volume of 0.8 ml, was used, and the 90” pulse lengths were 22.5 ps for both “P and *‘Al (cen- tral +1/2 t) --l/2 transition only). The samples were spun at 4.9-5.2 kHz.

The pulse sequences used for the DD, CP, TEDOR, and two-dimensional CP and TEDOR experiments have been described in detail previously.‘” For the one- and two-dimensional TEDOR experiments, the number of rotor cycles before and after the coherence transfer (n, and m, respectively) were optimized experimentally and are indicated in each figure legend.

RESULTS AND DISCUSSION

The “Al MYAS spectrum of AlPO,-5 (Figure I), acquired using a small pulse an

li le (less than 15”) to ensure a

quantitative spectrum, shows sharp resonances from tetrahedral Al at 38 ppm and octahedral Al at -16 ppm. Integration of the peak areas shows that 38% of the Al sites have octahedral coordination, whereas 62% have tetrahedral coordination. This is in agreement with a previous study of hydrated AlPO,-5, which found that up to 40% of the Al sites could be in an octahedral Al

I I I I I I I I I I I I I 150 loo 50 0 -50 -100 -150

Frequency (ppm from Al(NO&)

Figure 1 27AI MASspectrum of AIPO,-5, acquired using a pulse angle of 11.25”. Spinning side-bands are indicated by asterisks (“I.

A/PO,& z7A//3’P so/id-state n.m.r. connectivity: CA. Fyfe et al.

coordination.” The tetrahedral Al peak arises from thus single Al[4P] environment in the framework, wherea, the octahedral Al peak has been assigned previously” tc) an environment arising from the coordination of two water molecules to a tetrahedral Al site to give an octahedral coordination. Thus, the presence of the octahedral Al peak indicates that the sample is well hy - drated. In addition, there is a very small sharp peak at 8 ppm (which is not a spinning side band or due to the offset frequency) between the tetrahedral and octahedral Al resonances. This peak can be assigned to five,- coordinate Al arising from the coordination of one water molecule to a tetrahedral Al site in the framework. This environment is due to incomplete hydration and may be an intermediary between tetrahedral and octet- hedral Al.

The 31P MAS spectrum (F&uw 2) shows a single broad asymmetric resonance centered at -26 ppm, with a small downfield shoulder in the region from -10 to -20 ppm, similar to that presented previously6 for fully hydrated AlPO,-5. Careful inspection of the alP &US spectra of the hydrated and dehydrated samples presented in Ref 6 indicates that in these also, upon hv- dration, the 31P resonance shifts downfield, broadens, and displays a small downfield shoulder.

The conventional *‘Al spin-echo (SE) experimerit (Figure 3~) shows resonances from tetrahedral and octahedral Al and also a very small signal due to five- coordinate Al. The Tz relaxation times were determined to be 3.0 and 1.7 ms for tetrahedral and octahedral Al, respectively, and are therefore not short enough IO cause problems in SE type experiments over small IO moderate numbers of rotor cycles. In the *‘Al(“‘P) DD difference experiment, an “Al SE is recorded first with on-resonance 31P dephasing and then without “‘P dephasing (i.e., off-resonance). The difference between these two signals represents the *‘Al sites that are

Frequency (ppm from H3P04)

Figure 2 3’P MAS spectrum of AIPO,-5, with spinning side bands indicated by asterisks (*).

Zeolites 16:50-55, 1996 51

A/PO,-5, 27AW3’P solid-state n.m.r. connectivity: C.A. Fyfe et al.

The CP experiment is shown in &we 4a, using an optimized contact time of 2 ms, and the TEDOR experiment in Figure 4b, with the optimized conditions of n = 4 and m = 1 rotor cycles. The CP and TEDOR experiments both show as expected that the “P site, including the downfield shoulder, is connected to “Al. Note that in the TEDOR experiment, the spinning side bands (in dicated by *) have a nega$ve phase. This effect, observed previously in VPI-5, arises because the evolution of the side bands is different from that of the isotropic peaks. Using the optimized conditions for the one-dimensional experiments, the TEDOR and CP experiments can be extended to two dimensions to map the individual connectivities.

A sample of AlPO$ hydrated with H,O was first used for the two-dimensional 27Al + 31P TEDOR experiment. However, the correlation to octahedral Al was much weaker than expected and could not be ac- counted for by the effects of T2 relaxation alone. We feel that the presence of a third nucleus, namely ‘H in . the H,O molecules coordinated to octahedral Al, inter-

27A1 ( 3 ‘P) DD null feres with the reversal of the sign of the dipolar cou- pling in the TEDOR experiment, thereby reducing the efficiency of coherence transfer from octahedral Al to P. Therefore, the AlPO,-5 was exchanged with D,O

I I I I I I1 I I I I I I I

150 loo 50 0 -50 -100 -150 Frequency (ppm from Al(N03)3)

Figure 3 27AI experiments on AIPO,-5. In each experiment, the spinning rate was 4.88 kliz, and the total echo time was two rotor cycles (n = 2). (a) *‘AI SE signal (no 3’P dephasing), with spinning side bands indicated by asterisks (*). (b) *‘Ah3’P) DD difference experiment, with one rotor cycle of 3’P dephasing. 10,528 scans were acquired with a recycle delay of 0.25 s. (c) 27Al(3’P) DD difference null experiment acquired under the same conditions as (b), except that the 3’P dephasing was kept on- resonance.

dipolar-coupled (i.e., close in space or “connected”) to the 31P site. The ‘7Al(31P) DD experiment on AIPO,-5 (Figuw 3b) demonstrates that both tetrahedral and oc- t.ahedral Al are connected to P, and thus both are in the framework. This confirms the previous suggestion (Ref. 6) that octahedral Al is formed by coordination of two water molecules to an Al site. Since only one inequivalent tetrahedral Al site is observed, the octahedral Al must be formed at random sites in the framework. The corresponding 2’Al(31P) DD null experiment was performed (fipre 36) in which the 31P dephasing was kept on-resonance in all experiments. The null experiment confirmed that no “Al signals arose from imperfect sub traction in the difference experiment.

The one-dimensional 27Al 3 31P TEDOR and CP experiments involve coherence transfer from *‘Al to :‘iP and thus show only si nals from those 31P sites that are dipolar-coupled to ‘41 In addition, since 27Al is . the source of the magnetization, the experiments can be repeated very quickly (0.1 s), thereby increasing the efficiency of the experiments over a given time period.

h * * CP (a>

50 0 -50 -100 Frequency (ppm from H3P04)

Figure 4 One-dimensional “AI + 3’P coherence-transfer experiments on AIPO,-5, with spinning side bands indicated by asterisks (*). (a) “AI + “P CP experiment, with a contact time of 2 ms. 2,400 scans were acquired with a recycle delay of 0.1 s. (b) “AI + 3’P TEDOR experiment, with n = 4 and m = 1 rotor cycle. 2,048 scans were acquired with a recycle delay of 0.1 s.

52 Zeolites 16:50-55, 1996

A/PO& *7A//3’P so/id-state n.m.r. connectivity: C.A. fyfe et al.

(see the Experimental section) to reduce the interfer- ence from ‘H in H,O, and the DsO-exchanged sample was used for all of the TEDOR and CP experiments presented in this work.

The two-dimensional “Al + “‘P TEDOR experiment (F&X 5), acquired with n = 4 and m = 1 rotor cycles, shows clear correlations of both tetrahedral and octahedral Al to the P site, again confirming that octahedral Al is connected to P in the framework. The dashed Zirm show that the correlation of P to the tetrahedral Al resonance is centered at -26 ppm, whereas the correlation to octahedral Al is centered downfield at -22 ppm. The downfield offset of the octahedral Al cross- peak is consistent with growth of a downfield shoulder on the sip resonance in the region from -10 to -20 ppm as water is incorporated. T2 effects underestimate the octahedral Al intensity ( T2 = 1.7 ms for octahedral Al versus 3.0 ms for tetrahedral Al), giving a correlation to octahedral Al which is less intense than in the “Al SE experiment (Fiprr 3a), which itself is subject to T2 effects.

The twodimensional “Al 4 slP CP experiment (Fig- ure 6)) acquired with a contact time of 2 ms, also shows a clear correlation of P to both tetrahedral and octahedral Al. However, the correlation to octahedral Al is much stronger in the two-dimensional CP experiment than in the two-dimensional TEDOR experiment (Fig- ure 5). In the “Al projection of the two-dimensional CP experiment, the intensity of the octahedral Al is approximately one third that of the tetrahedral Al. Again, the dashed lines show that the correlation of P to tetrahedral Al is centered at -26 ppm, whereas the correlation to octahedral Al is centered downfield at -22 ppm. Interestingly, there is also a correlation of the previ-

I , , , I I I I I I I I1I I I I , I I ,

25 0 -25 -50 -15

Frequency (ppm from H $‘04)

Figure 5 Two-dimensional “AI + 3’P TEDOR experiment on AlPO,-5, with n = 4 and m = 1 rotor cycle. The dashed lines indicate the connected resonances. For each of the 128 experiments in t,, 4,096 scans were acquired. The recycle delay of 0.1 s resulted in a total experimental time of 14.6 h.

1111111~1~~1~111111II~1IIII~

25 0 -25 -50 -15 -100

Frequency (ppm from H sP0, )

Figure 6 Two-dimensional “Al + 3’P CP experiment on AIPO,- 5, with a contact time of 2 ms. The dashed lines indicate the connected resonances, and the spinning side bands are ind- cated by asterisks (*). For each of the 128 experiments in t,, 5,600 scans were acquired. The recycle delay of 0.1 s resulted in a total experimental time of 19.9 h.

ously assigned five-coordinate Al resonance to P, whic,h supports the assignment of this environment as arisiiig from the coordination of one water molecule to an %l site in the framework. This correlation to fi\e- coordinate Al is also present in the two-dimensional TEDOR experiment (Kpre 5)) although its intensity is much smaller than in the two-dimensional CP experiment.

The two-dimensional CP and TEDOR experimerits show correlations from tetrahedral Al to the large “P peak (at -26 ppm) and from octahedral Al to the dou,n- field shoulder (in the region from -10 to -20 ppni). This observation might suggest that there are only two P environments: one surrounded by tetrahedral Al, aitd one surrounded by octahedral Al. However, in this case the resonances should be sharper than observed. .\n alternative explanation can be formulated in terms of a random distribution. If the individual probabilities of finding tetrahedral and octahedral Al are approximately 62 and 38%, respectively, and the distribution is assumed to be random, there are actually five possible P environments surrounded by varying numbers of ret- rahedral and octahedral Al as first nearest neighbors. as shown in Table 1. The broad “‘P resonance, encomp;iss- ing the large peak (-26 ppm) and the small shoulcler (from -10 to -20 ppm), represents the distribution of the five possible P environments.

Fipre 7 shows “‘P rows of the two-dimensional “Al + s P CP experiment (F&re 6) taken through thetet- rahedral and octahedral “Al resonances. Figure 7u she ws that the correlation of P to octahedral Al is centered at -22 ppm, with a line width at half-height of 14 ppm. This distribution represents the four P environ-

Zeolites 16:50-55, 1996 53

A/PO,-5, 27Al/3’P solid-state n.m.r. connectivity: C.A. Fyfe et al.

Table 1 Probabilities of the five possible phosphorus environments in AIPO,-5

Phosphorus environment= Probability (%)*

P[4 tetl 14.8 P[3 tet, 1 octl 36.2 P[2 tet, 2 octl 33.3 P[l tet, 3 octl 13.6 P[4 octl 2.1

a tet and act refer to tetrahedral and octahedral Al, respectively. *The probabilities are calculated assuming that the individual probabilities of tetrahedral and octahedral Al are 62 and 38%, respectively, and that the distribution is random.

ments that contain octahedral Al, that is, from P[4 act] to P[3 tet, 1 act]. The center of the distribution (-22 ppm) is likely due to P[2 tet, 2 act] and not P[3 tet, 1 act], even though the former has a lower probability than the latter. The reason is that there should be a relative enhancement of each of the four P environments which is approximately proportional to the number of attached octahedral Al atoms, favoring those P environments with high Al content. Therefore, the intensity of the P[2 tet, 2 act] site is enhanced relative to the P[3 tet, 1 act] site. A similar behavior was observed in dipolar based 27Al + 2gSi connectivity experiments on zeolites,14 in which there was a relative enhance-

ment of a given resonance which was approximately linear with the number of Al atoms in the neighboring tetrahedral sites. Figure 7b shows that the correlation to tetrahedral Al is centered at -26 ppm, with a linewidth at half-height of 12 ppm. This distribution represents the four P environments that contain tetrahedral Al, that is, from P[l tet, 3 act] to P[4 tet]. Using the same arguments as for octahedral Al, the center of the distribution (-26 ppm) is likely due to P[3 tet, 1 act].

As a final check that the octahedral Al is in the framework, a one-dimensional 31P + *‘Al CP experiment was performed. The experiment (not shown) revealed signals from both tetrahedral and octahedral Al, further confirming that both Al environments are connected to P in the framework, but was much less efficient than the “Al -+ 31P CP experiment (Figure 4a), since the much longer 31P Tl relaxation time now determines the recycle delay of the experiment.

CONCLUSIONS

27Al/31P solid-state n.m.r. experiments prove directly that in hydrated AlPO,-5, both tetrahedral and octahedral aluminum are connected to phosphorus in the framework. Furthermore, the two-dimensional 27Al -+ slP TEDOR and CP ex set of the octahedral ’ 7p

eriments show a downfield off- Al correlation from the tetrahe-

dral “Al correlation, consistent with the growth of the downfield shoulder on the 31P resonance as water is incorporated. This observation can be interpreted in terms of a random distribution of tetrahedral and octahedral aluminum coordinations around the phospho- ros sites. The two-dimensional 27Al + 31P CP experiment also shows that the small peak between the tetrahedral and octahedral “Al resonances is connected to “P, confirming the assignment of this site to five- coordinate aluminum arising from the coordination of one water molecule to an aluminum site in the framework.

We observed an increased efficiency of coherence transfer from octahedral “Al to “P upon exchanging the sample with D,O. This suggests that, in general, an abundant third nucleus (in this case, ‘H in H,O) may interfere with such coherence transfer experiments.

o i;i \ii- ACKNOWLEDGMENTS

This work was supported by the Natural Sciences and Engineering Research Council of Canada in the form of operating and equipment grants (to C. A. F.), a post- graduate scholarship (to K. C. W.-M.), and a postdoc-

J 11 11 1 11 I I I 11 I 11 11 I L

toral fellowship (to Y. H.). We thank the Union Carbide Corporation for kindly providing the sample of AlPOh-

50 0 -50 -100 and T. Markus for assistance with probe electronics.

Frequency (ppm from H3P04) REFERENCES

Figure 7 3’P rows through octahedral Al (a) and tetrahedral Al (b), from the two-dimensional 27AI + 3’P CP experiment on AIPO,-5 (Figure 6). The spinning side bands are indicated by asterisks (*).

1 Meier, W.M. in SC/ Monograph on Molecular Sieves (Ed. R.M. Barrer) Sot. Chem. Ind., London, 1978

2 Breck, D.W. Zeolire Molecular Sieves: Structure, Chemistry and Use, Wiley, New York, 1974

54 Zeolites 16:50-55. 1996

AIPO,-5, 27A1/3’P so/id-state n.m.r. connectivity: C.A. Fyfe et al.

Wilson, ST, Lok, B.M. and Flanigen, E.M. US Pat. 4,310,440 (1982)

10

Bennett, J.M., Dytrych, W.J., Pluth, J.J. and Smith, J.V. Ze- elites 1987, 7, 160

11

Bialek, R., Meier, W.M., Davis, M. and Annen, M.J. Zeolites 1991,11,438

12

Meinhold, R.H. and Tapp, N.J. J. Chem. Sot. Chem. Com- 13 mun. 1990,219 McCusker, L.B., Baerlocher, C., Jahn, E. and Btilow, M. Ze- elites 1991, 11,308

14

Fyfe, CA. Solid-state NMR for Chemists, CFC Press, Guelph, 1984

15

Engelhardt, G. and Michel, D. High-resolution So/id-state NMR of Silicates and Zeolites, Wiley, Chichester, 1987

16

Fyfe, C.A., Feng, Y, Grondey, H., Kokotailo, G.T. and Gies, H. Chem. Rev. 1991,91,1525 Fyfe, C.A., Grondey, H., Mueller, K.T., Wong-Moon, K.C. and Markus, T. J. Am. Chem. Sot. 1992, 114, 5876 Fvfe, C.A., Mueller, K.T., Grondey, H. and Wong-Moon, K.C Chem. Phys. Left. 1992, 199, 198 Fvfe. C.A.. Mueller. K.T.. Grondev. H. and Wona-Moon. K.C J: Phys. ihem. 1963,9i, 13484 ‘. Fyfe, C.A., Wong-Moon, K.C., Huang, Y., Grondey, H. and Mueller, K.T. J. Phys. Chem. 1995, 99, 8707 Fyfe, CA., Huang, Y., Wong-Moon, K.C. and Grondey, H. unpublished work Fenzke, D., Freude, 0.. FrGhlich, T. and Haase, J. Chem Phys. Lett. 1984, 111, 171

Zeolites 16:50-55, 1996 55

Download - 27Al/31P solid-state n.m.r. structural investigations of AlPO4-5 molecular sieve

Top Related