1
Characterization of adsorbed xenon in Zeolite-A, X, and Y by high-pressure 129Xe NMR spectroscopy

2
The local structure of confined xenon in A-, X- and Y-type zeolites (molecular sieve 5A, 13X and zeolite NaY) was investigated by in situ high-pressure 129Xe NMR spectroscopy.

3
Xenon confined in dehydrated samples gives a resonance peak in the chemical shift range from 50 to 100 ppm at 0.1 MPa, and the chemical shift values increase as pressure increases.

4
The pressure dependence of 129Xe chemical shift on xenon confined in the micropores can be approximately described by the Langmuir adsorption model, resulting in the parameters being able to describe pore diameter, adsorption ability of the pore, and the local structure of the confined xenon atoms.

5
Using these parameters, the pore structure, the mobility and the cluster size of xenon in the micropore are discussed for A- and X-type zeolites.

6
Furthermore, the effect of pre-adsorbed water and the binder for pellet preparation on the pore structure is also examined.

7
In the Y-zeolite, a chemical exchange of xenon between the outside and inside of the micropores was observed above 0.75 MPa.

8
Spectral simulation assuming a simple two-site exchange model revealed the exchange rate constant and the population ratio of xenon in the exchange between the outside and the inside the micropores.

9
The exchange of xenon is optimum around the supercritical point of the bulk xenon (∼6 MPa), and decreases above 6 MPa, and this originates from the cooperative phenomenon in the supercritical fluid.

Introduction

10
Zeolites are crystalline porous aluminosilicates including alkali and/or alkaline earth metals, with the chemical composition of Mx/n·[(AlO2)x(SiO2)ywH2O. These compounds exhibit interesting properties such as molecular sieve ability and catalysis.1

11
These properties can be modified by chemical treatment such as ion exchange and/or de-alumination.

12
Many researchers have investigated these materials at the scientific and engineering points of view,2–4 and it has been found that the adsorption behavior, reactivity, and molecular sieve ability of zeolites are closely related to the framework, porous structure, and intermolecular interactions between the pore wall and the adsorbed molecule, as well as between the adsorbed molecules.

13
In addition, adsorbed water present in the micropores and the formation of macro- and mesopores due to the aggregation of fine particles of zeolites, will also affect the adsorption behavior of the zeolites.

14
The adsorbed water decreases the volume of the micropore to less than 0.5 nm, which is a so-called “ultra-micropore”, and screens the guest molecules from the pore wall, affecting the functional properties of zeolites such as catalysis, the molecular sieve and reactivity.

15
Furthermore, the adsorbent products consist of fine particles of synthetic zeolite with a μm diameter order, which are bound by binder agents such as clay, phosphates, and glass with a low melting point to prepare the pellets and/or beads for easy handling.

16
The addition of binder agents to form pellets leads to extra pores of 1–50 nm.

17
These meso- and macropores increases in the amount of adsorbate at the large value of the relative pressure and play an important role in the process of the diffusion of the adsorbates such as molecular diffusion and Knudsen diffusion.65

18
129Xe NMR is an important tool in the study of ultra-micro, micro-, meso-, and macropores.

19
Xenon-129 NMR spectroscopy has been developed to study the structure and topology of internal voids, adsorption sites, and intermolecular interactions in porous materials, since the potential of xenon-129 as a probe for the inner space of materials was proposed by Ito and Frassard5 in 1980.

20
Recently, this smart methodology has been widely used in the characterization of porous materials as a modern and sophisticated technique of porosimetry,5–9 and the local structure of materials such as zeolites,5,10–15 glass,16 fullerenes,17 carbon nanotubes,18 polymers,19–22 and clathrate compounds,23–26 liquid crystals,27–29 some organic solvents,9,30 and proteins in solution30–32 have been investigated.

21
In the past decade, hyperpolarized (HP) xenon NMR techniques have been developed to enhance the sensitivity of 129Xe NMR.33,34

22
However, to obtain information from ultramicro- and/or meso- and macropores using xenon as a probe, it is necessary to confine xenon in ultra-micropores of the same size as xenon, and/or to slow down the exchange between xenon adsorbed in meso- and macropores and the bulk.

23
It is difficult to achieve this situation under ambient conditions, and high-pressure conditions offer a solution.

24
We recently described a conventional and newly designed in situ variable pressure NMR probe, allowing studies with pressures up to 20 MPa,22 and it has been applied in the study of xenon–wall interactions in the free volume of polymers,35 the intermolecular interaction of xenon confined in 1D-nanochannels,36,37 and the supercritical phenomenon of xenon in mesoporous silica.38

25
Furthermore, Brunner et al. have reported the pressure dependence of 129Xe chemical shift in the supercritical state up to a pressure of about 70 MPa using a high-pressure cell made from a sapphire tube.39

26
This high pressure condition is expected to offer new insights into xenon–wall and xenon–xenon interactions for xenon confined in nanospaces, as well as the dynamic behavior of xenon in nanospaces.

27
In this study, we applied the high-pressure 129Xe NMR technique to some typical zeolites: molecular sieve 5A, 13X, and zeolite NaY.

28
Molecular sieve 5A has a type-A framework (LTA) with the chemical composition of (Ca4.8, Na2.4)[(AlO2)12(SiO2)12] ·wH2O, with the unit cell having a lattice symmetry of Pm3m, a lattice constant of a0 = 1.24 ± 0.001 nm, and a maximum diameter of 1.14 nm for the super cage which is linked each other by the window with a diameter of about 0.5 nm.40–42

29
In contrast, molecular sieve 13X and zeolite NaY have a faujasite X- and Y- type framework (FAU) with the chemical composition of Nax[(AlO2)x(SiO2)192-xwH2O, respectively, with the unit cell having a lattice symmetry of Fd3m, a lattice constant of a0 = 2.49 ± 0.01 nm and a maximum diameter of 1.37 nm for the super cage which is linked each other by the window with a diameter of 0.74 nm.42,43

30
We have investigated the adsorption behavior of xenon on these zeolites using high-pressure 129Xe NMR spectroscopy to clarify the influence of the framework, the adsorbed water and the binder agents on the adsorption behavior of xenon under high-pressure conditions.

Experimental

Chemicals

31
Powdered samples of molecular sieve 5A, 13X, and zeolites NaY were purchased from Aldrich Co. Ltd. Pellets of molecular sieves were purchased from Nakarai Co. Ltd. Sample dehydration was achieved by heating at 400 °C under reduced pressure (P < 2 × 10−2 Torr) for 3–24 h after roughly drying the specimens at ambient temperature under reduced pressure for 12 h.

32
All the specimens were characterized by thermogravimetric analysis (TG), nitrogen adsorption isotherm, 29Si MAS NMR spectrum, and powder X-ray diffraction (XRD) to confirm the amount of adsorbed water, the pore volume and the specific surfaces, and the ratio of SiO2 to AlO2 (Si/Al) in the framework.

Thermogravimetric analysis (TG)

33
TG analysis was carried out using SEIKO I SSC-5020M II TG/DTA 200-type thermal analysis equipment.

34
TG diagrams were recorded at a heating rate of 10 K min−1 up to 600 °C under nitrogen gas flow at a rate of 50 ml min−1.

35
α-Alumina, α-Al2O3, was used as a reference sample.

N2 adsorption isotherm measurements

Nitrogen adsorption isotherms were measured using Shimazu Gemini 2375 automatic specific surface measuring apparatus in the relative pressure range P/P0 from 0.001 to 0.95.

29Si MAS NMR measurements

29Si magic-angle spinning (MAS) NMR spectra were measured using a Bruker DSX-200 spectrometer with a Larmor frequency of 39.7 MHz.

36
Free-induction decay (FID) signals were recorded in a single pulse experiment using a pulse delay of 80 s, and π/2-pulse of 8 μs, under 4 kHz of MAS rate at room temperature.

37
The chemical shift was referenced by tetramethylsilane (TMS) as an external standard.

Powder X-ray diffraction (XRD) measurements

38
XRD measurements were carried out using a Rigaku Rotaflex Rint 2000 diffractometer on HP 9000 S712/60 system from 3° to 60° in 2θ.

39
The scan rate was 1.00° min−1, and the increment step was 0.01°.

High-pressure 129Xe NMR measurements

40
High-pressure 129Xe NMR measurements were carried out using a Bruker MSL-200 spectrometer with a Larmor frequency of 55.6 MHz, in which a home-built in situ pressure variable NMR probe22 was installed.

41
The powdered or pellet samples were packed into a glass tube of 5 mmϕ, with both ends capped with glass wool, and the tube and samples were placed in a pressure-resistant NMR cell made from ZrO2.

42
The prepared NMR cell was then mounted onto the high-pressure NMR probe.

43
The FID signals were recorded in a single-pulse experiment with a pulse delay of 5 s, π/2-pulse of 3 μs and accumulation of 8-8192, depending on the signal-to-noise ratio, over a pressure range of 0.01 to 10 MPa at 298 K. 129Xe chemical shift was referenced by a signal from xenon gas at zero density.

44
Temperature and pressure were controlled within the experimental error of ±0.5 K and 10% in MPa, respectively.

Results

Characterization of samples using TG, N2 isotherm, and 29Si MAS NMR

45
Thermogravimetric (TG) measurement of pre-heated samples up to 600 °C suggests that water adsorbed in the samples was removed by heating above 400 °C.

46
The amount of adsorbed water, defined by the ratio of the weight of water to the weight of the dehydrated zeolites, was determined from the weight loss as follows; 41% for molecular sieve 5A, 100% for molecular sieve 13X and 163% for Zeolite NaY.

47
After dehydration of the samples by heating at 400 °C under reduced pressure, the N2 adsorption isotherm was measured to characterize the amount of nitrogen at the maximum adsorption (Wsat), the specific surface area (Stotal), the pore volume (Vpore), the external surface area (Sex), and the internal specific surface area (Sin).

48
These samples give the typical I-type adsorption isotherm according to the IUPAC classification.44,45

49
The pore volume, external and internal surface areas are given by t-plot analysis.46

50
The parameters determined are listed in Table 1.

51
29Si magic angle spinning (MAS) NMR spectrum gives the atomic ratio of Si to Al, Si/Al for each sample as listed in Table .147–50

According to the results of TG and 29Si MAS NMR, we can determine the chemical composition of each sample: (Ca4.8, Na2.4)[(AlO2)12(SiO2)12]·(∼39)H2O for molecular sieve 5A, Na84[(AlO2)84(SiO2)108]·(∼750)H2O for molecular sieve 13X, and Na52[(AlO2)52(SiO2)140]·(∼1200)H2O for Zeolite NaY.

129Xe NMR spectrum for dehydrated powder samples

Figs. 1(a)–(c) show the pressure dependence of the 129Xe NMR spectrum in dehydrated powder samples of molecular sieve 5A, 13X and zeolite NaY, respectively.

52
The pressure dependence of the chemical shift values for each resonance line is shown in Fig. 2.

53
In each sample, the spectrum at 0.1 MPa mainly consists of two dominant peaks; one near 0 ppm and other at the chemical shift range from 50 to 100 ppm.

54
The resonance line near 0 ppm comes from bulk free xenon gas coexisting on the powder sample.

55
The chemical shift of this resonance line agreed with that for pure bulk xenon gas within the experimental error, suggesting that the observed peak for bulk free xenon is not affected by the chemical exchange with xenon adsorbed in the micropores.

56
On the other hand, the peak appearing in the lower field corresponds to xenon adsorbed in the micropore, and shifts toward the lower field side as pressure increases.

57
The δ-value, which is extrapolated to zero pressure using the data below 0.1 MPa, gives 90 ppm for molecular sieve 5A, 65 ppm for 13X and 52 ppm for zeolite NaY, and increases rapidly as pressure increases to 1 MPa.

58
For molecular sieve 5A and 13X, it reaches 260 ppm and 230 ppm at 6 MPa, respectively.

59
In addition to these two peaks, the extra peak appears at the chemical shift range 20–30 ppm higher than that of bulk-free xenon above 5 MPa.

60
This peak originates from partial components of xenon contributing to the chemical exchange between the inside and outside of the micropore.11,51

61
In contrast, in zeolite NaY, the resonance line shifts again toward the higher field above 1 MPa with the line broadening, and the δ-value for xenon in the micropore shows the minimum level of 110 ppm at 6 MPa.

62
Above 6 MPa, the resonance line shifts again to the low field with narrowing of the lines.

63
Furthermore, under 4 MPa pressure, the resonance line at 306 K (FWHM ∼70 ppm) becomes narrower than that at 298 K (FWHM ∼100 ppm) as shown in Fig. 3.

This anomalous pressure dependence is caused by the influence of the rapid chemical exchange of xenon between the inside and outside of the micropores.

129Xe NMR spectrum for hydrated powder samples

Figs. 4(a) and 4(b) show the pressure dependence on the 129Xe NMR spectrum for the powder sample of molecular sieve 5A and 13X before dehydration.

64
The chemical shift for xenon in the micropore entirely shifts the lower field compared to the dehydrated samples.

65
Fig. 5 shows the pressure dependence of the chemical shift value for each sample, in which the δ-value extrapolating to zero pressure is evaluated as 127 ppm for molecular sieve 5A, and 160 ppm for 13X.

66
These values are 37 ppm and 95 ppm greater than for each dehydrated sample.

67
The chemical shift value increases 100 ppm for 5A and 35 ppm for 13X as pressure increases up to 6 MPa.

In these samples, the extra broad peak, which appears at 20–30 ppm higher than the peak of free xenon in the dehydrated samples above 5 MPa (See Fig. 1(a) and 1(b)), was not observed.

129Xe NMR spectrum for pellet samples

Figs. 6(a)–(d) show the pressure dependence of 129Xe NMR spectra of xenon adsorbed in the pellet samples of molecular sieve 5A and 13X.

68
In molecular sieve 5A, the pressure dependence of 129Xe NMR spectra is similar in both the hydrated (Fig. 6(a)) and dehydrated (Fig. 6(b)) pellet samples.

69
The extra resonance peak, which appears above 4 MPa in the dehydrated powder sample in Fig. 1(a), was also observed in the pellet dehydrated sample.

70
It is considered that the existence of some additional porous structure such as micro-, meso- and macropores, which is caused by binder agents, is not clear in molecular sieve 5A.

71
On the other hand, in molecular sieve 13X, the resonance peak from xenon in the micropore is split into two peaks in the pellet sample.

72
It is considered that there are two types of micropores in the pellet samples.

73
The intensity of these peaks depends on the presence of adsorbed water.

74
In the partially hydrated sample (Fig. 6(c)), the peak in the higher field is dominant, whereas in the dehydrated sample (Fig. 6(d)), the intensity of these peaks is reversed.

75
This suggests that micropores corresponding to these two peaks have different affinity for water adsorption, that is, the micropore with a resonance peak in the lowest field can adsorb water easier than that with a resonance peak in the higher field.

76
As pressure increases, the line width of both peaks broadens, but these peaks did not coalesce.

Discussion

Pressure dependence of 129Xe chemical shift

77
The pressure dependence of the 129Xe chemical shift of xenon confined to micropores is similar to the xenon adsorption isotherm classified in type-I by IUPAC,44,45 suggesting that the adsorption of xenon is accelerated in low pressure regions due to the deep potential, which is emphasized by the micropore field.

78
When the contribution of local magnetic fields in the pore to the chemical shift is quite small and there are no strong adsorption sites for xenon, the 129Xe chemical shift for xenon adsorbed in the pore is described byδ = δS + δXe + δEwhere δS is the interaction between xenon and the wall of the pore, and δXe is the contribution of Xe–Xe interaction.10,52

79
The second term of the right hand side of eqn. (1) is assumed to be proportional to the density of xenon in the pore, and is approximated to be δXe = δXe–Xeρ.

80
Therefore, the plateau observed on the pressure dependence of 129Xe chemical shift in high-pressure regions indicates that the adsorption of xenon in the micropores of molecular sieve 5A and 13X has reached saturation.

81
The third term is contribution of local electric field caused by charge center such as Ca2+.

82
This term will contribute the chemical shift in molecular sieve 5A containing Ca2+.

83
Although the adsorption amount of xenon in molecular sieve 5A and 13X in high-pressure regions up to 10 MPa enables the discussion of the density dependence of the chemical shift, the xenon adsorption isotherm in these systems is unknown.

84
The appropriate assumption of the adsorption of xenon in high-pressure regions leads to the pressure dependence of xenon density in micropores.

85
In this study, this is assumed by the Langmuir equation, because the physisorption due to micropore filling (type-I) can be qualitatively approximated by the Langmuir equation.53,54

86
The surface coverage of xenon in micropores, θ, is defined by the density ratio at equilibrium pressure to the density when the adsorption is saturated, ρ(P)/ρ(∞).

87
In Langmuir-type adsorption behavior, the coverage is given bywhere K is the ratio of the rate constant for adsorption to desorption.

88
Furthermore, the change in the chemical shift value between 0 MPa and infinite pressure is represented by ΔδXe–Xe = δ(∞) − δS.

89
Using these relationships, the 129Xe chemical shift at pressure, P, is represented by

90
We can discuss the adsorption behavior of xenon using the parameters, δS, ΔδXe–Xe, and K, obtained from the pressure dependence of 129Xe chemical shift value.

91
In particular, parameters such as ΔδXe–Xe and K, which relate to xenon–xenon interaction in the adsorption process of xenon, are given as high-pressure 129Xe NMR measurements, in addition to δS which can be determined from conventional 129Xe NMR measurements covering only the low-equilibrium pressure range.10–15

92
The analysis of the pressure dependence of 129Xe chemical shift using eqn. (3) leads to three kinds of parameters for each sample as listed in Table 2.

93
In NaY zeolite, the analysis was carried out using chemical shift data below 0.75 MPa.

Pore size from δS

94
δS reflects the interaction between xenon and pore walls.

95
Fraissard et al. assumed that δS is given by the average value between the chemical shifts of xenon on the wall, δa, and in free space, δV, for de-aluminated zeolites and zeolites containing only Na+ as a cation as follows:55where δV depends on the volume of free space in the pores.

96
On the basis of this assumption,56 there is a relation between δS and the mean free path, l, of xenon in the pore as represented byThis equation suggests a linear relation between the reciprocal value of δS and the pore size.

97
Assuming the shape of the pore as a sphere and cylinder, the mean free path of xenon is given by l = (DSDXe)/2 and l = (DCDXe), respectively, using the pore diameter, DS and DC for spherical and cylindrical pores and the van der Waals diameter of xenon, DXe.

98
According to these equations, the following relations are derived between δS and DC and/or DS:56

99
These equations have previously been proposed to explain the relation between the chemical shift and the free space in micropores.

100
Although it has recently been attempted to elucidate the chemical shift of xenon in the pores using ab initio MO calculations on the basis of xenon–wall as well as xenon–xenon interactions,57 it is convenient to describe the relation between δs and the pore size of zeolites.

101
Using these equations, we can evaluate the pore diameter of molecular sieve 5A and 13X for hydrated and dehydrated samples, and dehydrated zeolite NaY as listed in Table 2.

102
In molecular sieve 5A, δE also contributes to δ at zero pressure.

103
Although the effect of δE on δ will leads to overestimation of δS, δ at zero pressure, which corresponds to δS + δE, enables us to roughly estimate the pore diameter.

104
The pore diameter evaluated from the crystal structure is 1.14 nm for molecular sieve 5A and 1.37 nm for molecular sieve 13X and zeolite NaY.

105
Comparing the pore size from δS assuming the spherical and cylindrical shape to that one, the spherical and cylindrical pore model seems to be valid for molecular sieve 5A and zeolite NaY, respectively.

106
For molecular sieve 13X, the pore size from the crystal structure is the intermediate value of the spherical and cylindrical pore model.

107
This aspect may concern the mobility of adsorbed xenon in the pore.

108
In molecular sieve 5A, xenon is strongly adsorbed in the super cage, resulting in the long residence time of xenon in the cage.

109
Consequently, the adsorbed xenon assumes a spherical shape in the occupying space.

110
As the rate of translational motion of xenon increases in the pore, the residence time of the adsorbed xenon in the super cage decreases and xenon translating from the super cage to others assumes a cylindrical shape in the occupying space.

111
The effective shape of the pore describing δS implies that the mobility of xenon decreases in the order; NaY > 13X > 5A.

Influence of adsorbed water on the xenon adsorption

112
The analysis of the pressure dependence of 129Xe chemical shift in the hydrated sample for molecular sieves gives the parameters, δS, ΔδXe–Xe, and K as listed in Table 2.

113
In fact, the influence of the adsorbed water on the xenon chemical shift was already examined in NaY.66

114
In NaY, both δwall and δ1, which correspond to δS and δXe in this paper, were affected by the amount of water.

115
The effect of the presence of water on the xenon environments is different between the adsorption and desorption processes of water: Adsorption of water leads to inhomogeneous distribution of xenon environments, whereas desorption leads to uniform one.

116
In molecular sieve 5A and 13X, xenon environment seems to be uniform because of the single and symmetric peak for confined xenon (See Fig. 4).

117
This implies that the samples used in this study are in the partially dehydrated condition.

118
The δS values in the hydrated samples are larger than the correspondence values for the dehydrated samples.

119
Since the reciprocal value of δS is proportional to the pore size, which depends on the pore model, this aspect suggests that the pore size felt by adsorbed xenon decreases due to water adsorption in the hydrated samples.

120
In molecular sieve 5A, the decreased pore diameter for the spherical pore shape is 0.30 nm, which corresponds to a 0.5 monolayer of water with 0.27 nm cross-section based on the van der Waals diameter of O- and H-atoms.

121
On the other hand, in molecular sieve 13X, the decreased pore diameter is 0.95 nm for the spherical pore and 0.45 nm for the cylindrical pore and the average of these values is 0.70 nm, suggesting that about 1.3 molecular layers of water are formed on the wall, that is, the amount of adsorbed water in 13X is 2.6 times greater than that in 5A.

122
This is consistent with the amount of water adsorbed in the pretreated samples determined by TG measurements, in which the amount of adsorbed water in 13X is 2.3 times greater than that in 5A; 41 wt.% for 5A and 100 wt.% for 13X.

123
The δwall value in NaY changes from 58 ppm for completely dehydrated sample to 158 ppm for the sample containing 85% of the saturated water content.66

124
This change is similar to that in molecular sieve 13X.

125
The ΔδXe–Xe value is related to the local density of xenon in the pore when the micropore is saturated by xenon.

126
In addition, at low density where the additivity of the 129Xe chemical shift value is valid for the amount of xenon occupying the nearest neighbors, this value will approximately reflect the number of surrounding xenon atoms for a xenon atom.

127
According to the discussion on the additivity of the 129Xe chemical shift to gaseous xenon adsorbed in the α-cage of zeolite A by Jameson et al.,58 the additivity is valid for the cluster sizes up to Xe6 and the shift value increases to about 22 ppm on average by the addition of one xenon atom.

128
For Xe7 and Xe8 clusters, the chemical shift increases by about 45 ppm and 44 ppm when a xenon atom is added to the cluster, because the increase in cluster size leads to a large contribution of the chemical shielding at a shorter interatomic distance.

129
On the basis of this discussion, the ΔδXe–Xe value for a dehydrated sample for 5A and 13X suggests the formation of a Xen cluster with an average size of n = 7, whereas for the hydrated sample, the decrease in ΔδXe–Xe value leads to the average size of cluster Xe6 for 5A and of Xe3 for 13X.

130
In addition, the hydration also decreases in K value, implying that the hydrophobic interaction between xenon and the wall consisting of SiO2 and/or AlO2 is more efficient than the xenon–water interaction.

Pore structure in pellet samples

131
Figs. 6(c) and 6(d) suggest that there is a space larger than the original micropore of the zeolites in molecular sieve 13X.

132
This may be the effect of the binder agents used in preparing the pellets, that is, relatively large micropores and/or mesopores are formed in the interspace between microcrystallites.

133
This effect is clear in molecular sieve 13X.

134
Above 0.1 MPa, the extra resonance line appears at a chemical shift 15 ppm smaller than the chemical shift corresponding to xenon in the original micropore.

135
Since the potential well for adsorption due to the pores is inversely proportional to the pore size,54,59 the larger the micropore diameter, the smaller the chemical shift value.

136
Therefore, the extra peak will correspond to xenon adsorbed in the micropores and/or mesopores formed by the process of fine particle binding.

137
In addition, the broadening of the line width of the peak for xenon in larger sized pores than the original, indicates the exchange of xenon between these pores, but these peaks did not coalesce.

138
This implies that the exchange of xenon between these sites is very slow compared with the frequency separation of about 800 Hz (∼15 ppm).

139
For xenon adsorbed in the extra pores, the spectrum depends on the hydration condition of the sample.

140
In the hydrated sample, the intensity of the peak from xenon in the original micropore is smaller than that in the extra-large pores.

141
In contrast, in the dehydrated sample, the intensities of these peaks are reversed.

142
This aspect suggests that the hydration water is adsorbed in the original micropore formed by the zeolite frameworks more efficiently than the extra-large pores formed in the intercrystallites.

143
This is caused by the deeper potential depth in the original micropore of 13X framework formed by the pore wall.

144
For molecular sieve 5A, this effect cannot be observed.

Chemical exchange of xenon in NaY zeolite

145
In NaY, the pressure dependence of the resonance line with line broadening is related to the chemical exchange of xenon between the outside and inside of pores.

146
The change in the chemical shift at pressures above 1 MPa in Fig. 2 is quite different from those for molecular sieve 5A and 13X.

147
As an initial approximation, the pressure dependence of the chemical shift in pressure regions from 0.02 to 0.75 MPa is analyzed by the Langmuir adsorption model, because in this region, the pressure dependence is similar to those for the other two samples.

148
The analysis leads to the following parameters: δS = 52 ppm, ΔδXe–Xe = 121 ppm, and K = 6.0 MPa−1.

149
In the pressure region higher than 0.75 MPa, the Langmuir model cannot explain the pressure dependence of the chemical shift.

150
In contrast, part of the xenon gas co-existing with adsorbed xenon in the pore gives rise to a resonance line which coincides with the resonance for pure bulk xenon gas under the same conditions, implying that the effect of the chemical exchange of xenon is rarely affected in the observed part as free xenon gas in the spectrum.

151
When the chemical exchange of xenon between the outside and inside of the pore occurs with the exchange rate faster than the frequency separation between two resonance peaks in relation to the exchange, a resultant resonance peak appears at a frequency given by the weighted average of these two resonance frequencies.

152
The line width of the averaged resonance line is efficiently affected by the exchange rate and the population of xenon in both sites.

153
Thus, the temperature dependence of the spectrum in Fig. 3 strongly suggests the existence of the chemical exchange of xenon.

154
Using the simple two-site exchange model,60 we then analyzed the pressure dependence of the spectrum to evaluate the exchange rate for xenon between the outside and inside of the pores.

155
Two sites for xenon, bulk-xenon gas and adsorbed xenon in the pore, are denoted by A and B, respectively, where the population of each site is pA and pB and the exchange rate constant from A to B site is kA and that from B to A site is kB.

156
According to the principal of the detailed balance, pAkA = pBkB is satisfied.

157
The residence time of xenon in each site is τA = kA−1 and τB = kB−1.

158
In the slow exchange limit, the line width of the Lorentzian line shape function is then given as follows: ΔνA = (πT2A)−1 + (πτA)−1 for site A and ΔνB = (πT2B)−1 + (πτB)−1 for site B, where T2A and T2B are the transverse relaxation times for each original resonance peak of xenon at sites A and B, respectively.

159
As the exchange rate increases, the line width also increases.

160
They are given by the inverse of the line width at half maximum for each resonance peak, when the chemical exchange is negligible.

161
In the fast exchange limit, the resonance line appears at a frequency given by the weighted average between the frequencies of site A and B as follows:ν = pAνA + pBνBand the line width is given by

162
The eqn. (8) indicates that the resonance frequency and the line width strongly depend on the population of xenon in each site as well as the exchange rates.

163
Spectrum analysis was carried out using the GAMMA simulation program.61

164
The variables in the simulation are the ratio of the population of xenon in the pore to that for bulk xenon, and the exchange rate constant between these two sites.

165
In simulating the spectrum, we used various assumptions for the following points: (i) The original peak position (νB) for xenon adsorbed above 0.75 MPa is assumed to be given by the curve using the Langmuir model for the chemical shift value below 0.75 MPa.

166
(ii) The line width (T2A and T2B) for each original peak without the effect of the chemical exchange is assumed to be identical to that of the resonance peak for pure bulk xenon (ca. 50 Hz).

167
(iii) For bulk xenon gas, the chemical shift values for pure bulk xenon gas are used as the original resonance peak position (νA) at the corresponding pressure.

168
The observed and simulated spectra are shown in Fig. 7(a) and 7(b), which focus on the chemical shift range from 40 to 200 ppm.

169
The simulated spectrum on each pressure well reproduces the observed spectrum.

170
The result of the simulation and the resultant parameters such as pA/pB and the exchange rate constant from the inside of the pore to outside are shown in Fig. 8(a) and 8(b), respectively.

171
The pA/pB value and the exchange rate constant increase linearly as pressure increases from 0.75 MPa to 6 MPa, indicating that the exchange of xenon is dominated by the density of xenon for bulk xenon more efficiency than that for adsorbed xenon.

172
Above 0.75 MPa, the pressure dependence of the chemical shift for adsorbed xenon seems to be saturated, and then the chemical exchange of xenon between the outside and the inside of the pores is accelerated by the increase in xenon density outside of the pores.

173
In contrast, pA/pB seems to slightly increase above 6 MPa and the exchange rate constant decreases as pressure increases after the maximum at 6 MPa, where xenon outside the pores becomes supercritical fluid (Tc = 289.8 K and Pc = 5.8 MPa).

174
This implies that under the supercritical fluid, a constant amount of xenon outside of the pores is concerned with the exchange of xenon inside of the pores.

175
The origin of this behavior can not be clearly identified, but it may be concerned with the co-operative phenomenon in supercritical fluid such as the large fluctuation of xenon density and the large correlation length between xenon atoms.62

176
The intermolecular coherence among xenon atoms in supercritical fluid may promote the co-operative aggregation of xenon atoms such as the cluster formation.63

177
This structure formation may contribute to decrease the efficiency of xenon exchange between the outside and inside of the pores.64

178
Finally, it is very interesting to show the different behavior of NaY on the chemical exchange from that of molecular sieve 13X, althought both zeolites have a same crystal structure.

179
This different behavior on the chemical exchange will come from the difference in Si/Al values between X and Y type and/or the position of cations in the unit cell.

180
In the present study, we can not examine what effect is dominant for the chemical exchange between the inside and the outside of the pore.

181
The further study is now in progress.

Conclusion

182
In this study, high pressure 129Xe NMR measurements were carried out to examine the local structure of confined xenon in the original micropores of the framework, as well as the effect of adsorbed water and the binding of polycrystalline on the xenon environments in molecular sieve 5A, 13X and zeolite NaY.

183
The observations made were as follows:

184
1.

185
The pressure dependence of the 129Xe chemical shift value of xenon confined in the micropores of molecular sieve 5A, 13X and zeolite NaY are analyzed by the Langmuir adsorption model, leading to information about the pore size, the amount of xenon at saturated pressure, and the ratio of the apparent rate constant between the adsorption and desorption of xenon in the pores.

186
2. 129Xe NMR spectrum and their pressure dependence enable us to detect decreases in the pore diameter due to the adsorption of water and the formation of additional micropores due to the binding of polycrystalline for pellet production.

187
The super cage in molecular sieve 13X (FAU type framework) can accommodate water more efficiently than that in molecular sieve 5A (LTA type framework).

188
3.

189
The analysis of the pore diameter assuming a spherical and the cylinder shaped pore suggests that the mobility of xenon adsorbed in the micropore of the framework increases in the order NaY > 13X > 5A.

190
In zeolite NaY, the chemical exchange of xenon between the inside and outside of the micropore can be observed at high-pressure regions above 1 MPa.

191
4.

192
The exchange rate constant and the population of xenon contributing to the exchange seem to be affected by the supercritical phenomena of bulk xenon gas co-existing with adsorbed xenon in the micropore.

193
In the supercritical state, the exchange rate decreases, although the population of xenon in the micropores and the bulk gas is almost constant.

194
Thus, the high-pressure 129Xe NMR technique offers new microscopic aspects on the pore structure, the intermolecular interaction of xenon confined in the pore, and the dynamical behavior of adsorbed xenon.

195
This method will be a useful and powerful tool to investigate the porosity of materials from a microscopic point of view, and will be developed as a compensative method of the adsorption isotherm in the characterization of pore structure.