Defect Structures in Aluminosilicate Single-Walled ... - ACS Publications

Erwan Paineau , Mohamed Salah Amara , Geoffrey Monet , Véronique Peyre , Stéphan Rouzière , and Pascale Launois. The Journal of Physical Chemistry ...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Defect Structures in Aluminosilicate Single-Walled Nanotubes: A Solid-State Nuclear Magnetic Resonance Investigation G. Ipek Yucelen,† Rudra Prosad Choudhury,† Johannes Leisen,† Sankar Nair,‡ and Haskell W. Beckham*,† †

School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332, United States School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332, United States

‡

S Supporting Information *

ABSTRACT: We report a detailed investigation of the defect structures in aluminosilicate single-walled nanotubes via multiple advanced solid-state NMR techniques. A combination of 1H−29Si and 1H−27Al FSLG-HETCOR, 1H CRAMPS, and 1 H−29Si CP/MAS experiments were employed to evaluate the proton environments around Al and Si atoms in the final nanotube structure. The 1H CRAMPS spectra of dehydrated aluminosilicate nanotubes revealed the proton environments in great detail. Integration of these results with the findings from the 1H−29Si and 1H−27Al FSLG-HETCOR and 1H−29Si CP/ MAS data allows the structural assignment of all the chemical shifts and the identification of various types of defect structures in the aluminosilicate nanotube wall. In particular, we identify five main types of defect structures arising from specific atomic vacancies in the nanotube structure. It is estimated that ∼16% of Si atoms in the nanotube inner wall are involved in a defect structure. The characterization of the detailed structure of the nanotube wall is expected to have significant implications for its chemical properties and applications.

■

is a broad resonance at −90 ppm that extends up to −110 ppm, representing a small but significant fraction (as much as 20% of total Si) of less ordered silicon environments. In one study, this broad 29Si resonance was speculated to originate from a more polymerized phase containing silicate or aluminosilicate units.12 In 27Al NMR spectra, this disordered phase was also assumed to be the source of a weak Al(IV) peak near 60 ppm12 (Figure 1d and inset), which occurs in addition to the sharp peak at ∼6 ppm from octahedrally coordinated Al(VI) in the nanotube wall (Figure 1b). The disordered phase was interpreted to contain Q4 silicate tetrahedra, each connected to fewer than 3 Al atoms [Si(nAl), n < 3], with the Al atoms also being in tetrahedral coordination.12,13 However, such a disordered phase has never been observed as an impurity in extensive TEM and cryo-EM imaging of the SWNTs.5,14,15 In fact, more recently, Al(IV) has been proposed to exist as a part of the SWNT structure, either bound to the Al(VI) outer wall or substituting Si(IV) tetrahedral on the inner wall.16 Other evidence also suggests that the aforementioned resonances [0 and 60 ppm 27 Al, −90 ppm 29Si] originate from the nanotube structure itself. For example, the terminal groups of the SWNTs could consist of relatively disordered Si and Al environments and make up 2% of the Si and the Al atoms in a nanotube sample of ∼100 nm average length. Second, the SWNTs form by

INTRODUCTION Aluminosilicate (AlSiOH) single-walled nanotubes (SWNTs), with diameters in the range of 2.2−2.8 nm and lengths of ∼100−500 nm, exhibit interesting structures and unique properties.1−4 The nanotube structure is comprised of isolated silanol groups [O3SiOH] on the interior wall, connected to a curved gibbsite octahedral framework [(OH)3Al2] on the outer surface (Figure 1a,b). The well-defined and structurally complex nanopore channel formed by the material, as well as its precisely controllable diameter and functionalizable inner and outer surface hydroxyl groups,5−7 leads to molecular recognition properties and a range of emerging applications in separation, molecular storage, and catalysis. The model of the “perfect” structure (Figure 1a,b) of this nanotube material has been built up over the years via a combination of techniques.8,9 Although this structure is an accurate model of the nanotube material, several aspects of the structure remain to be fully understood and are critical to the surface properties and chemical reactivity of the nanotube, e.g., the location and dynamics of the protons on the inner and outer surfaces, and the types of structural defects existing in the nanotube.10 So far, 29Si and 27Al MAS NMR (e.g., Figure 1c,d)2,11,12 have been used to reveal the coordination environments of aluminum and silicon atoms. In 29Si NMR spectra, only two chemical shifts have been discussed. The first is a sharp resonance at −80 ppm and is conclusively ascribed to Q3(6Al) coordination5 of a silanol group connected to six Al wall atoms through bridging O atoms (Figure 1b). The second © 2012 American Chemical Society

Received: June 18, 2012 Revised: July 25, 2012 Published: July 25, 2012 17149

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

Figure 1. (a) Aluminosilicate single-walled nanotube structure, (b) the nanotube building unit composed of aluminums bridged via hydroxyl groups and silicate tetrahedra attached to the aluminum atoms through three oxygen bridges, (c) 1D 29Si CP MAS spectrum (3 ms contact time), and (d) 1D 27Al MAS NMR spectrum of a purified nanotube sample.2

information about the hydroxyl groups present in aluminosilicate SWNTs. Overall, this paper presents an updated model of the structure of the aluminosilicate SWNTs with a clearer understanding of its defect structures.

rearrangement of a number of nanoscale precursors such that the complete absence of defects would be unexpected.2 Third, as mentioned earlier, cryo-EM studies show no impurity phases but only pure nanotube products.5 Finally, Si and Al species represented by similar chemical shifts are also observed in allophanes (single-walled aluminosilicate nanoshells of ∼4 nm in diameter) and are modeled as being part of the nanoshell structure.17 In view of all the above reasons, it is clear that the minority Si and Al species are indeed part of the SWNT structure and can be considered as defect sites existing in the mostly perfect SWNT structure. In defect-free nanotubes from which adsorbed water molecules have been evacuated, only two proton environments are identified: bridging outer surface hydroxyls (Al−OH−Al) and inner-wall silanols (Si−OH) (Figure 1a,b). In this paper, our primary hypothesis is that the dynamics of protons near the defect structures can be used to interpret the local structure of the defect sites by means of NMR techniques that distinguish such environments based on those dynamics. The hydroxyl groups were first evaluated on the basis of their heteronuclear coupling with 27Al or 29Si nuclei by means of 2D FSLG HETCOR NMR, which incorporates frequency-switched Lee− Goldburg (FSLG) decoupling into the HETCOR pulse sequence and achieves excellent resolution in the proton dimension. 1H → 29Si CP/MAS NMR experiments were used to distinguish Si signals that are coupled with protons. The information obtained from these techniques was then integrated into the analysis of high-resolution 1D 1H MAS and 1H CRAMPS NMR experiments to obtain further

■

EXPERIMENTAL DETAILS The synthesis of aluminosilicate SWNTs was carrried out as described in detail elsewhere.2 Aluminosilicate nanotubes were synthesized from an aqueous solution of 0.1 M aluminum secbutoxide and 0.05 M tetraethyoxysilane, with 0.05 M HClO4 being used as the acid medium. The solution was aged at room temperature for 18 h, and the system evolved to an aluminosilicate colloid with a local structure close to that of the nanotube. Upon heating to 95 °C, nanotubes continuously form and grow by mechanisms elucidated in recent publications.2,3 After 96 h of heating, the obtained nanotube gel was dialyzed against DI water with a 15 kDa membrane to obtain a purified nanotube dispersion. Freeze-drying preserves the molecular structure of species in solution.18 NMR experiments were performed on freeze-dried materials sampled from the reactor at various stages of the nanotube synthesis and also on freeze-dried nanotube dispersions after dialysis. Dehydration of a pure freeze-dried nanotube sample was carried out in a Schlenk line at 250 °C for 24 h under vacuum. Samples were then transferred to zirconia rotors under a nitrogen atmosphere in a glovebox and sealed with an O-ring cap prior to NMR measurements. All NMR spectra were collected at room temperature on a Bruker AV3-400 spectrometer operating at 400 MHz for 1Hs 17150

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

Figure 2. 2D 1H−29Si FSLG-HETCOR spectra of (a) freeze-dried nanotubes after purification by dialysis and (b) nanotubes dehydrated in vacuo at 250 °C. Contact time is 2500 μs in all experiments. Asterisk indicates center spin-lock frequency.

experiments is challenging to calibrate accurately.25 Instead of quantitative chemical shifts, here we focus on understanding the different types of defect-related proton environments in comparison to the ideal proton environments. Of the two proton environments identified in freeze-dried nanotubes (see Figure 2a), one of the 1H resonances is assigned to silanol units that are hydrogen bonded to water oxygens (Si−OH···OH2). The other 1H resonance represents the molecular water hydrogen-bonded to silanol groups (Si− OH···OH2). Previous multilayer adsorption simulations showed that water molecules immobilized on the nanotube inner wall can form water layers by hydrogen bonding.6 Moreover, Figure 2a indicates that the Si−OH···OH2 species is also found at a variety of different Si environments represented by ca. −90 ppm 29Si NMR resonances that are different from the dominant Q3(6Al) configurations. Molecular water is therefore also bound to defect areas in nanotube. As mentioned in the Introduction, it has been previously proposed that the broad 29 Si resonance could belong to polymerized Q4 silicate or aluminosilicate environments,12 in which case the second coordination sphere of Si is occupied by Al or Si atoms and not by −OH groups.11−13 On the other hand, Figure 2a,b shows strong proton correlations with Si in the species represented by the broad peak centered around −90 ppm in 29Si MAS spectra. This indicates the existence of −OH groups in the second coordination sphere of Si (i.e., Q3 or Q2 Si species) or structures that are more open than a polymerized unit, thereby allowing binding of water protons and resulting in the observed proton correlations. Dehydration of the nanotubes at 250 °C under vacuum completely removes adsorbed water molecules and isolates the Si−OH structures that are at various Si chemical shifts including defects as shown in Figure 2b. For a successful modification of interior nanotube walls, it was shown that adsorbed water molecules should be removed completely by an appropriate heat treatment condition.6 Our present NMR study also confirms that heat treatment at 250 °C completely removes the water molecules while preserving the nanotube structure.5,6 As discussed before, the 29Si NMR spectra of aluminosilicate nanotubes exhibit resonances from about −80 to −110 ppm. This indicates the presence of disordered 29Si environments in addition to the perfectly ordered configuration.11 In fact, our 2D 1H−29Si FSLG-HETCOR studies indicates the existence of

and with a 10 kHz spinning frequency. The 27Al chemical shifts are reported with respect to aluminum nitrate, whereas the 29Si and 1H chemical shifts are reported with respect to 3(trimethylsilyl)-1-propanesulfonic acid sodium salt. 1H → 29Si CP/MAS NMR experiments were performed with a 5 s repetition rate, 1024 scans, and varying contact times. 1H CRAMPS spectra were recorded using a windowed phasemodulated Lee−Goldburg (wPMLG) pulse sequence with 2.8 μs π/2 pulses. 1H CRAMPS spectra were calibrated using glycine. This calibration was performed to obtain pulse lengths, chemical shift scales, and power levels prior to the experiment. 1 H MAS NMR spectra were collected using single-pulse experiments with a 3 μs π/2 pulse, a 2 s repetition time, and 16 scans. Two-dimensional HETCOR (heteronuclear correlation) experiments were initially limited by 1H−1H dipolar coupling, which caused line broadening in the 1H spectra. Therefore, frequency-switched Lee−Goldburg (FSLG) decoupling was incorporated into the HETCOR sequence.19 This pulse sequence is based on cross-polarization from one nucleus to another20 and was adapted from the literature.21 The 1H → 29Si experiments were recorded with a ramped-amplitude CP sequence using a contact time of 2.5 ms; 128 scans were collected. Spectra were acquired with 64 t1 increments of 63 μs, a 1H π/2 pulse of 3 μs, a dwell time of 20 μs, and a recycle delay of 5 s. 1H → 27Al FSLG HETCOR experiments were performed with a 1H π/2 pulse of 3 μs, a 200 μs contact time, and a 4 s repetition rate. A set of 64 free induction decays was obtained with a t1 increment of 63 μs, states-TPPI (time proportional phase incrementation) phase cycling,22 and a dwell time of 12 μs. The protons were decoupled from the nuclei of interest using a TPPM proton decoupling23 [twopulse phase modulation] scheme.

■

RESULTS AND DISCUSSION H → 29Si FSLG-HETCOR and 2D 1H → 27Al experiments were performed to identify Al-correlated or Si-correlated 1H chemical shifts in the final nanotube structure. We also simultaneously carried out 29Si NMR cross-polarization contact time experiments to identify previously unresolved Si chemical shifts in the nanotubes. 1 H−29Si FSLG-HETCOR and 29Si CP NMR Experiments. In hydrated nanotubes (see Figure 2a), two different proton environments were identified, both of which belong to Q3(6Al) Si−OH units (at −80 ppm). The 1H scale in FSLG-HETCOR 1

17151

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

Figure 3. 1H → 29Si CP/MAS NMR spectra of freeze-dried aluminosilicate nanotubes at contact times of (a) 1 ms and (b) 15 ms.

species with various 29Si chemical shifts that appear in the approximately −80 to −110 ppm region. The grouping of Si atoms with hydroxyl groups (−OH) can be facilitated by the use of 1H → 29Si cross-polarization (CP) between protons and silicon atoms that are within the van der Waals contact distance (0.33 nm).18 The application of CP enhances selectively the Si signals of atoms that are coupled with protons by internuclear 1 H → 29Si dipole interactions. 29Si CP NMR contact-time experiments (cf. Figure 3) were carried out to clarify the dominant silicon species that are coupled with protons represented by the broad −90 ppm resonance and to identify the chemical shifts of unresolved silicon species under this broad peak. Each spectrum was deconvoluted with a minimum necessary number of Lorentzian peaks (also see magnified insets). Our CP experiments show that the number of detectable peaks depends on the contact time. There are five Si environments (at −80.4, −86.6, −92.1, 100, and 105.5 ppm) detected at a short contact time of 1 ms (Figure 3a), of which only three (at 80.4, −86.6, and −92.1 ppm) are seen at a higher contact time of 15 ms (Figure 3b). One explanation could be that protons sufficiently close to silicon atoms allow intermolecular polarization transfer at short contact times. Therefore, the peaks near −100 and −105.5 ppm belong to silicon species that have protons in close proximity. At longer contact times, those peaks disappear possibly due to the absence of polarization transfer from distant protons. Additionally, the relaxation of the protons is also important and should be taken into account.18,26 Structural interpretation of chemical shifts in the −80 to −110 ppm region is complicated because this region is composed of overlapping signals from Q4, Q3, and Q2 environments.11,18,27−29 Both silicate and aluminosilicate environments11 were reported for peaks included in this broad region. However, previous studies can provide guidelines in our interpretation of these peaks. 29Si chemical shifts are highly sensitive to the second coordination sphere. For example, a 5.5 ppm chemical shift difference within pairs of 29 Si signals (−86.6 and −92.1 ppm or −100 and −105.5 ppm) indicates aluminum deshielding in the second coordination sphere of Si atoms in a Q4 aluminosilicate unit.30 Moreover, when one or two hydroxyl groups are attached to Si atoms (i.e., Q3 and Q2) in a silica gel, chemical shifts of −99.8 and −90.6 ppm, respectively, are reported with a 10 ppm difference.28,29 The fact that the tail of the broad peak tail never reaches −110 ppm indicates that there is no Si(0Al) or pure polymerized silica in the sample.29

Table 1 shows all the possible Si configurations consistent with the above discussion. It also shows likely configurations of water binding to the Q3 Si(6Al)−OH units as suggested by our 2D HETCOR studies described earlier in this report. As summarized in Table 1, the Si environments (−86.6 and −92.1 Table 1. Chemical Environments Detected by HETCOR and CP Experiments

17152

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

Figure 4. 2D 1H−27Al FSLG-HETCOR spectra of (a) freeze-dried AlSiOH nanotubes and (b) nanotubes dehydrated in vacuo at 250 °C. Contact time is 200 μs in all experiments. The asterisk indicates center spin-lock frequency and spinning side bands.

Figure 5. 1H MAS NMR spectra of freeze-dried hydrated (20 wt % water) aluminosilicate nanotubes: (a) with indicated spinning side bands (∗) showing phase correction before deconvolution and (b) the center peak with deconvolved peaks. 1H MAS NMR spectra of nanotubes dehydrated at 250 °C under vacuum for 24 h: (c) the center peak with deconvolved peaks and (d) CRAMPS spectrum showing deconvolved peaks.

ppm or −100 and −105.5 ppm) correspond to a range of Q4 Si(nAl) units, where n = 1−4 is the number of Al atoms in the coordination sphere. However, the peaks at ca. −92.1 and −100 ppm could also be assigned to Q2 and Q 3 silanols, respectively.18,27 The peak at −105.5 ppm is assigned to resonances arising from coordinations to three Si atoms (3Si, 1Al).26,27 We also note that the proton environments bound to the Si configurations are confirmed by 1H NMR studies, as described later in this report.

1

H−27Al FSLG-HETCOR Experiments. High-resolution heteronuclear 1H−27Al correlation NMR spectra of nanotubes are shown in Figure 4. Tetrahedrally coordinated Al atoms are not visible in 1H−27Al HETCOR spectra even though their presence is confirmed by 1D 27Al spectra.2 This does not necessarily preclude a correlation between proton and tetrahedral Al, which comprises less than 1% of total detectable Al in the nanotube. Therefore, it likely does not appear above noise level at a contact time of 200 μs. Further studies may be necessary to reveal proton correlations of tetrahedral Al. The 17153

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

Table 2. Structural Representations of 1H MAS (Left Column) and 1H CRAMPS NMR Peak Assignments (Right Column)

1

H−27Al correlation spectrum of freeze-dried nanotubes is shown in Figure 4a. Three different proton environments are clearly identifiable, whereas the main 27Al signal is at 8 ppm, arising due to Al in the nanotube configuration. Molecular depictions of chemical environments resulting in each proton signal are given in Figure 4a. The dominant signal is assigned to physisorbed water molecules on bridging Al hydroxyl groups [Al2(OH)···H2O] and the remaining signals tentatively assigned to bridging Al hydroxyl groups [Al2(OH)···H2O] and water molecules, which either directly bound to Al atoms [Al−H2O] or in between nanotubes connecting outer gibbsite walls. When the nanotubes are dehydrated completely, the bridging hydroxyl protons (Al2−OH) are isolated leaving only their proton signal correlated to Al atoms (Figure 4b). 1 H MAS and 1H CRAMPS NMR. 1D 1H MAS and 1H CRAMPS NMR can give direct information about different hydroxyl groups and their environments in the aluminosilicate nanotubes. However, no previous study exists on this aspect, and very little 1H chemical shift data are available for singlewalled metal oxide nanotubes.5 The nanotubes are hydrated up to 20 wt % at room temperature after freeze-drying.5 Because of the dominance of signal from water molecules, or because of strong proton dipolar couplings and chemical exchange between hydroxyl groups, only a single peak is observed in 1 H MAS NMR spectra. Hence, 1H MAS NMR has not been commonly employed as a method of characterizing the

nanotubes. In the present work, the 2D HETCOR experiments described earlier, in conjunction with 1D 1H MAS and 1H CRAMPS NMR discussed in this section, allowed us to evaluate the molecular environments in the nanotubes. These spectra are shown in Figure 5 with peak assignments summarized in Table 2. Figure 5a displays the 1H MAS NMR spectrum of the freezedried hydrated aluminosilicate nanotubes, which includes spinning side bands, showing how baseline correction of spectra is done before spectral deconvolution. Figure 5b shows the main peak in the 1H MAS NMR spectrum along with deconvolved peaks that were obtained by fitting with a minimum number of Lorentzian and Gaussian peaks using the DMfit software.31 Four fitted peaks are shown in Figure 5b at 4, 5, 6, and 7 ppm. Assignments of these peaks are based on the chemical environments indicated by our 2D HETCOR studies and from the literature. The peaks at 5−6 ppm are due to water molecules hydrogen-bonded to surface hydroxyl groups at Si and Al sites in the intratube and intertube channels (cf. Table 2, left column).32−35 An intense broad peak centered at 7 ppm could be due to the inner-wall silanol (Si− OH···OH2) groups with bound water molecules in a variety of possible configurations (cf. Table 1). The reason for the large breadth of this peak could be the motional restrictions induced by adsorbed water layers.6 The 4 ppm peak is assigned to bridging hydroxyl (Al2−OH···OH2) groups with bound water 17154

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

structures likely to exist in the aluminosilicate nanotubes in Figure 6. As discussed before, any kind of defect would

molecules, as also suggested by the HETCOR measurements. For nanotubes dehydrated at 250 °C under vacuum for 24 h, two peaks (at 3 and 5 ppm) were sufficient to deconvolute the 1 H MAS spectrum. The main peak (area = 60% of the total) of the 1H MAS NMR spectrum shifts to 3 ppm (Figure 5c) and is due to the bridging hydroxyl (Al2−OH) groups on the outer wall, whereas the 5 ppm peak (40%) is due to the inner-wall silanols (Si−OH). Although the deconvolution of 1H NMR spectra is given and discussed here in detail based upon the minimum number of peaks required for a good fit, their evaluation depends somewhat on the number of peaks selected for deconvolution and could therefore be subject to minor differences in interpretation. The use of 1H CRAMPS data allowed us to accurately resolve the number of proton environments in nanotubes. 1H CRAMPS increases the spectral resolution of 1H MAS NMR spectra by suppressing homonuclear dipolar interactions between protons. We found that the technique was only successful after the elimination of water molecules. This could be due to the removal of water-mediated chemical exchanges/ motions on the time scale of CRAMPS experiments upon dehydration of the nanotubes. Six clearly distinguishable 1H CRAMPS resonances were identified after complete removal of bound water by in vacuo dehydration at 250 °C for 24 h (Figure 5d and Table 2, right column). Slight changes in the interpretation of ppm values in comparison to the hydrated sample can be expected after the removal of physisorbed water. The relative percentage of each peak was used in the interpretation of the identified peaks. In the solid form, the nanotubes are organized into a network of bundles within which the individual nanotubes are packed into a locally ordered arrangement.5,7 The resonance near 3 ppm is assigned to Al bridging protons involved in hydrogen bonding to oxygen atoms on the outer wall of an adjacent nanotube (Al2− OH···O). The peak at 2 ppm arises from protons in isolated bridging (Al2−OH) groups on the outer nanotube circumference. The peak at 6 ppm is assigned to the protons associated with the isolated Si−OH groups on the inner walls. The ratio of the area percentages of Al bridging protons (36%) to silanol protons (10%) is about 3, further supporting our interpretation based on the chemical formula of the nanotube material [(OH)3Al2O3Si(OH)]. The above assignments cover the expected proton environments in the perfect nanotube structure. In addition, there is a sharp peak at 3.7 ppm which possibly belongs to mobile water or hydroxyl ligands in extraframework Al or Si species such as those observed in zeolites.36 A broad resonance near 4.5 ppm is assigned to hydroxyl protons that bridge Si and Al atoms (Al− OH−Si) in lattice defects (2% of total spectral area).27,37 The −0.5 ppm peak is probably due to (Al−OH) protons in lattice defects at the outer nanotube wall or nanotube ends.24,38 In summary, 1H CRAMPS spectra of dehydrated nanotube revealed the existence of a variety of proton environments in the nanotube structure and also confirmed that dehydration at 250 °C removes the bound water without any significant effect on the nanotube structure.5 It is suggested that the simplicity of this method may render it very useful in studying structural transformations (e.g., induced by heating at higher temperatures) or chemical surface modifications of nanotube materials. Proposed Defect Structures. As evinced by a combination of NMR techniques, several types of disordered regions exist in aluminosilicate nanotubes. By integrating our NMR findings described in the previous sections, we depict the defect

Figure 6. Wall structure of the single-walled aluminosilicate nanotube, depicting the different defect environments. Colored areas indicate disordered Al, Si, and H environments created by defects (numbered 1 to 5 and discussed in the text). The NMR confirmation details are shown in Figure S2.

introduce lattice strain, which is a possible cause for the broadening of 27Al NMR line shapes. 29Si chemical shifts are more sensitive to changes in the surrounding atoms and thereby provided more definite clues regarding the structure of the defects. Our studies confirm that the disordered regions are part of the nanotube walls and not an impurity phase. The resonances seen in our 27Al, 29Si, and 1H NMR spectra that do not originate from “perfect” bonding configurations in the nanotube represent ∼10% of the overall nanotube structure based on their average NMR percentages. The most probable defect structures are those that incorporate species represented by the −80 to −90 ppm 29Si chemical shift region, due to their dominance in the 29Si NMR spectra over other defect peaks. Figure 6 illustrates the most frequent structural defects in nanotubes as deduced from our NMR data, together with the corresponding chemical environments which are confirmed by our comprehensive 1H, 27Al, and 29Si NMR investigation. The remaining, less probable, defects are shown in the Supporting Information (Figure S1). A brief discussion of each type of defect depicted in Figure 6 is given as follows: Defect 1 involves the absence of an Al atom from the nanotube wall, resulting in voids similar to those observed in allophanes.12,43,44 Such Al vacancy defects might be the source of dangling Al−OH bonds in an otherwise perfectly hydroxyl-bridged Al lattice.24,38 One such Al−OH bond is colored and circled with a solid line in Figure 6. It is the source of the 1H NMR signal observed at −0.5 ppm.38 This defect also leads to Si−OH−Al hydroxyls (1H NMR signal at 4.5 ppm) bridging the inner and outer walls of the nanotubes, as indicated by a dashed circle in Figure 6. We note that similar Al−OH and Si−OH−Al environments can be observed as a result of nearly all defects depicted in Figure 6, and the resulting 1H NMR signals are therefore not limited to defect 1. This is a further confirmation that the broadening of the 27Al 17155

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article 29

line shape is likely due to outer wall defects. Defect 1 results in a 29Si NMR signal at around −86.6 ppm. Defect 2 involves the presence of an additional Si atom in the inner wall, causing the formation of Si−O−Si bonds. This defect causes the formation of two irregular chemical environments around Si [(3Al, 3Si), (5Al, 1Si)] as indicated in Figure 6. Both can give rise to 29Si NMR signal near −86.6 ppm. In any type of defect, as in this one, an asymmetric environment around aluminum emerge, possibly the cause for quadrupolar broadening. Moreover, the broadening of peaks centered around −90 ppm in 29Si spectra is possibly due to such defect 2 induced lattice strains. Defect 3 indicates the absence of two Al atoms in the outer wall that could cause instability or lattice strain, compensated by the condensation between two adjacent silanols along the nanotube circumference. This results in the formation of defective Q4 and Q3 Si sites, giving rise to 29Si NMR signals near −92.1 ppm. Defect 4 involves the absence of an Al atom and an extra Si atom in the framework, resulting again in the formation of Si− O−Si bonds. Defect 5 depicts the formation of Q3(4Al) Si−OH groups due to missing two Al atoms in the outer framework. Both defect 4 and defect 5 are evinced by the broad signal centered at −92.1 ppm. Please see Supporting Information (Figure S2) for the chemical environments around each defect and NMR confirmation. Figure 6 shows each type of defect only once, but it is seen that vacancy-type defects result in multiple distorted chemical environments around the vacancy. This explains the observation of a high defect percentage in the nanotube structure (which involves as much as ∼16% of the Si atoms). Most of the defects result from vacancies of Al atoms in the outer nanotube wall. These types of defects might be expected because of the formation mechanism of these nanotubes. It has recently been shown that pure aluminate species together with aluminosilicate precursors remain in solution throughout the synthesis, whereas there are no pure silicate species.2 Therefore, the yield of complexation between Al and Si is not expected to be 100%; that is, not all Al atoms introduced into the synthesis take part in nanotube formation, thereby resulting in a deviation from the Al/Si ratio of 2 of the “perfect” nanotube structure and the formation of Al vacancy defects. The existence of such structures has interesting implications for the nanotube properties and potential applications.7,45

Si FSLG-HETCOR measurements gave direct information about protons coordinated either with Al or Si. Our 2D correlation studies also confirmed 1H NMR chemical shift assignments. Finally, integration of all our NMR findings resulted in the first comprehensive molecular model of defect structures and associated proton environments in aluminosilicate single-walled nanotubes.

■

ASSOCIATED CONTENT

S Supporting Information *

Close-up 3D models of defect structures with corresponding NMR confirmations and table of additional possible defect configurations. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (CAREER, CBET-0846586). REFERENCES

(1) Yucelen, G. I.; Kang, D.-Y.; Guerrero-Ferreira, R. C.; Wright, E. R.; Beckham, H. W.; Nair, S. Nano Lett. 2012, 12, 827. (2) Yucelen, G. I.; Choudhury, R. P.; Vyalikh, A.; Scheler, U.; Beckham, H. W.; Nair, S. J. Am. Chem. Soc. 2011, 133, 5397. (3) Mukherjee, S.; Bartlow, V. M.; Nair, S. Chem. Mater. 2005, 17, 4900. (4) Farmer, V. C.; Fraser, A. R.; Tait, J. M. J. Chem. Soc., Chem. Commun. 1977, xxxx. (5) Kang, D.-Y.; Zang, J.; Wright, E. R.; McCanna, A. L.; Jones, C. W.; Nair, S. ACS Nano 2010, 4, 4897. (6) Kang, D.-Y.; Zang, J.; Jones, C. W.; Nair, S. J. Phys. Chem. C 2011, 115, 7676. (7) Bonelli, B.; Bottero, I.; Ballarini, N.; Passeri, S.; Cavani, F.; Garrone, E. J. Catal. 2009, 264, 15. (8) Ildefonse, P.; Kirkpatrick, R. J.; Montez, B.; Calas, G.; Flank, A. M.; Lagarde, P. Clays Clay Miner. 1994, 42, 276. (9) Cradwick, C. P. G.; Farmer, V. C.; Russell, J. D.; Masson, C. R.; Wada, K.; Yoshigana, N. Nat. Phys. Sci. 1972, 240, 187. (10) Levard, C.; Masion, A.; Rose, J.; Doelsch, E.; Borschneck, D.; Olivi, L.; Chaurand, P.; Dominici, C.; Ziarelli, F.; Thill, A.; Maillet, P.; Bottero, J. Y. Phys. Chem. Chem. Phys. 2011, 13, 14516. (11) Barron, P. F.; Wilson, M. A.; Campbell, A. S.; Frost, R. L. Nature 1982, 299, 616. (12) Goodman, B. A.; Russell, J. D.; Montez, B.; Oldfield, E.; Kirkpatrick, R. J. Phys. Chem. Miner. 1985, 12, 342. (13) Wilson, M. A.; Lee, G. S. H.; Taylor, R. C. J. Non-Cryst. Solids 2001, 296, 172. (14) Yucelen, G. I.; Kang, D.-Y.; Guerrero-Ferreira, R. C.; Wright, E. R.; Beckham, H. W.; Nair, S. Nano Lett. 2012, XX, xxxx. (15) Yang, H.; Wang, C.; Su, Z. Chem. Mater. 2008, 20, 4484. (16) Hu, J.; Kamali Kannangara, G. S.; Wilson, M. A.; Reddy, N. J. Non-Cryst. Solids 2004, 347, 224. (17) Hiradate, S.; Wada, S.-I. Clays Clay Miner. 2005, 53, 401. (18) Burkett, S. L.; Davis, M. E. J. Phys. Chem. 1994, 98, 4647. (19) Burum, D. P.; Bielecki, A. J. Magn. Reson. 1991, 94, 645. (20) Bielecki, A.; Kolbert, A. C.; Levitt, M. H. Chem. Phys. Lett. 1989, 155, 341. (21) van Rossum, B. J.; Förster, H.; de Groot, H. J. M. J. Magn. Reson. 1997, 124, 516.

■

CONCLUSIONS In this work, we have used advanced NMR characterization tools to understand and elucidate local defect structures in a complex nanoscopic material, i.e., an aluminosilicate singlewalled nanotube. Such information is of great value in predicting and controlling the functional properties of the nanotube, which depend critically on the structure and composition of the walls. In particular, we performed comprehensive 1H CRAMPS, 1H MAS, 29Si CP/MAS, 1H → 27 Al, and 1H → 29Si FSLG-HETCOR measurements on asmade and dehydrated nanotubes. The 1H CRAMPS and 1H MAS experiments revealed the interaction between bound water molecules and surface hydroxyl groups (Al2OH and Si−OH sites) and specificially showed the existence of six types of protons in the aluminosilicate nanotubes. 29Si CP/MAS experiments selectively enhanced the Si signals with variable contact times, exposing a variety of previously obscured Si environments and indicating differences in the proton pools surrounding these environments. Our 1H → 27Al and 1H → 17156

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157

The Journal of Physical Chemistry C

Article

(22) Marion, D.; Ikura, M.; Tschudin, R.; Bax, A. J. Magn. Reson. 1989, 85, 393. (23) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.; Griffin, R. G. J. Chem. Phys. 1995, 103, 6951. (24) Arancibia-Miranda, N.; Escudey, M.; Molina, M.; GarcíaGonzález, M. T. J. Non-Cryst. Solids 2011, 357, 1750. (25) Coelho, C.; Rocha, J.; Madhu, P. K.; Mafra, L. J. Magn. Reson. 2008, 194, 264. (26) Maciel, G. E.; Sindorf, D. W. J. Am. Chem. Soc. 1980, 102, 7606. (27) Zhang, W.; Bao, X.; Guo, X.; Wang, X. Catal. Lett. 1999, 60, 89. (28) Jaymes, I.; Douy, A. J. Eur. Ceram. Soc. 1996, 16, 155. (29) Xu, M.; Arnold, A.; Buchholz, A.; Wang, W.; Hunger, M. J. Phys. Chem. B 2002, 106, 12140. (30) Merzbacher, C. I.; McGrath, K. J.; Higby, P. L. J. Non-Cryst. Solids 1991, 136, 249. (31) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magn. Reson. Chem. 2002, 40, 70. (32) Xu, M.; Harris, K.; Thomas, J. Catal. Lett. 2009, 131, 16. (33) Gunawidjaja, P. N.; Mathew, R.; Lo, A. Y. H.; Izquierdo-Barba, I.; García, A.; Arcos, D.; Vallet-Regí, M.; Edén, M. Philos. Trans. R. Soc., A 2012, 370, 1376. (34) Gun’ko, V. M.; Turov, V. V.; Shpilko, A. P.; Leboda, R.; Jablonski, M.; Gorzelak, M.; Jagiello-Wojtowicz, E. Colloids Surf., B 2006, 53, 29. (35) Gun’ko, V. M.; Turov, V. V. Langmuir 1999, 15, 6405. (36) Huang, J.; Jiang, Y.; Marthala, V. R. R.; Thomas, B.; Romanova, E.; Hunger, M. J. Phys. Chem. C 2008, 112, 3811. (37) Dieter, F. Chem. Phys. Lett. 1995, 235, 69. (38) Jiang, Y.; Huang, J.; Dai, W.; Hunger, M. Solid State Nucl. Magn. Reson. 2011, 39, 116. (39) Malicki, N.; Mali, G.; Quoineaud, A.-A.; Bourges, P.; Simon, L. J.; Thibault-Starzyk, F.; Fernandez, C. Microporous Mesoporous Mater. 2010, 129, 100. (40) Lee, S. K.; Stebbins, J. F. J. Phys. Chem. B 2000, 104, 4091. (41) Hatakeyama, M.; Hara, T.; Ichikuni, N.; Shimazu, S. Bull. Chem. Soc. Jpn. 2011, 84, 656. (42) Xue, X.; Kanzaki, M. Solid State Nucl. Magn. Reson. 2007, 31, 10. (43) J. Phys. Chem. C 2011, 110726095844008. (44) Creton, B.; Bougeard, D.; Smirnov, K. S.; Guilment, J.; Poncelet, O. J. Phys. Chem. C 2007, 112, 358. (45) Levard, C.; Masion, A.; Rose, J.; Doelsch, E.; Borschneck, D.; Olivi, L.; Chaurand, P.; Dominici, C.; Ziarelli, F.; Thill, A.; Maillet, P.; Bottero, J. Y. Phys. Chem. Chem. Phys. 2011, 13, xxxx.

17157

dx.doi.org/10.1021/jp3059728 | J. Phys. Chem. C 2012, 116, 17149−17157