Surface nano-architecture of a metal2013organic framework

Surface nano-architecture of a metal2013organic framework

(Parte 1 de 2)

Surface nano-architecture of a metal–organic framework

The rational assembly of ultrathin films of metal–organic frameworks (MOFs)—highly ordered microporous materials1–5 —with well-controlled growth direction and film thickness is a critical and as yet unrealized issue for enabling the use of MOFs in nanotechnological devices, such as sensors, catalysts and electrodes for fuel cells. Here we report the facile bottom-up fabrication at ambient temperature of such a perfect preferentially oriented MOF nanofilm on a solid surface (NAFS-1), consisting of metalloporphyrin building units. The construction of NAFS-1 was achieved by the unconventional integration in a modular fashion of a layer-bylayer growth technique coupled with the Langmuir–Blodgett method. NAFS-1 is endowed with highly crystalline order both in the out-of-plane and in-plane orientations to the substrate, asdemonstratedbysynchrotronX-raysurfacecrystallography. Theproposedstructuralmodelincorporatesmetal-coordinated pyridine molecules projected from the two-dimensional sheets that allow each further layer to dock in a highly ordered interdigitated manner in the growth of NAFS-1. We expect that the versatility of the solution-based growth strategy presented here will allow the fabrication of various well-ordered MOF nanofilms, opening the way for their use in a range of important applications.

MOFs are highly ordered crystalline coordination polymers with well-defined porous networks. As a result of their potential applications associated with the presence of pores and cavities, such materialshavebeenexaminedforstorage,separations,sensors,sizeand shape-selective catalysis, molecular recognition and nanoscale reactors2–5. Metalloporphyrins are very stable pi-conjugated macrocyclic molecules forming a diverse class of multifunctional materials, which exhibit interesting and useful catalytic, optical, electronic and biological properties6. Metalloporphyrin molecules are excellent candidates to act as components of molecular building blocks for the construction of MOFs because of the huge variety of available derivatives incorporating different centre-coordinated metals and various functional group substituents in the periphery of the molecule, thereby allowing fine control of the linkage motif. Bulk crystals of MOFs comprising porphyrin molecules as building blocks7–9 and their highly selective adsorption and/or catalytic properties10–12 have been reported.

However, sensors, catalytic devices and other related nanotechnological devices and applications using porous materials depend critically on the availability of thin films and their

1Department of Chemistry, Faculty of Science, Kyushu University, Hakozaki 6-10-1, Higashi-ku, Fukuoka 812-8581, Japan, 2CREST, Japan Science and Technology Agency, Sanbancho 5, Chiyoda-ku, Tokyo 102-0075, Japan, 3Department of Applied Chemistry, National Defense Academy, Yokosuka, Kanagawa 239-8686, Japan, 4Japan Synchrotron Radiation Research Institute, Kouto 1-1-1, Sayo-cho, Sayo-gun, Hyogo 679-5198, Japan, 5Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan, 6INAMORI Frontier Research Center, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-3095, Japan. †Present address: Nanoscience and Nanotechnology Research Center, Osaka Prefecture University, Gakuencho 1-2, Naka-ku, Sakai, Osaka 599-8570, Japan. *e-mail:r-makiura@21c.osakafu-u.ac.jp; kitagawa@kuchem.kyoto-u.ac.jp.

integration with other components such as electrodes—this is a major challenge that needs to be addressed13,14. As a first step towards this goal, two-dimensional (2D) nanostructures (monolayer) comprising tailored pore sizes and functionalities15–17 or composed of porphyrin molecules18 on a solid surface have been constructed. In addition, attempts to fabricate MOF films (multilayers) have been reported by applying direct crystal growth from solvothermal solution or by layer-by-layer growth13,19–23. Although these reports show promising results in obtaining partially ordered systems or with some ordering preference, the assembly of thin MOF films with completely controlled size and growth direction has not been achieved as yet, and the structural details, especially in the film orientation parallel to the substrate surface, remain unknown.

The basis of the layer-by-layer growth technique is that the reaction components are combined in a sequential manner. Unreacted or physisorbed components are removed between successive deposition steps by rinsing with an appropriate solvent. Such a layer-by-layer protocol has been widely used to grow supramolecular architectures on surfaces24 but only recently it has been applied for the fabrication of MOF (refs 21–23) or crystalline coordination polymer thin films25. A distinct advantage of this growth mode is that it provides good control of the film orientation and the number of layers in the vertical direction to the substrate (out-of-plane). However, comparable control is not established in the horizontal direction parallel to the substrate (in-plane), although a periodic structure sometimes occurs as a consequence of van der Waals or pi–pi interactions, for instance in self-assembled monolayers26. On the other hand, the Langmuir–Blodgett method is an excellent technique for the fabrication of well-ordered monolayers on liquid surfaces that can be deposited and stacked onto solid substrates. The Langmuir–Blodgett technique has been applied to various functional molecular systems, leading to the successful assembly of well-organized 2D arrays27–30.

The rational bottom-up assembly of MOF thin films on substrates, where crystalline order is endowed in both out-of-plane and in-plane orientations, has remained an elusive target. Here, we demonstrate how the layer-by-layer growth and the Langmuir–Blodgett methods can be adeptly integrated in a modular fashion to allow fabrication on a solid surface under mild conditions of a perfectly crystalline MOF thin film comprising metalloporphyrin building units and metal ion joints, as schematically shown in Fig. 1. The procedure involved

NATURE MATERIALS | VOL 9 | JULY 2010 | w.nature.com/naturematerials 565

LETTERS NATUREMATERIALSDOI:10.1038/NMAT2769

Molecular building units Top view of CoTCPP-py-Cu 2D array

Rinse and immerse2D array formation by Langmuir¬Blodgett method and deposition onto the substrate

CuCl2.2H2O, 3 aqueous solution

Substrate Layer-by-layer growth

NAFS-1

CoTCPP, 1py, 2

Figure 1 | Schematic illustration of the fabrication method of NAFS-1. The solution mixture of CoTCPP (1) and pyridine (2) molecular building units is spread onto an aqueous solution of CuCl2·2H2O, 3 in a Langmuir trough. Pressing the surface with barrier walls leads to the formation of a copper-mediated CoTCPP 2D array (CoTCPP-py-Cu) (Langmuir–Blodgett method). Surface-pressure/area (pi–a) isotherms were measured to evaluate the molecular area (Supplementary Fig. S1). The 2D arrays are deposited onto the substrate by the horizontal dipping method at room temperature. The substrate is then immersed in the pure solvent to remove excess starting materials or physisorbed components and dried. The repetitive process of successive sheet deposition and rinsing/solvent immersion leads to the sequential layer-by-layer growth of NAFS-1 with any desired thickness. C atoms are shown in grey, N in blue, O in red, Co2+ ions in pink and Cu2+ ions in green.

spreading a solution of the molecular building units comprising 5,10,15,20-tetrakis(4-carboxyphenyl)porphyrinato-cobalt(i) (CoTCPP; 1) and pyridine (py; 2) in chloroform/methanol onto an aqueous solution of CuCl2·2H2O (3) as a subphase, leading to the formation of a 2D copper-mediated CoTCPP array (CoTCPP-py-Cu). The 2D CoTCPP-py-Cu sheets were then deposited on a Si(100) (or quartz) substrate and stacked by a sequential layer-by-layer growth procedure that included

566 NATURE MATERIALS | VOL 9 | JULY 2010 | w.nature.com/naturematerials

NATUREMATERIALSDOI:10.1038/NMAT2769 LETTERS a b

Number of cyclesWavelength (nm)

Maximum absorbance

Calc. for monolayer 2

9 cycles

Figure 2 | Layer-by-layer film growth of NAFS-1 followed by ultraviolet–visible absorption spectroscopy. a, Ultraviolet–visible absorption spectra of NAFS-1 films on a quartz substrate after successive cycles of sheet deposition, rinsing/solvent immersion and drying. Each spectrum is labelled by the corresponding cycle number (1–9). Absorption spectra of the mixed solution of CoTCPP/py and CuCl2 and the CoTCPP-py array formed on the pure water subphase are shown in Supplementary Fig. S2. b, Plot of maximum absorbance of the CoTCPP Soret band versus the number of film growth cycles. The linear increase in absorbance signifies that each cycle of layer stacking leads to roughly the same amount of deposited material. The dotted horizontal line marks the calculated maximum absorbance for a NAFS-1 monolayer with 100% surface coverage.

intermediate rinsing, solvent (water) immersion and drying steps, thereby resulting in the generation of MOF nanofilms of any desired thickness (NAFS-1: nanofilm of metal–organic frameworks on surfaces no. 1).

Ultraviolet–visible absorption spectroscopy was used to follow the successful layer-by-layer film growth of NAFS-1 on a quartz substrate by monitoring the evolution of the characteristic Soret band of the porphyrin units as a function of the number of sheet deposition and solvent immersion cycles (Fig. 2a). The measured absorbance of the monolayer is in good agreement with the calculated value taking into account the molar absorption coefficient of CoTCPP and the number of molecules present after considering the molecular orientation (the porphyrin 2D plane is parallel to the surface). This indicates that the substrate surface is well covered with the CoTCPP-py-Cu monolayer—the surface coverage is nearly 100%. In addition, the linear increase of the absorbance with increasing number of cycles as shown in Fig. 2b provides unambiguous evidence that, to a good approximation, each step of the layer stacking protocol leads to deposition of the same amount of material. NAFS-1 was also characterized by X-ray photoelectron spectroscopy (XPS) (Supplementary Fig. S3) and infrared spectroscopy (Supplementary Fig. S4).

Detailed insight into the crystallinity, preferred orientation and homogeneity of the fabricated NAFS-1 films was obtained from synchrotron X-ray diffraction (XRD; λ = 1.554Å) carried out in two different scattering geometries—out-of-plane mode, which is sensitive to the lattice parameter in the growth direction, and in-plane(grazingincidence)mode,whichissensitivetothein-plane lattice dimensions. Figure 3a shows the XRD pattern recorded in the out-of-plane geometry for NAFS-1 with 20 stacking cycles fabricated on a Si(100) substrate. Three diffraction peaks are observed revealing the highly oriented nature of the material. They can be indexed as (001), (002) and (003), leading to a value of 0.9380(3)nm for the size of the interlayer spacing (Supplementary Fig. S6). The average crystalline domain size, which corresponds to the total thickness of the 20-layer NAFS-1 is also estimated as about 20nm from the full-width at half-maximum (FWHM) of the observed peaks using Scherrer’s equation. We also carried out complementary X-ray reflectivity measurements on the same film. Simulation of the reflectivity oscillatory pattern allowed evaluation of the total film thickness as 21(1)nm. Both estimates are in excellent agreement with the expected film thickness when we consider the number of layers and the magnitude of the interlayer spacing (0.938nm×20 layers ≈ 19nm; Supplementary Fig. S6). To obtain further structural information on the ordering of the individual layers in NAFS-1, we also carried out rocking curve

(θ-scan) and sample direction dependence (azimuthal angle, φ, scan) measurements at the (001) peak position in the out-of-plane orientation. Observation of a single peak with a small FWHM of ∼0.3◦ in the rocking curve (Fig. 3b) indicates that the individual sheets are tidily stacked in the film and are almost parallel to each other with an average tilting angle of 0.3◦ (Supplementary Fig. S6). Moreover, the absence of azimuthal angle dependence within the surface plane (Fig. 3c) implies that the sheets form uniformlywithoutanypreferenceinthedirectionofthesubstrate.

Figure 4a shows the in-plane XRD pattern for the NAFS-1 sample measured at an incident angle, α = 0.1◦. Seven diffraction peaks are observed up to a scattering angle, 2θ, of 45◦ and index as (110), (200), (320), (400), (330), (440) and (550)/(710) on a metrically tetragonal unit cell with basal plane dimensions, a=b=1.6460(3)nm. An important point to note here is that none of the Bragg reflections observed in grazing-incidence geometry coincides with any of the peaks measured in the out-of-plane XRD pattern and all in-plane XRD peaks index as (hk0)—no (hkl) (l 6= 0) reflections are present. This not only signifies that NAFS-1 is characterized by highly crystalline order in the in-plane orientation but that it also grows in a perfect preferentially oriented directionally controlled manner. The average crystalline domain size is evaluated as ∼18nm from Scherrer’s equation using the FWHM of the intense (110) reflection. Atomic force microscopy (AFM) images of the NAFS-1 film also show domains with a size of around 20nm (Supplementary Fig. S8).

Investigation of the crystalline structure of NAFS-1 was initiated by considering as a starting model the reported crystal structures of bulk porphyrin MOFs built by MTCPP (M = Co3+, Zn2+, Pd2+) units9. These consist of 2D layers of metalloporphyrin linkers and paddle-wheel M2(COO)4 secondary building units that are pillared by 4,4′-bipyridine (bpy) molecules to give different packing arrangements. Our structural model for NAFS-1—shown in projection along different unit-cell axes in Fig. 4c–e—also consists of a similar 2D ‘checkerboard’ motif of CoTCPP units linked by binuclear Cu2(COO)4 paddle-wheels (Supplementary Fig. S9) for each stacked sheet. The comparable lattice size of

NAFS-1 (1.65nm) to those of the bulk systems (∼1.66nm; ref. 9) together with the XPS (Supplementary Fig. S3) and infrared results (Supplementary Fig. S4) demonstrates the formation of the paddlewheel structure. In contrast to the pillaring role played by bpy in linking the layers in the bulk porphyrin MOFs (ref. 9), in the NAFS-1 structural model two monodentate pyridine molecules coordinate axially to both the copper dinuclear block and CoTCPP tocompletethecoordinationsphereandretaintwo-dimensionality. As shown in Fig. 4a, the simulation of the in-plane XRD pattern

NATURE MATERIALS | VOL 9 | JULY 2010 | w.nature.com/naturematerials 567

LETTERS NATUREMATERIALSDOI:10.1038/NMAT2769

Intensity (arb. units)Intensity (arb. units) cb θθ φ 2 at (001) at (001)2 at (001)

Figure 3 | Out-of-plane synchrotron XRD patterns of NAFS-1. a, Observed (filled circles) and fitted (red solid line) out-of-plane synchrotron XRD profiles (θ−2θ-scan, 2θ-step =0.1◦, λ=1.554Å) for a NAFS-1 thin film (20 deposition cycles) on a Si(100) substrate. The inset shows further higher statistics fine scans (blue rhombi, 2θ step =0.025◦ or 0.05◦) collected in the vicinity of the three diffraction peaks (see also Supplementary Fig. S5). b, θ-scan at the (001) peak position. The rocking curve clearly shows a single peak with a FWHM of ∼0.3◦, implying that the individual sheets are tidily stacked almost parallel to each other with an average tilting angle of ∼0.3◦ (Supplementary Fig. S6). c, φ-scan at the (001) peak position. The (001) peak intensity does not change when the sample is rotated from 0◦ (arbitrary starting angle) to 90◦ while keeping the out-of-plane orientation fixed. The insets in a–c show schematic representations of the scattering geometries in each of the XRD experiments.

(calc. 1) (including only (hk0) peaks) is in excellent agreement with experiment for this structural model. The simulated powder XRD pattern of the proposed model (including all (hkl) reflections) is shown for comparison in Supplementary Fig. S10. The calculated in-plane pattern in the absence of any coordinated pyridine molecules bound to the metal ions (calc. 2) is also included in Fig. 4a,b. The clear observation of the (200) Bragg reflection in the experimental profile (Fig. 4b), which appears in calc. 1, but is absent in calc. 2, supports the existence of axially coordinated pyridine moleculesinNAFS-1(seealsoSupplementaryFig.S11).

Stacking of the layers by physical adsorption, which would have also led to poor preferred orientation along the growth direction can be ruled out, as the observed interlayer spacing of 0.94 nm is significantly smaller than the calculated layer thickness (including the axially coordinating pyridine molecules) of ∼1.24nm. Instead, an interlayer stacking order exists in which neighbouring sheets are shifted along the a axis by 1/4 of the unit cell (Fig. 4c–e). This specific interdigitated layer stacking in NAFS-1 can be driven by the pi–pi interactions between pyridine molecules coordinating axially to the paddle-wheels (Fig. 4c–e). Such a layer stacking was taken into account in the simulation of the in-plane XRD pattern (calc. 1). When it is absent, the simulated XRD profile (calc. 3) shows the appearance of an intense (220) peak that is absent from the experimental data (Fig. 4b).

Therefore, our results suggest that the pi–pi interaction plays an important stereoelectronic role in defining the perfect directional control in the growth of NAFS-1 as the metal-coordinated pyridine molecules projected from the 2D sheets allow each further layer to dock in a highly ordered interdigitated manner in the growth of NAFS-1. As there was no control applied in matching the linking sites in the layer-by-layer technique, we suggest that the initialinteractionbetweensuccessiveCoTCPP-py-Cu2Dsheetsjust after deposition may be that of physical adsorption. The key step at which the interlayer pi–pi interactions are introduced should be the solvent immersion process. During the film immersion in the solvent, a structural rearrangement of the nanosize sheets can be initiated, leading to the pi–pi-interaction-stabilized perfectly ordered stacking architecture.

The results presented here show the first successful construction of a crystalline MOF film (named NAFS-1) on a surface with complete structural growth control in both the out-of-plane and in-plane geometries using an unconventional fabrication protocol at room temperature. Moreover, the proposed crystalline packingfortheNAFS-1structure—interdigitated2Dmotifsstacked by pi–pi interactions between parallel aligned pyridine ligands (Fig. 4c–e)—is reminiscent of that in the subclass of bulk MOFs that exhibit ‘gate opening’ transitions4, that is, abrupt structural changes between closed and porous crystalline states induced

568 NATURE MATERIALS | VOL 9 | JULY 2010 | w.nature.com/naturematerials

NATUREMATERIALSDOI:10.1038/NMAT2769 LETTERS

0.938 nm

1.235 nm e

Intensity (arb. units) 0

Intensity (arb. units) Intensity (arb. units)

2 in-plane

NAFS-1 exp.

calc. 1 calc. 2 calc. 3 calc. 1 calc. 2 calc. 3

Cu ¬ interaction a b a b c a c b

(Parte 1 de 2)

Comentários