Research Article www.acsami.org
Large-Scale Synthesis of Monodisperse UiO-66 Crystals with Tunable Sizes and Missing Linker Defects via Acid/Base Co-Modulation Yajing Zhao, Qing Zhang, Yali Li, Rui Zhang, and Guang Lu* Institute of Functional Nano & Soft Materials (FUNSOM), Jiangsu Key Laboratory for Carbon-Based Functional Materials & Devices, Soochow University, Suzhou 215123, China S Supporting Information *
ABSTRACT: Beyond their pore structures and surface chemistry, precise controls over other attributes of metal−organic frameworks (MOFs) such as shapes, sizes, and defects are also favorable to their fundamental studies and applications but still remain challenging. Herein, we reported an acid/base comodulation strategy to the large-scale synthesis of monodisperse UiO-66 crystals with acetic acid for modulating crystal shape and with triethylamine (TEA) as a base for controlling the nucleation of crystallization and tuning the formation of missing linker defects via promoting presumably the singe deprotonation of terephthalic acid linkers. The obtained monodisperse MOF crystals have a well-defined octahedral shape, tunable sizes ranging from ∼500 nm to ∼2 μm, and high thermal stability. Their assembled-monolayers are responsive to methanol vapor with the crystal size-dependent and defectrelevant sensing performances. KEYWORDS: metal−organic framework, UiO-66, single crystals, microporous materials, modulation coordination synthesis
■
INTRODUCTION Metal−organic frameworks (MOFs) are a class of porous crystalline materials composed of metal ions (or clusters) linked by organic ligands.1−3 Due to their large surface areas, regular pore sizes, and tailorable chemistry, these materials have shown great promise for a variety of applications.4−18 Beyond the pore and chemistry nature, many other attributes of materials such as crystal shapes, sizes, and defects are also gaining increasing concerns along with the rapid development in the field of MOFs.19−25 The effective controls over these attributes facilitate the fundamental understanding of crystallization processes,26−34 study of sorption kinetics,35,36 tuning of properties,37,38 enhancement of performances,39−41 and extension of applications of MOF materials.19 Monodisperse MOF crystals with uniform shapes and sizes are also promising building blocks for constructing oriented thin films and devices with one-dimensional (1D), two-dimensional (2D), and threedimensional (3D) ordered mesostructures.42−46 Nevertheless, reports regarding the synthesis of monodisperse MOF crystals with uniform shapes but tunable sizes are comparatively few.36 Just as other crystal materials, various types of defects do exist in MOFs. On the one hand, defects can be engineered, which provides an alternate powerful manner to manipulate the chemical and physical properties of MOF materials.25 On the other hand, admittedly, defects of some types (for example, the missing linker defects) within MOFs will lead to the decreased thermal and mechanical stabilities of materials.47−49 Intensive efforts have been made recently to create different types of defects in MOFs, but there are few reports on how to reduce defects.47 Finally, the reproducible and large-scale synthesis of © XXXX American Chemical Society
MOF crystals with controlled attributes is also critical to their applications. UiO-66, a zirconium-carboxylate MOF (Zr6O4(OH)4(BDC)6, BDC = terephthalic acid), has been intensively studied recently due to its excellent chemical, mechanical, and thermal stabilities since it was first reported by Lillerud et al.50 After the work reported by Behrens et al.,51 the modulation coordination strategy has been extensively adopted for the synthesis of UiO-66.52 With different monodentate acids such as benzoic acid, acetic acid, and formic acid as modulators, crystals with shapes of octahedron, cube, and cuboctahedron and with sizes ranging from 14 nm to 300 μm have been successfully prepared, respectively.51,53,54 In previous work, we also reported an acetic acid-modulated synthesis of monodisperse UiO-66 crystals and their self-assembly.46 However, it is limited to the small scale reactions and is incapable of tuning crystal sizes. In addition, it has also been noted recently that the addition of acid modulators in the synthesis promotes dramatically the formation of missing linker defects in UiO66 crystals.53,55−59 This finding has been successfully exploited to increase the surface areas,60 enlarge the pore sizes,61 and enhance the catalytic activity62 and gas sorption56,63 of UiO-66 materials. However, the increase in missing linker defects in UiO-66 crystals commonly results in the decrease of their thermal and mechanical stabilities.47−49 Although Lillerud et al. have reported the synthesis of “ideal” UiO-66 with high thermal Received: February 27, 2017 Accepted: April 11, 2017
A
DOI: 10.1021/acsami.7b02887 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
ACS Applied Materials & Interfaces
■
stability,47 the product consists of intergrown crystals with a cuboctahedral morphology.58 Here, we reported an acid/base co-modulation strategy that is capable of synthesizing monodisperse UiO-66 octahedral crystals with tunable sizes and in large scale. In this strategy, acetic acid was used to modulate the shape of crystals and triethylamine (TEA) served as the co-modulator to control the nucleation of crystals and thus their final sizes (Scheme 1).
RESULTS AND DISCUSSION To gain a better understanding of how the reaction scale affects the size distribution of UiO-66 crystals synthesized with acetic acid as the sole modulator, we first visually monitored reactions (with 4 mM ZrCl4, 4 mM BDC, and 2.4 M acetic acid in N,Ndimethylformamide (DMF)) occurring in a 10 mL and a 150 mL scales, respectively. When the clear precursor solutions were heated at 120 °C, visible turbidness was observed after ∼40 min for the 10 mL reaction but ∼4 h for the 150 mL reaction, indicating the slower formation (and nucleation) of UiO-66 crystals in the larger scale reaction. Scanning electron microscopy (SEM) measurements (Figure S1) reveal that the 10 mL reaction yields monodisperse crystals with an average size of ∼600 nm but the 150 mL reaction does not, although all crystals display a well-defined octahedral shape. These two observations suggest that the broad size distribution of crystals observed in the large scale reaction is presumably a result of the longer nucleation period for the crystallization of UiO-66 where secondary nucleation occurs inevitably. It was known that the formation (nucleation and growth) of UiO-66 is based on the coordination of Zr4+ ions with BDC linkers, which, however, is initiated via deprotonating linkers by the organic base, dimethylamine, gradually formed during the reaction due to the thermal decomposition of solvent, DMF.64 Given the relative slower heat transfer and thus slower decomposition of DMF in the large scale reaction, it is reasonable that a prolonged nucleation (and growth) period of UiO-66 was observed therein. To shorten the nucleation period of crystallization and yield crystals with narrow size distribution, we added TEA to the acetic acid-modulated synthesis solutions to increase the concentration of base in the early stage of a large scale reaction. We also anticipated that an effective control over the sizes of UiO-66 crystals could be achieved further by tuning TEA concentrations. The process of the acid/base co-modulated synthesis of UiO66 was detailed in the Experimental Section. The synthesis was commonly carried out in 150 mL scale, where the concentrations of ZrCl4 (4 mM), BDC (4 mM), and acetic acid (2.4 M) were kept constant and the concentrations of TEA were systematically varied, but it was also applicable to a larger scale (for example, 300 mL, see Figure S2). After 6 h of reaction, similar yields (62−65%) were observed for all the large-scale syntheses. SEM measurements reveal that reactions in the presence of 0.5−8 mM TEA produce well-defined octahedral crystals (Figure 1a−f) whereas higher TEA concentrations result in the formation of intergrown crystals (Figure S3). These octahedral crystal samples display narrow size distributions with small values (0.038−0.046) for coefficient of variation (CV) (Figure S4). Quantitative data extracted from samples obtained in four independent batches of reactions with the same experimental parameters establishes the high reproducibility (with batch-to-batch CV values of 0.033− 0.079) of current synthesis approach (Table S1). The statistical analysis also reveals a strong dependence of crystal sizes on the concentrations of TEA used in the synthesis (Figure 1g). The crystal sizes decrease noticeably from 1949 to 583 nm as TEA concentrations increase gradually from 0.5 mM to 4 mM but do slightly to 502 nm as further enhancing the base concentrations. Although the octahedral shape of the obtained crystals has manifested the participation of acetic acid in the MOF formation reactions via competitive coordination with Zr4+, the existence of acetate in all digested samples was also
Scheme 1. Illustration of the Proposed Acetic AcidModulated and Acetic Acid/Triethylamine (TEA) Comodulated Syntheses of UiO-66 Crystals in Large Scale
Monodisperse UiO-66 crystals with sizes ranging from ∼500 nm to ∼2 μm can be reproducibly synthesized with reaction scale up to 300 mL. To the best of our knowledge, this is the first time that organic bases were used for the controlled synthesis of UiO-66. Our results indicate that bases, indeed, play important roles in the crystallization of UiO-66 and the formation of missing linker defects via promoting probably the formation of singly deprotonated linkers. Compared to product synthesized with only acetic acid as modulator, the monodisperse crystals obtained in the presence of TEA contain fewer missing linker defects and display higher thermal stability. These uniform MOF crystals can be easily assembled to construct oriented monolayer-based sensors with size-dependent and defect-relevant sensing performances for methanol vapor, where missing linker defects in UiO-66 show more obvious effect on the desorption of molecules than their adsorption.
■
Research Article
EXPERIMENTAL SECTION
General Procedure for the Acetic Acid/TEA Co-modulated Synthesis of UiO-66. The final concentrations of BDC (4 mM), ZrCl4 (4 mM), and acetic acid (2.4 M) were kept constant in all syntheses whereas that of TEA was varied from 0 to 8 mM. Typically, BDC (0.6 mmol) and TEA (0−1.2 mmol) were charged in a 250 mL round-bottle flask containing 140 mL DMF and allowed to reaction at room temperature under stirring (600 rpm) for 10 min before acetic acid (0.36 mol) was added. Subsequently, the above mixed solution was heated to 120 °C in an oil bath and a 10 mL ZrCl4 (0.6 mmol) solution in DMF were added, and the mixed solution was allowed to react without stirring at this temperature for 6 h. After the solution was cooled to room temperature, the product was collected by centrifugation at 8000 rpm for 5 min, washed three times with DMF and three times with methanol, and then soaked in methanol for 3 days with replacing the soaking solvent every 12 h to exchange DMF. Other experiment and characterization details are provided in the Supporting Information (SI). B
DOI: 10.1021/acsami.7b02887 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
based on the pKa values in aqueous solutions due to the lack of data for terephthalic acid in DMF. Nevertheless, considering the possible similar change trends from aqueous to DMF solutions observed for the pKa values of acetic acid (from 4.76 to 13.5) and o-phthalic acid (from 2.89 to 6.7 and from 5.11 to 16.5), an isomer of terephthalic acid,65 it might be instructive to some extent to discuss the deprotonation equilibria involved with terephthalic acid in DMF solutions with the pKa values in aqueous solutions as reported previously.30,58 In our experiment, terephthalic acid (H2BDC) was first allowed to react with TEA for 10 min before the addition of acetic acid. Given the difference between pKa1 (3.51) and pKa2 (4.82) for H2BDC linkers (4 mM), HBDC− is expected to be the predominated deprotonated form and its concentration increases as TEA concentrations increase from 0 to 4 mM. More TEA will decrease the concentration of HBDC− and increase that of the doubly deprotonated species (BDC2−). However, the subsequent addition of acetic acid (pKa = 4.76) decreases the concentration of BDC2− to yield HBDC− again and acetate. Accordingly, the smallest molar ratio of acetate/BDC observed for the 4 mM sample is presumably a result of the maximum concentration of HBDC− and minimum concentration of acetate in the initial stage of its reaction. This speculation also suggests that the rapid formation of HBDC− in the presence of TEA might be responsible for the promoted nucleation process of crystallization and its concentration also influences the nucleation density, resulting in the formation of monodisperse crystals with tunable sizes. The noticeable decrease in crystal sizes in the TEA concentration range of 0.5−4 mM is caused by the gradual increase in HBDC− concentration which, however, does not change obviously as TEA concentrations further increase in the presence of acetic acid, resulting in the slight change of crystal sizes correspondingly. The crystallinity of samples was studied by powder X-ray diffraction (XRD) measurements (Figure S6). All samples display similar diffraction patterns which are identical to that of UiO-66. Thermal stability of samples in air was investigated by thermogravimetric analyses (TGA) (Figures 3a and S7−S13). The TGA traces (normalized with final weight to 100%) for all samples show obvious weight losses in the ranges of 25−200 °C, 200−390 °C, and 390−600 °C due to the volatilization of solvents (water, methanol, and DMF), the dehydroxylation and the elimination of monodentate modulators, and the decomposition of BDC linkers in framework structures, respectively.55,58 Samples synthesized in the presence of TEA display higher thermal stability (with rapid weight losses at 524−529 °C) than that obtained in the absence of TEA does (with rapid weight loss at ∼514 °C). Following the method reported by Lillerud et al.,58 the TGA data was further analyzed to evaluate the deficiencies of BDC linkers in samples. The weight at 390 °C (W390) and the final weight (100%) were assigned to the dehydroxylated- and modulator-eliminated-Zr6 formula unit, Zr6O6+x(BDC)6−x (where x is the deficient value of BDC linkers per Zr6 formula unit), and its decomposition product, 6ZrO2, respectively. For the ideal UiO-66, x is zero and thus a theoretical W390 of its dehydroxylated formula unit, Zr6O6(BDC)6, is 220.2%. The observed W390 values for all synthesized samples are smaller than 220.2%, indicating that there are missing linker defects present. As shown in Figure 3b and summarized in Table S3, samples synthesized in the presence of TEA have smaller calculated x values than that synthesized with acetic acid alone does. However, the x values do not decrease monotonously as TEA concentrations increase.
Figure 1. SEM images of UiO-66 crystals synthesized in the presence of 4 mM ZrCl4, 4 mM BDC, 2.4 M acetic acid, and TEA with concentrations of a) 0.5, b) 1, c) 2, d) 4, e) 6, and f) 8 mM, respectively. The average crystal sizes are 2140, 1134, 742, 554, 529, and 512 nm for samples shown in a−f), respectively. g) Statistical average crystal sizes of samples obtained from four independent batches of reactions with the same experimental parameters versus TEA concentrations.
confirmed by proton nuclear magnetic resonance (1H NMR) measurements (Figure 5S). As TEA concentrations vary gradually from 0 to 8 mM, the calculated molar ratios of acetate/BDC for samples initially decrease monotonously, then increase, and display the minimum value for the 4 mM sample (Figure 2 and Table S2). To gain an insight into the roles of TEA in reaction, the deprotonation equilibria were taken into account. It should be noted that the following discussion is
Figure 2. Molar ratios of acetate/BDC determined by 1H NMR measurements of the digested samples versus TEA concentrations. C
DOI: 10.1021/acsami.7b02887 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 4. a) Nitrogen sorption isotherms of samples synthesized in the presence of TEA with concentrations of 0, 0.5, 1, 2, 4, 6, and 8 mM, respectively, at 77 K up to 1 bar. b) The calculated BET surface areas of samples versus the concentrations of TEA.
Figure 3. a) TGA traces (solid) and the corresponding first derivative curves (dashed) of samples synthesized in the presence of 0 and 0.5 mM TEA. b) The calculated deficient values of BDC linkers per Zr6 formula unit of samples versus the concentrations of TEA.
areas of UiO-66 samples, which is consistent with previous reports.58 The monodisperse UiO-66 crystals were also used to produce 2D oriented monolayers on silicon substrates with the aid of liquid−air interfacial assembly technique (Figure S16).46 These oriented MOF thin films can serve as potential sensors for chemical gases and vapors.9,11,13 As shown in Figure 5a,b, the large-crystal film shows more interference peaks with narrower widths in its reflectance spectrum than the smallcrystal one does. Exposure to methanol vapor results in a red-
Instead, a minimum value of 0.15 was observed for the 4 mM sample. As discussed above, the increase in the concentration of HBDC− as TEA concentrations change from 0 to 4 mM might be favorable to decreasing the missing linker defects which, however, increase, in turn, with higher TEA concentrations due to the formation of more acetate. Comparison between the deficient values of BDC linkers derived from TGA measurements and molar ratios of acetate/BDC determined by 1H NMR results also indicate that there are other negatively charged species besides acetate that counterbalance the positive charges on Zr6 clusters with missing linkers. Among the possible species suggested by previous studies,47,53,55,58,59 formate (originating from the thermal decomposition of DMF) and chloride were ruled out based on the NMR measurements (Figure S6) and the energy dispersive X-ray microanalysis (EDX) results (Figure S14), respectively. Thus, oxide or/and water/hydroxide were speculated to compensate for the residual defect sites in Zr6 clusters with missing linkers. The porosity of activated samples was investigated by nitrogen sorption measurements. As shown in Figure 4a, all samples display the type I isotherms with rapid increases in nitrogen uptake at low relative pressure (
