A microscopy technique has been used to study the formation and growth of crystals of porous solids known as metal–organic frameworks in real time. The findings will aid the design of methods for making these useful compounds.
Materials called metal–organic frameworks (MOFs) have sparked intense interest over the past few decades. In particular, those that form permanently porous architectures have tremendous potential for applications such as chemical sensing, gas storage and catalysis. But techniques for synthesizing these compounds are still often developed through trial and error — in part because the mechanisms that dictate the self-assembly of MOF unit cells from their constituent metal ions and ligands, and their subsequent growth into nanoparticles, are largely unknown and difficult to observe. Writing in the Journal of the American Chemical Society, Patterson et al.1 help to solve this problem by reporting the first observations of the crystallization of MOFs made in real time, using a technique called liquid-cell transmission electron microscopy.
Previous studies of MOF crystallization have made use of various ex situ and in situ analytical techniques, including high-resolution transmission electron microscopy (HRTEM) and energy-dispersive X-ray diffraction (EDXRD). For example, HRTEM has been used to examine the crystallization of the well-characterized MOF-5, by analysing samples taken at various time intervals early in the compound's synthesis2. Time-resolved in situ EDXRD has been used to determine the kinetics of MOF crystallization as a function of parameters that included pH, temperature and ligand length3,4. But the challenge of observing the crystallization process in real time persists.
Patterson and colleagues' use of liquid-cell transmission electron microscopy (LCTEM) is a big step forward. This technique allows dynamic processes that occur in liquids to be imaged as they happen. It has been used to observe systems such as biological structures5 and growing nanocrystals6, but had not previously been applied to MOF syntheses.
Because analytical samples can be damaged by the electron beam used in LCTEM, the authors began by performing a series of control experiments using a zirconium-based MOF called UiO-66, to decouple the effects of beam irradiation on MOF synthesis from the effects of the reaction mechanism. UiO-66 was a good choice for a control because it is easy to synthesize and extremely stable, which meant that it could be prepared ahead of the LCTEM experiments without any risk of it degrading before use. The authors observed that the dissolution or growth of UiO-66 particles depends on the voltage of the electron beam. A threshold dosage of 40,000 electrons per square nanometre was also established — as long as experiments were performed below this limit, damage and particle motion during crystallization were negligible.
Patterson et al. then chose another MOF, ZIF-8, as the ideal candidate for demonstrating the LCTEM method. The growth mechanisms of ZIF-8 have previously been studied7 using transmission electron microscopy on samples removed from synthetic solutions of the MOF, which provided a good comparison with the LCTEM results. ZIF-8 can be synthesized at room temperature in methanol, using zinc nitrate as the metal source and 2-methylimidazole molecules as the organic ligands. The authors observed the growth of ZIF-8 in real time over 11 minutes — the first particles detected were 15 nm in diameter, with subsequent growth observed up to 50 nm.
The authors went on to record videos of the particle formation and used them to determine the growth kinetics of the MOF, uncovering several important features of the crystallization process. First, they proved by direct observation that ZIF-8 particles form through the growth of smaller subunits, rather than by particles coalescing. They also found that an excess of ligand molecules leads to the formation of ZIF-8 particles that are smaller than those formed when the metal-to-ligand ratio is 1:1. The researchers had predicted this using ex situ methods, but LCTEM enabled them to observe the process as it occurred.
A series of careful growth experiments was then performed under various accumulative electron doses. The results convincingly show that LCTEM can be applied effectively to study nanoparticles that are easily damaged by electron beams. It remains to be seen whether the technique will be effective at the higher temperatures (typically greater than 100 °C) at which most MOFs form.
A general conclusion from this work is that MOF growth occurs through the transport of metals and ligands to a nascent particle, followed by their movement to an edge or surface site, where bonding between the metal and ligand finally occurs (Fig. 1). The attachment of metal–ligand monomers to a surface site is therefore the controlling factor in particle growth, and the process is not diffusion-limited.
The development of this ability to watch particle formation in situ during MOF self-assembly should enable a variety of complicated synthetic questions to be answered. One example is how the addition of 'modulator' compounds, which are sometimes used in MOF syntheses to control the crystallinity of the products, affects the growth kinetics. For MOFs that can adopt different structures, LCTEM could also shed light on what dictates whether kinetic products — those that crystallize most quickly — form during reactions, rather than the most thermodynamically stable products. LCTEM is a much-needed addition to the MOF-characterization toolkit, and its use in conjunction with other methods will no doubt lead to the specific control of crystal morphology, compositions and defects.