The porous, three-dimensional (3D) crystalline structure of metal-organic frameworks (MOFs) gives them adsorbent and other properties that have made them useful, for example, for gas storage, gas separation, and catalysts. In particular, multivariate MOFs have either multiple organic functionalities or metal ions (variate units) incorporated into their backbone, allowing for highly selective separations and catalysis, compared to the less functionalized “simple” MOFs. Now Omar M. Yaghi, the inventor of MOFs, and his graduate student Zhe Ji at the University of California, Berkeley have been able to determine the spatial arrangements of the metal ions in one family of multivariate MOFs, MOF-74. Until now, the necessary level of characterization has been elusive, and Yaghi turned to collaborator Tong Li, a professor at Ruhr-Universität Bochum in Germany and an expert in atom probe tomography (APT) in order to achieve this.

The arrangement of the variate units in the MOFs introduces heterogeneity, similar to the well-known sequences of nucleotides in DNA. This results in different spatial arrangements of the variate units covalently bound to the ordered backbone. The researchers explored infinite metal-oxide rods, used in many MOFS, and mixed metal species within these. The metals are known to be arranged in a chain configuration in these rods. Various configurations included sequences of random, short and long duplicates, and insertions.

“The ‘reading’ of sequences in MOFs [has thus far been] largely limited by characterization techniques that can only acquire average structures,” the research team told MRS Bulletin. “Therefore, we employed APT to reveal the MOF sequences in real space, and at the atomic scale.”

In APT, under laser pulsing, atoms are evaporated from the surface of the sample and projected onto a position-sensitive detector. The information from the position-sensitive detector is obtained layer-by layer so that a series of two-dimensional maps are obtained, which are then integrated into a 3D map.

The researchers recognized the potential that the size difference of various metals and the temperature of synthesis can affect the spatial arrangement of the sequences of atoms. The research team synthesized crystals of the well-known MOF-74: Co,Cd-MOF-74 (120°C), Co,Cd-MOF-74 (85°C), Co,Pb-MOF-74 (85°C), and Co,Mn-MOF-74 (85°C), varying the metal combination and synthesis temperature. 

“For each type of metal sequence, we measured three crystals and found that a specific synthesis condition always led to the same type of metal sequence,” the research team told MRS Bulletin. “These results indicate that the chemical effects can introduce biases into the metal sequences during synthesis and provide a practical basis for us to customize a desired sequence type by judiciously choosing the appropriate synthesis parameters.”

As reported in a recent issue of Science, by achieving an APT detection efficiency of 52%, the research team was able to identify four types of metal sequences. With a combination of Co and Cd synthesized at 120°C, the sequence was random, whereas when synthesized at 85°C, the sequence consisted of short duplicates of two to four metals. Long duplicates of more than four metals was achieved with a combination of Co and Pb synthesized at 85°C. With a combination of Co and Mn synthesized at 85°C, the sequence consisted of insertions of a single metal into duplicates of another metal. 

“A specific sequence of open-metal sites in multi-metal MOF-74 can enforce a unique pattern of polarization of guest molecules along their diffusion trajectories, enabling effective separation of target molecules from a mixture,” says the research team. “Our work in characterizing multivariate MOFs provides the first step toward addressing real-world applications using metal sequences.”

The researchers told MRS Bulletin that they believe APT can be applicable to many MOFs for which the spatial arrangement of metals that can be distinguished from the ionized pieces of organics is of interest.

“This work outlined the basic methodology of applying APT to MOFs,” says Qiaowei Li of Fudan University, China, who was not involved in this study. “In the future, it will not be difficult to know what particular sequence encompasses a unique function of a MOF, and to make a structure containing that sequence. For example, specific metal sequences that encode maximized interaction differences toward two guests in a mixture can be incorporated into a MOF, and enhanced gas separation can be achieved.”

The research team says, “We believe the recent development of sample preparation techniques—for example, cryogenic focused ion beam milling—and APT detectors of improved detection efficiency will make this technique more and more useful for MOF samples.” 

Read the abstract in Science.