Big insights from tiny crystals

Most compounds form crystals so small that scientists cannot experimentally determine their atomic structures using X-ray crystallography. Microcrystal electron diffraction now provides a unique solution for this challenge.

Modern chemistry routinely requires the investigator to obtain atomic resolution structures of their molecules of interest. These are used to validate preliminary theoretical models, as well as to explain the molecule’s chemical and physical properties — information that is invaluable to a diverse range of fields, such as drug design and the development of new materials. Several methods have been developed over the past century to facilitate this, with nuclear magnetic resonance (NMR) and X-ray crystallography gaining particular prominence.

NMR currently stands as the primary means by which structures of small molecules are routinely elucidated in chemistry laboratories. It offers the advantage of requiring little material for analysis, is rapid to set up and the sample can be recovered after the spectra have been acquired. One great drawback, however, is that NMR only provides limited information to assign the relative stereochemistry of the molecule of interest and to ultimately determine its absolute configuration. This is where X-ray crystallography comes into play. Given the frequently achieved sub-Ångstrom resolution of X-ray crystal structures, this method has earned the reputation of being the gold standard for small-molecule structure determination. The main disadvantage of X-ray crystallography is implied in its name: it requires the molecules of interest to form relatively large crystals. Many compounds are either incapable of this or require a significant time investment into optimization of crystal growth, which prevents X-ray crystallography from becoming a routine high-throughput method in the lab. However, if electrons are used instead of X-rays, significantly smaller crystals are able to provide a diffraction pattern from which the structure can be reconstructed, as electrons interact with crystal matter more strongly than X-rays. Two recent articles, one published in Angewandte Chemie International Edition1 and the other in ACS Central Science2, make clever use of this particular property of electrons to determine small-molecule structures from extremely tiny crystals (Fig. 1).

The beauty of the methodology developed by these studies is that a significant portion of compounds exist naturally in a microcrystalline state3, rendering a separate crystal growth step unnecessary, and directly allowing rapid structure determination after purification. In the first paper, a team led by Tim Gruene started by making a customized electron diffractometer by combining a transmission electron microscope with a hybrid pixel detector typically used for imaging X-rays1. The benefit of this approach compared to using conventional complementary metal–oxide conductor detectors is that the data collection speed, dynamic range and signal-to-noise ratio are greatly increased — allowing lower, less damaging electron doses to be used. This is a potentially critical advantage for many radiation-sensitive nanocrystals, especially for very small ones. The team then proceeded by directly analysing the powder blend from a pill of the cold medicine Grippostad, which contains both active and inactive components. The active ingredient paracetamol is present in microcrystalline form, whereas the other compounds are amorphous. The paracetamol microcrystals, far too small to diffract X-rays, produced an electron diffraction pattern to a resolution of 0.81 Å, which enabled determination of the polymorph of the compound and use of automated atom assignment when solving the structure. Due to the fact that electrons interact more strongly with matter than X-rays do, some of the hydrogen atoms of paracetamol could also be observed, which is a rare occurrence in crystal structures. Gruene and co-workers then used the same technique on a new methylene blue derivative that generally produces crystals that are unfavorable for X-ray diffraction studies. The structure was solved within four hours of sample preparation and its validity was later confirmed by X-ray crystallography using larger crystals of the same compound. The team thereby successfully demonstrated that electron diffraction can be used to extract structural information from crystals that are too small to be analysed using X-rays, even in the presence of non-crystalline accompanying material. Their future research will focus on the development of hardware and refinement algorithms for electron diffraction data.

In the second paper, Christopher G. Jones and Michael W. Martynowycz et al. performed a similar investigation that demonstrates many interesting findings and techniques complementary to the first study2. Whereas the investigations of Gruene et al. use a custom hardware setup with its own advantages, Jones, Martynowycz and co-workers show that a transmission electron microscopy setup currently used worldwide can be implemented to obtain electron diffraction data directly from powders, thereby making the method immediately accessible to a broad audience. Here, the authors used cryogenic temperatures for their experiments, which reduces radiation damage to the targeted crystals4, allowing more complete datasets to be collected — thereby improving the accuracy of the reconstructed structures. The team started by applying seemingly amorphous progesterone powder onto an electron microscopy grid and then collected data from one of the thousands of available nanometer to micrometer size crystals. According to the study, the entire duration of the process took less than three minutes. Soon after a file conversion step, the output was ready to be processed in popular X-ray crystallography software packages, yielding a 1 Å resolution structure just thirty minutes after taking the sample out of the reagent bottle. Jones, Martynowycz and co-workers were “astounded by the ease with which such high-quality data were obtained”, and promptly proceeded to directly test various powders from their chemical cabinet at the microscope. This resulted in obtaining eleven structures, ten of which were achieved at a resolution of 1.1 Å or better. Similar to the study by Gruene et al., hydrogen atoms, normally invisible in X-ray crystallography, could be observed in the structures.

Eager to obtain further results, Jones, Martynowycz and co-workers set out to discover if this method is adequate for the study of heterogeneous mixtures — a task for which NMR and X-ray crystallography are poorly suited. After crushing a mixture of four compounds together and depositing them on an electron microscopy grid, individual nanocrystals were targeted, and within minutes their identity was confirmed by unit cell parameters and high-resolution structures were obtained for all components of the mixture. They also wanted to ensure that the recrystallization procedure potentially used for purification of the tested compounds during their production did not affect the study by favouring their existence in nanocrystalline form. In order to do so, they purified four new compounds using the popular combination of silica gel chromatography followed by rotary evaporation, and then applied the compound residue to the microscopy grid. The resulting two structures at 1 Å resolution proved that it is possible to perform routine structure determination of many purified compounds without resorting to any dedicated (re)crystallization steps. This tour de force electron diffraction study represents an important step forward by illustrating how high-quality small-molecule structures can be rapidly obtained from powders directly after purification.

Fig. 1: Setup showing how electron diffraction can be used to solve the atomic structure of microcrystalline powder components.

Many powders initially appear amorphous but high magnification shows they are actually micro- or nanocrystalline. Placing nanocrystals on an electron microscopy grid enables an electron diffraction pattern to be obtained using a transmission electron microscope (TEM). This diffraction pattern can then be processed using specialized software in order to reconstruct an atomic resolution model of the compound’s structure.

Neither study is the first of its kind to use electron diffraction for structure determination. This method has been long used in 2D crystallography5, later pioneered for examining 3D crystals by researchers investigating small molecules6,7 and macromolecules8. What distinguishes the results of these two reports is how they demonstrate the achievability of obtaining sub-Ångstrom structures directly from seemingly amorphous powders (that are in fact nanocrystalline) without any extra crystallization efforts. Even more remarkably, they reveal the possibility of obtaining structures of purified compounds directly after solvent removal using only tiny amounts of sample, in combination with truly amorphous components and in complex nanocrystalline mixtures. By all means, this method is not without potential limitations as the solvent that may be necessary to maintain the crystalline state might not be compatible with use in a (cryo)electron microscope. Also, the question still remains if it is feasible to obtain the absolute configuration of target molecules with electrons as is possible with X-rays, where a phenomenon called anomalous dispersion is used for this purpose. Nonetheless, it is clear that the prospect of obtaining structures of small molecules and determining their polymorphs directly after purification has the potential to greatly accelerate and simplify the workflow of researchers working in the fields of organic, inorganic, analytical, physical and medicinal chemistry, leading to new breakthroughs in areas ranging from drug design to nanotechnology.


  1. 1.

    Gruene, T. et al. Angew. Chem. Int. Ed. 57, 16313–16317 (2018).

    CAS  Article  Google Scholar 

  2. 2.

    Jones, C. G. et al. ACS Cent. Sci. 4, 1587–1592 (2018).

    CAS  Article  Google Scholar 

  3. 3.

    Wagner, T. High through-put chemical structure determination. In SHELX Symposium (University of Göttingen, 2011).

  4. 4.

    Glaeser, R. M. & Taylor, K. A. J. Microsc. 112, 127–138 (1978).

    CAS  Article  Google Scholar 

  5. 5.

    Longley, W. J. Mol. Biol. 30, 323–327 (1967).

    CAS  Article  Google Scholar 

  6. 6.

    Kolb, U., Gorelik, T., Kübel, C., Otten, M. T. & Hubert, D. Ultramicroscopy 107, 507–513 (2007).

    CAS  Article  Google Scholar 

  7. 7.

    Wang, Y. et al. Chem. Commun. 53, 7018–7021 (2017).

    CAS  Article  Google Scholar 

  8. 8.

    Shi, D., Nannenga, B. L., Iadanza, M. G. & Gonen, T. eLife 2, e01345 (2013).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Stefan Raunser.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Sitsel, O., Raunser, S. Big insights from tiny crystals. Nature Chem 11, 106–108 (2019).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing