Using our Ocean system, scientists demonstrate the power of combining multiple electron microscopy methods, showing for the first time how the desilication mechanism of zeolite crystals proceeds in 3D on the single particle level.

Original article by Hanglong Wu, Teng Li, Sai P. Maddala, Zafeiris J. Khalil, Rick R. M. Joosten, Brahim Mezari, Emiel J. M. Hensen, Gijsbertus de With, Heiner Friedrich, Jeroen A. van Bokhoven and Joseph P. Patterson
Thomas article feature image
Liquid phase electron microscopy (LPEM) possesses the potential to revolutionize materials science by providing direct evidence of processes occurring in solution at the nanoscale. However, the use of high-energy electrons can be problematic as it can significantly change the solution chemistry and the liquid sample being imaged. Therefore, a primary challenge in LPEM is understanding the role the electron beam plays in the observed mechanism. As such, researchers find themselves stuck on the same question: “Which features of the native process are being captured?”

Dr. Hanglong Wu and his colleagues from the Eindhoven University of Technology, ETH Zurich and University of California, Irvine provide a robust answer to this pressing question by developing a novel approach for studying complex reactions in solution. This concept, which they term a “distributed electron microscopy (EM) method” combines information from multiple EM methods, including LPEM, cryogenic EM and cryo-electron tomography. To demonstrate this powerful method, Dr. Wu and his colleagues use our dedicated LPEM solution to study the desilication mechanism of zeolite crystals. They show for the first time with an exceptional level of confidence how the reaction proceeds in 3D on the single particle level. Moreover, they show how LPEM can be combined with cryo-EM to study a complete reaction which is essential to rule out beam-induced effects in LPEM experiments. 

The desilication study

Dr. Wu and his colleagues employ a distributed electron microscopy method to study the desilication process in zeolite crystals, showing how the reaction proceeds in 3D and what controls the desilication kinetics. This distributed approach enables them to explore and understand the native mechanism at hand, exploiting the power of each imaging modality – LPEM, cryogenic EM (cryo-EM) and cryo-electron tomography (cryo-ET). The researchers use Al-zoned ZSM-5 zeolites, which have an aluminum-rich rim and an aluminum-poor core, as a model system, to study the desilication process. The reason this is an ideal model system is that ZSM-5 crystals have a complex 3D morphology and are also considerably beam-sensitive.

The two videos below respectively show the automated sample loading process assisted by a liquid handling machine called SciTEM, and how the desilication process was initiated via the manual flowing of an NaOH solution into the cell.

Movie 1: Movie showing the SciTEM-assisted sample loading process, where a tiny droplet (~ 300 pL) containing zeolite crystals was deposited on the centre of the viewing window area.

Movie 2: Movie showing that the liquid cell is gradually filled with 0.6 M NaOH

Using LPTEM to observe the desilication process

Using LPTEM, Dr. Wu and his colleagues monitored the desilication process of individual parent zeolite crystals in multiple areas of a single cell. In the movie below, the desilication process of three pre-etched zeolite crystals can be observed over a 3-hour period. In contrast, for the crystals without the pre-etching treatment, the researchers find that the zeolites were only partially desilicated after 30 h in the liquid cell.

Movie 3: Movie showing the desilication process of three pre-etched zeolite crystals. On the left is the raw data movie and on the right a 30-frame-averaged data movie

The LPTEM data collected by the researchers provide evidence that desilication propagates along the c-axis beginning either at the Al-rich/Si-rich interface or in the crystal center. Moreover, it is observed that for the intergrowth structures, the process begins at the intergrown interface. Importantly, the LPEM experiments show that the confinement in the liquid cell slows down the desilication kinetics and that the electron beam does in fact influence the morphological evolution. This then raises the question that Dr. Wu and his colleagues sought out to answer: which features of the native desilication process are actually being captured?


The LPSTEM results (Figure 1) observed by the researchers were significantly different from the LPEM observations. Dr. Wu and his colleagues believe that the differences mainly originate from contrast formation mechanisms and the localized electron distribution of each imaging mode. On one hand, LPTEM enables the quick control of the localized low electron flux homogeneously across the field of view. However, the appearance of diffraction contrast from the defects can be problematic for quantitative contrast interpretations. On the other hand, LPSTEM enables the better understanding of sample contrast evolution due to being less affected by diffraction effects. Simultaneously, however, the focused high flux electron probe might bring localized structural changes caused by radiolysis of the sample itself and water.

Figure 1 Hanglong article11

Figure 1: Desilication process of Al-zoned ZSM-5 crystals imaged by pulsed LP-STEM. Scale bars: 200 nm.

The LPSTEM results provide evidence that desilication proceeds by a sequential two-step process. Because the STEM probe scanning accelerates the etching process, this enabled the desilication process to be visualized with apparently “native” kinetics, thereby overcoming the confinement effects in the liquid cell. However, as beam effects were observed at all applied electron doses, the question remains: which features of the native desilication process are being captured?

Using cryo-EM and cryo-ET to observe the desilication process

Monitoring the desilication process using cryo-EM was the next step for Dr. Wu and his colleagues. Unlike LPEM, cryo-EM enabled the researchers to monitor the process in the absence of confinement and electron beam effects. In practice, cryo-TEM is generally performed prior to an in situ study, as it provides key information on what intermediate structures are present and, most importantly, how they evolve statistically over time. The figure below shows the cryo-EM results presented for crystals at different time points, where the stages of desilication are in agreement with the collective LPEM data and the general sample evolution observed by cryo-EM.

Figure 2 Hanglong article

Figure 2: Desilication process of ZSM-5 zeolites monitored by cryo-TEM (a) and cryo-STEM (b). Scale bars: 200 nm.

To verify the 3D interpretation of the desilication process, cryo-ET was performed on four crystals at an intermediate state of desilication. The movie below shows the cryo-ET reconstruction of a 12-hour desilicated zeolite crystal and Figure 3 shows the corresponding segmented surface rendering of the crystal.
Figure 3 Hanglong article

Figure 3: Segmented surface rendering of the 12 h desilicated zeolite crystal.

Movie 4: Movie showing the cryo-electron tomography 3D reconstruction of a 12 h desilicated zeolite crystal

Collectively, with the cryo-EM results, Dr. Wu and his colleagues were finally able to distinguish which information provided by LPEM observations represents the native desilication and which is an artifact. The cryo-EM data supported the hypothesis from both LPTEM and LPSTEM that desilication propagates along the c-axis. Moreover, it confirmed that it should occur in a two-step process, where both processes begin at the Al-rich/Si-rich interface and propagate toward the center of the crystal. This suggests that the LPTEM data either represents an artifact caused by the electron beam, confinement within the liquid cell, the sample preparation, or that the intermediates of this process are too transient to be captured by the cryo-EM data.

Novelty in findings

This study is a major step forward in understanding the role that the electron beam plays in LPEM experiments. Ultimately the collective results reveal that the desilication mechanism of zeolite crystals occurs via a two-step anisotropic etching process. This complex mechanism was only discovered due to the novel distributed approach that Dr. Wu and his colleagues employed. Particularly, the combination of LPEM and cryo-EM was essential to rule out the effects of the electron beam in the LPEM experiments.

Considering the strength of this distributed method, Dr. Wu and his colleagues anticipate that the combination of LPEM and cryo-EM will become a standard approach for understanding reaction mechanisms in aqueous solutions and could even be extended to non-aqueous solutions. Moreover this method will enable scientists to draw robust conclusions on the 3D reaction mechanisms of other important beam-sensitive material systems, such as biomaterials and proteins, whereby some grand challenges in materials science and life sciences could be potentially solved.

Hanglong Wu portrait
“DENSsolutions offers a simple but reliable solution for us to look into materials processes in liquid at the nanoscale. Currently, it is still difficult to observe the same phenomena in different liquid cells. That’s why people always joke that each LPEM experiment is “unique”. However, using the Ocean system, we have almost made every single “unique” LPEM experiment reproducible.”
Dr. Hanglong Wu
Postdoctoral Researcher |  Eindhoven University of Technology

Original article:

Discover our LPEM solutions:

Discover more publications made possible by our LPEM solutions:

Thomas article feature image

Scientists present novel approaches to map and control liquid thickness in LPEM experiments

Using our Ocean system, scientists present new approaches to overcome the limitations in LPEM experiments by quantitatively mapping and controlling liquid layer thickness.

Subscribe to our newsletter to stay up-to-date with the latest in situ microscopy news.