The paper in Nature Communications is available here: https://www.nature.com/articles/s41467-018-04296-4
Those working in the area of zeolite synthesis can attest that the mechanisms of crystallization are a black box that renders the ability to tailor growth more of an art than a science. People have lightheartedly questioned my sanity for tackling such complex systems, and to some extent their sentiment is merited given that zeolites, analogous to many other materials that grow via nonclassical pathways, are difficult to characterize in situ. For many crystalline materials, mechanistic insight is often limited to the interpretation of ex situ electron micrographs and/or other indirect techniques, which has given birth to the axiom “morphology does not equal mechanism” – meaning that restrain should be exercised when interpreting mechanistic details from ex situ data. In the book “The Discovers” written by historian Daniel J. Boorstin, he asserts that the greatest obstacle to discovery is not ignorance, but rather the illusion of knowledge. To this end, intuition-based approaches can be used as substitutes for direct evidence when formulating theories and/or rationalizing new discoveries. Taking Boorstin’s advice, we have worked to develop in situ methods to close knowledge gaps in zeolite crystal engineering by tracking surface growth at unparalleled spatiotemporal resolution; and while our findings have not completely illuminated the black box, we have shown that the complex mechanisms of zeolite crystallization involve multifaceted pathways that often diverge from conventional (or intuitive) conclusions that would be reasonably anticipated on the basis of classical theories.
The development of solvothermal atomic force microscopy (AFM) for in situ imaging at temperatures higher than conventional AFM instruments was one of the first projects I started once beginning my tenure-track position in the Fall of 2009. As a postdoc, I had worked extensively with AFM to track the dynamics of crystals that grow rapidly at room temperature; however, my goal was to use this same technique to track the growth of zeolites in situ, which requires high temperature and long imaging time. I was convinced that AFM instruments could operate under such conditions, although no liquid sample cells capable of exceeding 60°C were available commercially at the time. As such, I worked with Asylum Research to develop a liquid sample cell that could operate at higher temperature, thus beginning a collaboration that took nearly five years to design, trouble shoot, and optimize the system to pave the way for the first publication in 2014 on zeolite surface growth monitored in real time (DOI:10.1126/science.1250984).
Trouble shooting the system took much longer than anticipated due to several factors: the cell and cantilever had to physically withstand highly alkaline solutions; bubble formation had to be extensively mitigated to permit continuous imaging; feedback control software had to be developed to minimize lateral drift; and the progressive accumulation of silica on AFM tips required extensive calibration. All of these problems were addressed in our first manuscript analyzing the growth of silicalite-1 (or MFI), which set the stage for the paper we published in Nature Communications on zeolite A (or LTA); however, each new system introduces unexpected problems that require addition trouble shooting. For zeolite A, identifying growth mixtures was nontrivial owing to the presence of amorphous colloidal precursors that render solutions opaque and inoperable for AFM. Substantial time was spent figuring out how to extract transparent supernatants with tuned silica supersaturation for AFM measurements.
In our studies of zeolite A, we uncovered several unexpected findings – the most surprising being that zeolites are capable of growing via the formation of gel-like layers from solutions containing molecularly-dispersed solute. To our knowledge, this pathway is the first reported for nonclassical crystal growth. The discovery of this mechanism came from a standard test to identify if AFM tip rastering influenced surface topography. We noticed that the AFM tip progressively removed material from the surface with continuous imaging, which prompted more detailed measurements to confirm that this phenomenon was correlated to a unique gel-like property of surface layers. Extending this analysis to a broader range of conditions led to additional observations of layer-by-layer growth that in some respects confirmed, and in others contradicted, hypotheses of zeolite A crystallization in literature.
Prior to our in situ study, it was difficult to envision alternative mechanisms for zeolite A beyond the expectation of classical growth owing to the presence of layers in ex situ images of crystal surfaces. This seemingly gives credence to Boorstin’s caution against intuition-driven hypotheses; however, it is also important to keep in mind the words of Albert Einstein, who stated that “I believe in intuition and inspiration. Imagination is more important than knowledge. For knowledge is limited, whereas imagination embraces the entire world, stimulating progress, giving birth to evolution. It is, strictly speaking, a real factor in scientific research.” Given the many questions that still remain in zeolite crystallization, let us hope that future advancement of in situ techniques coupled with a little imagination may give us unhindered access to peer more openly into the black box!