Tracing nanoscale variations of optical properties of 2D heterostructures with multidimensional optical imaging
(Nanowerk Spotlight) Biosensing technologies are key to medical diagnosis, life science research, pharmaceutical discovery, and food/water safety assurance. By using label-free or label-based biosensors, researchers can study interactions among biomolecules; observe the activity of cells; or detect drug molecules, viruses and bacteria from bodily fluids.
The development of multimodal biosensors has allowed the creation of entirely new biosensing functionalities. For instance, this made it feasible to differentiate and identify multiple distinct biomolecules from a mixture containing numerous unknowns, as opposed to only detecting for the presence of a specific type of known molecules.
The development of label-free biosensing tools has benefited greatly from two-dimensional (2D) materials and their vast library of van der Waals heterostructures. These materials' high surface-to-volume ratio and atomic thinness yield a strong response to surface adsorption events. Combined with a broad spectrum of optical phenomena, tunable band structure, and favorable electronic properties, this makes 2D systems strong candidates for future optical, electrical, and electrochemical biosensing platforms.
There is one catch, though: Atomically-thin 2D materials often suffer from a range of surface non-uniformities at the nanometer scale – atomic impurities, adsorbates, defects, wrinkles, ruptures – that can cause a change in their optical properties. Unfortunately for biosensor developers, the importance and explicit role of these non-uniform surfaces in producing a 2D material's variability need yet to be fully understood.
However, this is not a trivial materials problem to solve. Whereas optical microscopy tools can provide information about specific properties of a nanomaterial (but at insufficient resolution), electron microscopy allows to explore the morphology of a sample – but not its optical and electronic properties.
Tackling this issue, a combined team from The Pennsylvania State University and the University of North Carolina at Greensboro, has developed a novel multidimensional characterization approach that combines several characterization techniques, allowing to get materials properties of the heterostructure with (almost) nanometer resolution.
"We have developed a method to combine several microscopy tools into a multidimensional imaging tool that allows us to inspect materials' optical properties at a resolution much higher than with regular optical microscopy," Slava V. Rotkin, a professor of Engineering Science and Mechanics at Penn State, tells Nanowerk. "At the core of this method are nanometer-resolution optical maps that we obtained with scattering Scanning Near-field Optical Microscopy (sSNOM)."
"Being familiar with the capabilities of all these tools, we looked at the best of each method and tried to compensate for their drawbacks," says Rotkin. "This is what I believe we were able to achieve with our novel imaging tool."
The researchers anticipate this novel multidimensional imaging method will allow a better understanding of the physics of 2D materials and enable their device applications in optoelectronics, biosensing and information technologies.
Expanding on this, they applied it to a particular example of 2D materials: a heterostructure made of graphene, synthesized at Penn State's 2DCC-MIP, and MoS2 (a classical transition metal dichalcogenide) fabricated in Ignatova's lab at UNCG.
They saw how graphene, placed over the MoS2 triangles, makes a coating and preserves the latter from the oxidation and thus from changing its useful optical properties. To achieve this, the team performed a two-year long aging study and showed the evolution with perfect encapsulation and with scratched graphene.
A high-resolution sSNOM map of the heterostructure: three MoS2 islands are clearly seen, covered with graphene monolayer (there are graphene wrinkles resolved in the vertical direction). Image clearly resolves also small bright MoO-triangles (top right) and MoOS nanoscale regions inside the MoS2 islands (bright spots over large triangles). (Image: Prof. Slava Rotkin, Penn State) (click on image to enlarge)
The researchers were able to map, with sub-wavelength resolution, strain and doping in the heterostructure. By doing this, they determined how much charge is transferred between the atomic layers of graphene and MoS2. The latter influences the excitons in MoS2, which changes the device photoluminescence. It also modifies vibrational modes of graphene, thus changing the Raman response.
They also observed how the optical response of their samples changed, e.g., small Mo oxide (MoO) crystallites that formed during the synthesis together with MoS2 and provide their own response. In addition, they saw partial oxidation of MoS2 into Mo oxy-sulfate (MoOS) dots, appearing as a few-nanometer wide bright spots in optical sSNOM image inside the triangular islands.
"By combining different tools in multidimensional imaging, we were able to observe how all various morphological and elemental non-uniformities of the heterostructure result in the formation of mechanical strain areas and charge doping areas," Rotkin notes. "We mapped how the latter produce the non-uniformity (variability) of optical response, which we analyzed and assessed with regard to the ultimate performance of a particular material for biosensing."
In summary, this new method of multidimensional imaging allows a better understanding of the physics behind optical responses of 2D materials' heterostructures. Enabled with such a detailed knowledge of 2D nanomaterials, researchers can now perform materials synthesis with a quality inspection tool and provide explicit recommendations.
Furthermore, as demonstrated in this work, it is now possible to fine-tune the design of biosensors upon understanding the detailed physics of the optical response of the used 2D material structures. Previously, the optical response you can expect from a new material was only known on average and from a large area. Now, not only you can measure the range of optical response to be expected from a new material; you can tell how this range (variability of response) can be tuned through controlling a particular type of non-uniformity – including doping, strain, charge transfer, impurity, wrinkles, tears, etc.
To demonstrate the power of their new tool with a practical example, the team applied correlated multidimensional imaging, including Raman and near-field microscopies, scanning probe and electron microscopies, to unveil physical processes behind label-free multimodal detection of doxorubicin (DOX, an anthracycline cancer drug) by vertical 2D heterostructures.
"We inspected the morphology (by AFM and SEM) and correlated it to the strain of the heterostructure and to the 2D materials' elemental contents (by Raman and sSNOM)," explains Ignatova. "We detected that the heterostructure has both uniform (hydrostatic) and shear strain components – that is, the strain, which generates a non-uniformity of material properties, is not uniform itself, but rather varies from area to area. These dependencies reflect on the Raman response of the 2D heterostructure" (and plasmonic response, not studied in this work).
"All of the above was correlated to biosensing performance," she continues. "Optical response of heterostructure: photoluminescence, Raman and graphene-enhanced Raman scattering were three channels, which we used for the label-free detection of DOX molecules."
In the next stages of their investigations, the researchers will apply their multidimensional imaging to a variety of 2D materials because it is common that their optical properties vary at the very small scale, which cannot be detected with regular tools.
Furthermore, they plan to develop theoretical foundations for sSNOM spectroscopy. " Here practical applications of the new microscopy are still ahead of the physics foundations – a very interesting and challenging task," Rotkin points out
"Something we would like to add in the future are modern computer science tools, such as Artificial Intelligence methods and/or Machine Learning algorithms," Rotkin concludes. "In fact, doing multidimensional imaging is hard due to the need to correlate a lot of data from different channels. This is where new AI/ML methods could save time and efforts."