AFM and TERS

Reply

Thomas Mueller replied on Wed, May 12 2010 2:47 PM

rated by 0 users

Verified by Stephen Minne

I think the combination of AFM with Raman is very interesting and there are good reasons why tip-enhanced Raman scattering, despite some challenges, has been a hot research topic in recent years. Below are a few thoughts on its benefits and challenges, please let me know in the forum or offline (tmueller@veeco.com) if this is useful and if you have additional questions. I would be interested to learn more about your application.

Just the combination of AFM and (conventional, confocal) Raman microscopy already provides the benefits of co-localizing complementary information, with AFM being intrinsically strong at providing nanoscale topography and mechanical, electrical, and thermal properties. By mapping out vibrational transitions and thus providing chemical and crystallographic information, Raman spectroscopy complements the AFM information ideally. (side note: You can functionalize the AFM tip to encode chemical information in the interaction force in special cases but I would hesitate to see that approach as a robust and versatile tool for interrogating unknown samples) I think this benefit of more complete, spatially resolved information is already valuable for the analysis of a wide range of complex samples, from biological to engineered materials.

With normal micro-Raman the spatial resolution is realistically on the order of 500nm at best, so what the conventional combination is missing is nanoscale resolution for the chemical information. This very capability of sub-100nm spatial resolution for the chemical information itself is exactly what TERS provides. This is quite unique, to have a tool for chemical information at the nanoscale. There are not many alternatives other than rather involved optical approaches (e.g., STED, PALM) that require the a priori presence of fluorescent molecules thus limiting their applicability. (side note: By chemical I mean chemically distinct functional groups in organic chemistry, as opposed to elemental information, for which there are well established techniques, e.g., those based on dispersing energy of electrons and/or X-rays). And I would say this capability is generally quite relevant and intriguing given the recognition in the last 10 years or so, that understanding physics and deliberately engineering structures at the nanoscale is powerful, as demonstrated for example by biological organisms.

In principle, one can of course construct a whole range of nanoscale spectroscopies based on using the AFM tip as a scattering center, providing a localized optical field. One research group that has been very active developing TERS as well as a range of related techniques is the Raschke group at the University of Washington in Seattle. Of the various candidates, Raman spectroscopy (and thus TERS) has the advantage of the convenience of typically using visible light, not requiring continuous tuning of the excitation wavelength, and providing a signal, that is in principle background free, so one can work with small signals. Unfortunately, it has the disadvantage that those signals are indeed very small, as the cross section for Raman scattering is typically many orders of magnitude smaller than the cross section for inducing an allowed delta-v=1 vibrational transition directly with energy-matched infrared radiation. So when localizing the optical field, better enhance that field strongly, employing plasmonics on a sharp metal object. At visible wavelengths, silver and gold are most suitable. If you implement TERS based on STM, you just need a wire. But of course ideally you like to leverage the versatility of AFM, so the “T” in TERS then needs to be an AFM tip. There the most common approach may be coating a silicon probe with either silver or gold, this is what some of the leading groups in the field are doing.

In terms of an AFM platform that is suitable for developing and using TERS, I see three major requirements. First, you want to start with a high quality AFM that can provide rich high resolution information in those areas where AFM shines as a technique, such as mechanical, thermal, electrical. Otherwise you lose an important reason for combining the techniques in the first place. Second, there are of course the requirements imposed by the integration. These include the realtime communication coordinating the elements of the integrated measurement. And they include the need for excellent, high numerical aperture optical access to the tip-sample junction. For opaque samples from top+side facing the front end of the cantilever. For transparent samples from below. In the latter case, if the transparent samples are biological, you want the AFM to be tailored to accommodating them in their typical carriers under their required conditions and accommodating other techniques typically employed for their interrogation. Retaining the alignment of the optics to the tip-sample junction during AFM imaging implies that the imaging needs to take place through sample scanning, not tip-scanning. A third point I would make is the need for less obvious AFM capabilities (which also must not be compromised in creating the optical integration) such as closed-loop and no drift, enabling accurate positioning during optical integration times. Finally, you need of course a high quality Raman system.

While TERS data has been reported for 10 years now (using Veeco AFMs) with Zenobi an early pioneer, interest in it has been increasing as reflected e.g. in the issue of the Journal of Raman Spectroscopy dedicated to it in 2009. Some of the investigation addresses the specific challenges that TERS is facing. The overall combined experimental setup, while not without complexity, has been documented well in publications. An area of active area of investigation are unique signal observations and their origin in contamination, single molecule behavior, and the nature of the optical near field. Making appropriate probes is a challenge but progress on doing that with high yield has been reported. And the need to distinguishing near-field signals from far-field background can be mitigated through multiple measurements and is also being addressed in recent publications through smart probe designs (e.g., see recent papers by Raschke and by DiFabrizio).