Multimodal and Spectral AFM Applied to Problems in Biomedical Device Compatibilization

We describe industry-collaborative research applying multimodal/photothermal AFM-IR to soft-material technologies that aid the body’s acceptance of biomedical devices. Much of our work has been in polymer-drug coatings (e.g., for dexamethasone elution from poly n-alkyl methacrylates) or polymeric fabric (polyethylene terephthalate) engineered to “buffer” a body/metal interface. In the process we are developing understandings of AFM-IR analytical methodology, such as the extent of depth-integration of (chemical) signal, as well as issues of signal/noise (s/n), heating via the IR laser, sample shape/geometry relative to irradiation direction, and more. We present a subset of these topics, both to inform newcomers to AFM-IR and foster discussion among advanced users.<br/><br/>In our core AFM-IR methodologies we utilize pulsed IR irradiation, and resultant AC photothermal expansion (from absorbance), to excite (i) the fundamental contact resonance while under contact-mode Z feedback or (ii) either the fundamental or next higher free eigenmodal resonance while under AC Z feedback, the latter implemented at either the fundamental flexural eigenfrequency or the next higher eigenfrequency. Method (ii) further utilizes <i>heterodyning</i>, by pulsing the IR laser at the <i>difference</i> of the two eigenfrequencies and taking advantage of the nonlinear tip-sample interaction, which causes frequency mixing. In separate submethods of (ii), the IR laser is either (a) pulsed to excite the next highest eigenmode while the fundamental eigenmode is mechanically driven for Z feedback (the latter being traditional AC/“tapping” mode) – what we dub “forward heterodyning”; OR (b) pulsed to excite the fundamental eigenmode while the next higher eigenmode is mechanically driven for Z feedback – what we dub “reverse heterodyning”. In either of these submethods it has been reported that the depth-integration of signal can be much shallower than the case of contact resonance (i), these depths being further a function of parameter settings such as “duty cycle”: the exact IR pulse length in time relative to the pulse repetition period. As such, the depth integration of signal can range from micron-scale at the high end down to tens of nanometers, albeit with accompanying differences in s/n.<br/><br/>We choose to take advantage of these differences in depth integration of signal per<i> research context</i> – whether for polymer-drug coatings, or biomedical device fabric. For the former, the depth-location of drug is an important engineering variable (e.g., to affect “burst” versus longer-term release). Thus depth integration as an <i>analytical variable</i> is useful. For the latter, problems of fabric curling relate to bulk/internal composition, such that we seek to suppress the analytical impact of surface contamination and favor a large signal integration depth. There are also differences in the availability of multimodal tribo-mechanical contrast via images of friction, contact resonance frequency (whereby stiffness), and AC phase (i.e., under conventional tapping). Indeed our selection of method (i) or (ii) is strongly impacted by the presence or absence of sliding friction: highly useful for surface contrast in some cases (favoring method i), while deleterious in other cases, such as very soft materials (favoring method ii). A further consideration in case (ii) is the greater propensity for tip contamination in the net repulsive regime (though yielding higher s/n) compared to the attractive regime. To exemplify, we include a brief example of excellent AFM-IR performance in the AC attractive regime using forward heterodyning, in a comparison of polyethylene spherulitic content (which aids oxygen barrier performance) for the cases of HDPE, LLDPE and blending thereof.