Nanoscale chemical characterization of polymers with nano-FTIR

article
Max Eisele1, Tobias Gokus1, Sergiu Amarie1, Andreas Huber1 and Dalia Yablon2
1. Neaspec GmbH, Munich Germany   
2. SurfaceChar LLC, Sharon MA USA
 
Biography
Dr. Max Eisele finished his undergraduate in the Max-Planck Institute of Quantum Optics in Germany working on laser-induced field emission from sharp metal tips. Fascinated by the idea to combine nanoscale solid-state physics with quantum optics he then did his PhD in nano-optics, developing one of the first ultrafast, pump-probe near-field optical microscopes. Following his PhD, Max joined neaspec GmbH where he is now working as a Senior Application Engineer supporting and advising leading scientist to use and develop neaspec’s near-field imaging and spectroscopy technology.
 
Abstract
Nano-FTIR offers a convenient and easy-to-use platform to measure infrared (IR) spectra with 10nm resolution, a significant 1000x fold improvement over resolution with conventional IR microscopes and spectroscopy.  This technology builds on AFM technology with a sharp metallized tip to deliver highly localized spectra.  Various imaging modes are possible including collection of absorption maps at a given frequency and hyperspectral imaging where a full spectrum is collected at every pixel point. The chemically sensitive IR information is collected simultaneously with the classical AFM data such as topography and multiple mechanical properties to provide a complete characterization of the surface. Nano-FTIR is applied here to the study of polymer materials including diblocks and complex films.
 
Corresponding author
 
INTRODUCTION
Nanoscale characterization of materials has proliferated over the past few decades as various properties can now be reliably measured including topography and mechanical, electrical, and magnetic properties.
 
One area that has eluded the nanoscale for the most part has been spectroscopy to provide chemically sensitive information. This limitation is due to the Abbe diffraction limit, which states that sample features can be resolved on a length scale of approximately λ/2 where λ= wavelength of incident light. So, for example, a wavelength of blue light of approximately 470nm allows resolution of features of only 235nm.
 
For analytical chemistry, infrared spectroscopy is one of the most useful analysis tools to study a wide variety of primarily organic liquids and solids. Famous for its chemical fingerprinting region in the mid-infrared region of 400-4000cm-1 (25µm to 2.5µm), infrared light is absorbed by key molecular vibrations to elucidate important structures key to chemical and biological processes. Because of its long wavelength, the resolution of IR is particularly challenging with ~10µm being the limit in the mid-infrared region. 
 
Significant resolution enhancement in the infrared is now available through a novel method of nano-FTIR that collects IR spectra with an unprecedented 10nm resolution, representing a 1000x fold improvement over traditional IR methods. nano-FTIR builds on the success of the atomic force microscopy (AFM) platform to provide high resolution nanoscale chemical identification of a material. The nano-FTIR spectra can be compared with those from standard libraries for accurate chemical identification.
 
This article reviews nano-FTIR, a complete tool that explores a wide variety of surface characteristics adding nanoscale IR characterization to the traditional properties measured by AFM. This tool enables correlative microscopy of a wide variety of materials including polymers, photonic devices, and biological materials where these various properties are mapped together.
 
Finally, nanoscale hyperspectral imaging with nano-FTIR is demonstrated where a spectrum is collected as a function of lateral position. This measurement yields a three-dimensional data cube with a wealth of highly localized chemical information that can be further analyzed by multivariate analysis common in the spectroscopy field. 
 
All the chemical, topographic, and mechanical information is provided with 10nm resolution – only limited by the radius of curvature of the AFM probe. nano-FTIR is applied to a wide variety of polymer materials including a diblock of PS/PMMA, a multilayer of polyethylene bound by polyamide layers, and a carbon fiber nanocomposite. 
 
EXPERIMENTAL SECTION
 
The nano-FTIR (neaspec, GmbH) spectroscopy system combines state-of-the art AFM technology based on very sharp metal-coated AFM tips with standard FTIR spectroscopy methods. As shown in Figure 1, a broadband infrared laser beam (L) illuminates the tip, which functions as a kind of nano-antenna to focus the light into a highly localized region at the probe’s apex. Because of the optical interaction between the tip and the sample (as the tip and sample are so close to one another) the elastically backscattered light (S) contains information about the local chemical information of the sample material directly below the tip. 
 
Figure 1. Concept of nano-FTIR. Light (L) is focused with a sharp, metallized AFM tip.  The scattered (S) can be analyzed for chemically specific information on the sample.
 
The backscattered light is detected by a Michelson interferometer to generate an interferogram, which then is processed with a fast Fourier Transform (FFT) resulting in a simultaneous measurement of the absorption (complex part of the measured spectrum) and reflectance (real part of the measured spectrum). The absorption measured in this highly localized, near-field measurement correlates directly with conventional absorption spectra collected in the far-field with no further data processing needed. The illumination lasers are broadband and span the spectral range from 650cm-1 to over 6500 cm-1 covering the entire traditional chemical fingerprinting region in the mid-infrared. 
 
A comparison of spectra collected with standard FTIR spectroscopy and nano-FTIR is shown in Figure 2. In the middle of Figure 2 is a spectrum of a polycarbonate thin film collected with nano-FTIR (blue spectrum) showing the peaks associated with the C-O-C stretch (1160cm-1, 1189cm-1, 1227cm-1), the aromatic C=C stretch at 1498 cm-1 and the carbonyl stretch at 1772cm-1
 
Figure 2. IR spectra from nano-FTIR can be directly compared with IR spectra from conventional ATR-FTIR spectroscopy.   Both absorbance (proportional to the imaginary part of the refractive index) and reflectance (proportional to the real part of the refractive index) spectra are collected where the peaks in the nano-FTIR (blue spectra) can be mapped to peaks in the conventional FTIR spectra (black spectra).
 
A spectrum of bulk polycarbonate collected with ATR-FTIR (attenuated total reflection FTIR spectroscopy) is shown for comparison directly above (black spectrum) revealing identical spectral peaks as those obtained with the nano-FTIR. Thus, nano-FTIR can be used in combination with standard infrared databases of vibrational spectra to perform chemical identification with an unprecedented confidence level.
 
In addition to the IR absorption spectra, which corresponds to the traditional FTIR absorbance spectra, the reflectance channel is collected simultaneously in nano-FTIR without the need for Kramers-Kronig relations. The reflectance data is proportional to the real part of the refractive index (where the absorption is proportional to the imaginary part of the refractive index), and offers another mechanism of contrast for imaging in addition to the traditional absorbance image. The reflectance spectra on polycarbonate are shown in the bottom panel of Figure 2 where the nano-FTIR spectrum (blue) again matches that collected with standard ATR-FTIR (black).
 
Because nano-FTIR is based on an AFM platform, the chemical information is obtained as an additional channel to the conventional measurement performed by AFM. The nano-FTIR operates in tapping mode where the tip is oscillated at a resonance frequency and “taps” along the surface for a gentle interaction with the surface suitable for both soft and hard materials. It employs conventional silicon based AFM cantilevers that have been coated in metal where platinum/iridium is a commonly used, inexpensive durable, chemically inert, and widely available coating. 
 
The nano-FTIR can operate in three major modes. In the first mode, individual spectra can be collected at highly localized points on the surface. 
 
The most common implementation of this mode is to image the surface first to collect a conventional AFM image together with spectrally integrated optical reflectance image, and then position the probe at areas of interest for individual spectroscopy measurements.
 
The second mode of operation involves collecting the absorption and reflectance at a fixed wavelength at each pixel point as the tip raster scans across the surface resulting in an absorption image and a reflectance image at a fixed frequency.
 
The final mode involves collecting a full spectrum at every pixel point in the image, which offers a complete set of data and with that the ultimate insight into the sample composition. Spectra can be recorded at an acquisition time as fast as 0.2 seconds/spectrum. Thus, for a hyperspectral data cube with 50 x 50 pixels, the full data set would be acquired in only 8 minutes. This kind of mapping is referred to as hyperspectral nano-FTIR imaging. 
 
RESULTS AND DISCUSSION 
 
1. Absorption imaging of a PMMA/PS diblock
 
The chemical identification provided by nano-FTIR allows unambiguous identifications of materials in a blend. Figure 3 shows a 1µm x 1µm nano-FTIR image of a diblock thin film of PMMA/PS. In the topography channel in grayscale on the bottom, two components are evident where one is topographically high (white) and one is topographically low (dark). However, it is unclear which component is PMMA and which is PS and so the nano-FTIR data is collected simultaneously to address this question.
 
Figure 3. 1μm x 1μm nano-FTIR image of a diblock thin film of PMMA/PS. On the left, the topography image is on the bottom (grayscale) and is collected simultaneously with the reflectance image (brown/yellow) and absorbance image (red/blue), both of which were collected at 1730cm-1. On the right is a cross sectional profile (see black dashed line in the topography image) from the height and absorbance image
 
The absorbance image is collected at 1730cm-1 as this is the absorption peak for the carbonyl functional group (C=O) present in PMMA and is shown in the top red/blue panel. In this image, one of the components clearly exhibits high (red) absorption while the other component exhibits low (blue) intensity, indicating that the “red” component is PMMA while the “blue” component is PS.
 
A cross-sectional profile across a part of the absorbance image together with the height cross sectional profile is shown on the right. Note that the absorbance profile resolves features that are a mere 10nm in width which do not correlate to any topography changes and prove the ultimate resolving power of this technology. The length scale of these features clearly requires a technique like nano-FTIR to spatially differentiate these two components. Finally, the reflectance image in the middle (brown/yellow) also shows excellent contrast between the two diblock materials. 
 
2. Correlation spectroscopy of polymer stacked film
 
A powerful feature of nano-FTIR is that it builds on the AFM platform, and thus on all the other information provided by AFM techniques including topography and material properties. Figure 4 shows a multilayer of low density polyethylene (LDPE) which is bound by thin, 20nm polyamide adhesion layers. The top image shows the absorbance collected at 1660cm-1, the frequency corresponding to the amide I absorption band (C=0 + C-N stretching mode) where the polyamide layers show strong absorption (white) with respect to the LDPE layers. The reflectance layer below also shows strong contrast between the polyamide and LDPE. 
 
Figure 4. nano-FTIR of LDPE layers bound by polyamide adhesion layers. The top absorbance image (blue) was collected at 1660cm-1, the amide I absorption band. The corresponding reflectance (yellow/brown) and topography image (grayscale) are shown below. Sample kindly provided by Royal DSM.
 
In addition to the chemical information extracted by the nano-FTIR absorbance and reflectance images, additional data is provided simultaneously by correlating nano-FTIR with state-of-the art microscopy modes: (i) conventional atomic force microscopy enables topographic imaging as well as material contrast (phase imaging), thermal analysis, and stiffness/adhesion by subsequent force spectroscopy measurements. (ii) tip-enhanced Raman spectroscopy (TERS) and nano-photoluminescence (nano-PL) complement the chemical identifications by nano-FTIR and give additional information about the vibrational and electronic structures of materials. The ability to collect and correlate all this information of sample topography, chemistry, material and mechanical properties at the nanoscale provides a complete understanding of a sample for the first time.
 
3. Hyperspectral infrared nanoimaging of a carbon fiber composite
 
Collection of a full IR spectrum at every pixel point in an image results in what is termed hyperspectral infrared nanoimaging, which is a large data cube that can be mined for information from many different directions. Figure 5 shows a hypercube of spectra measured by nano-FTIR on a composite of carbon fibers embedded in epoxy. A full spectrum was collected at every pixel point in this 10µmx10µm image and exemplary spectra collected on the epoxy and carbon fiber are shown in red and grey, respectively.
 
Figure 5. 10μm x 10μm, hyperspectral datacube of a carbon fiber composite in epoxy on left. Full spectra are collected at each pixel. On right, ‘slices’ of the cube at 1577cm-1 (epoxy peak) and 1100cm-1 are shown for both the absorbance and reflectance images.
 
The epoxy spectrum shows a peak at approximately 1577cm-1 while the carbon fiber spectrum has no peaks in the 1400-1600cm-1 region. Reflectance and absorbance ‘image-slices’ from the hyperspectral datacube at 2 frequencies, 1577cm-1 and 1100cm-1 are shown in the right of Figure 5. 
 
Focusing on the absorbance data on the top, the 1577cm-1 image clearly shows strong (pink) absorbance in the epoxy region and weak (blue/black) absorbance in the carbon fibers providing unambiguous identification of these two components.
 
The 1100cm-1 absorbance image shows a weaker contrast distinction between the two materials since neither epoxy nor carbon fiber has strong absorption at this frequency. In contrast, the reflectance image at both 1577cm-1 and 1100cm-1 shows excellent contrast between the two components.
 
This is a good example where the reflectance channel at individual frequencies is a useful additional source of contrast and is collected simultaneously with the absorbance data without requiring any further parameter optimization or processing.
 
Hyperspectral imaging represent the future direction of these kinds of measurements since hyperspectral datacubes provide rich spectral data with spatial localization.
 
In addition, multivariate data analysis based on established procedures known from far-field IR spectroscopy for principle component analysis can be performed on this hyperspectral data set. Full spectra on the nano-FTIR can be acquired as fast as 0.2 seconds/spectrum resulting in an extremely fast acquisition time for a 50 x 50 pixel image of only eight minutes.
 
CONCLUSION
 
Nano-FTIR is a powerful platform for complete nanoscale characterization of a material’s chemistry, topography, and material properties. This technique provides highly localized chemical identification on the 10nm length scale through IR spectra, which compare directly with IR spectra collected with conventional, far-field IR spectroscopy.
 
The spectra collected with nano-FTIR require no modeling or post-processing, and they can be analyzed with established multivariate analyses and procedures from far field spectroscopy. Applications of nano-FTIR for polymer characterization are described here including analysis of block copolymers, stacked polymer film, and hyperspectral imaging of a nanocomposite. 
 
REFERENCES
 
1. F. Keilmann and R. Hillenbrand, Phil. Trans. R. Soc. Lond. A. 362, 20 04, 787-805
2. F. Huth, A. Govyadinov, S. Amarie. W. Nuansing, F. Keilman, R. Hillenbrand Nano Letters, 12, 2012,  3973 - 3978
3. I. Amenabar, S. Poly, M. Goikoetxea, W. Nuansing, P. Lasch, R. Hillenbrand, Nature Communications, 8, 2017, 14402
 
Website developed by S8080 Digital Media