- Email: [email protected]
Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes
Polymer xxx (2015) 1e8
Contents lists available at ScienceDirect
Polymer journal homepage: www.elsevier.com/locate/polymer
Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes Oleksandr Trotsenko a, Roland Koestner b, Yuri Roiter c, 1, Alexander Tokarev a, Sergiy Minko a, * a b c
Nanostructured Materials Lab, University of Georgia, Athens, GA, 30602, USA GM Powertrain Engineering, Pontiac, MI, 48340, USA Department of Chemistry, Clarkson University, Potsdam, NY, 13699, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 July 2015 Received in revised form 3 November 2015 Accepted 9 November 2015 Available online xxx
Optimizing Nafion loading and surface distribution of Nafion in the fuel cell electrode is critical for the fuel cell performance for minimizing ohmic and mass transport overpotentials. An atomic force microscopy method is used here for a qualitative and a quantitative discrimination between the ionomer and Pt in the fuel cell electrode. This work describes a methodology for the analysis of complex composite surface of fuel cell electrodes and discrimination of different materials on the electrode surface. The reported methodology could be extended for imaging composite rough surfaces when contrast is based on mechanical properties, adhesion and electrical conductivity. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Nafion Adhesion Fuel cell electrodes AFM PeakForce QNM
1. Introduction Commonly used methods for mapping composite surfaces on the nanoscale include energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) associated with electron microscopy methods. However, probing materials with high energy electrons could be distractive for the material structures especially made of organic substances. For example, fluorinated organic materials including fluoropolymers are sensitive to electron beam radiation [1]. Damage caused by interactions of the electron beam with the polymer includes bond breaking, free radical formation and crosslinking, and eventually formation of amorphous carbon [2]. This damage induces structural and chemical changes, as was demonstrated for polymer electrolyte fuel cell (PEFC) membranes [3]. At the same time, high doses are required for EDS or EELS to obtain a spectroscopic map of the surface with a high resolution. A number of strategies were developed [1,3,4] to minimize beam induced damages including cooling samples and selection of special substrates. However, electron microscopy
* Corresponding author. E-mail address: [email protected] (S. Minko). 1 Current address: The Procter & Gamble Company, Cincinnati, Ohio, 45224, USA.
techniques remain destructive yet to the fluoropolymer samples and could change their structure and chemical composition. This problem has attracted recent attention in the light of the development of fluoropolymer, specifically Nafion based, electrodes for hydrogen fuel cells. The PEFC electrodes are complex composite materials made of carbon black nanoparticles decorated with Pt nanoclusters coated with a Nafion ionomer. The tiny details of the structure of the composite electrode are critically important for understanding of the mechanisms of mass transport to the electrode surface. The latter problem is currently considered as a major source of energy losses associate with fuel cell cathodes. Thus, less damaging methods for mapping of the ionomer on the surface of fuel cell electrodes are highly demanded. Atomic force microscopy (AFM) is a powerful and versatile tool to study interfaces and surfaces [5,6]. For the last decades, many AFM based methods have been developed for estimation of mechanical, electrical, magnetic and chemical properties of materials surfaces at nano- and microscopic levels [7e10]. Recently developed software packages enabled simultaneous acquiring combinations of physical properties of the mapped surfaces. For example, adhesion characteristics, mechanical response, and electrical properties can be measure while mapping with Bruker PeakForce QNM and TUNA [11,12] extensions.
http://dx.doi.org/10.1016/j.polymer.2015.11.021 0032-3861/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
2
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Adhesion forces measured with AFM in the air originate from chemical and physical interactions of the probe with the sample. The latter includes van der Waals forces and capillary force generated by the humidity condensed on the probe tip and the sample. The capillary force strongly alternates the tipesurface interactions and could result in inconclusive results of AFM mapping in the humid atmosphere. However, even in the presence of condensed water a chemical contrast was reported based on pulloff forces [13]. It was demonstrated that adhesion measurements at relative humidity (RH) above 50% are not essentially dependent on humidity [13]. The chemical contrast originates from the fact that the pull-off force depends on the water contact angle on the surface of the sample. Hydrophobic surfaces demonstrate lower experimental pull-off force in the air than hydrophilic surfaces. According to Sedin and Rowlen [13], the measured force (Fmeas) includes four contributing forces: the tipesample interaction in vapor (Fstv), the tip-sample interaction in liquid (Fstw), and the capillary force (Fcap) which is a sum of the surface tension (Ft) and the force caused by pressure difference (Fp). The terms Fstv and Fstw may be calculated using Derjaguin approximation for separation of a spherical tip (1) with an apex radius Rt and a flat sample (2) in the presence of the third phase (3, vapor or liquid) [14]:
Fstv=w ¼ 2pRt ðg13 þ g23 g12 Þ where g13 is the tip-vapor and g12 is the tip-water interfacial tension, and g23 is the water-vapor interfacial tension. For both hydrophilic tip and sample, the capillary force dominates over Fstw [15] and:
Fmeas zFstv þ
Ft þ Fp ½ðRHRH 0 Þ=m
1þe
where RH is the relative humidity, RH' is the transition point relative humidity when the capillary forces prevail, m e is the slope of the transition. Ft and Fp depend on contact angles of water (Fig. 1):
Ft ¼ 2pRt gw sinðjÞsinðj þ q1 Þ Fp ¼ 2pHgw R2t sin2 ðjÞ where Rt is the radius of curvature of the tip, gw is the surface tension of water, j is the filling angle, q1 is the contact angle for water on the AFM tip, q2 is the contact angle for water on the flat substrate, and H is the local mean curvature as defined for a circular
Fig. 1. Schematic of a capillary bridge formed between the AFM tip and the sample with hydrophilic and hydrophobic surface properties. Rt is the radius of curvature for the AFM tip, j is the fill angle, r1 and r2 are the principal radii of curvature of the meniscus, q1 is the contact angle for water on the AFM tip, and q2 is the contact angle for water on the sample surface.
approximation of the meniscus (Fig. 1)
1 1 r1 r2 1 cosðq1 þ jÞ þ cosðq2 Þ sinðq1 þ jÞ þ ¼ 2Rt 1 cosðjÞ sinðjÞ
H¼
Both types of interactions Fstv/w and Fcap are surface specific. Therefore, the measured pull-off force depends on contact angle for water on the surface. The force is greater for a hydrophilic surface than for a hydrophobic surface. This simple analysis explains the origin for the chemical contrast when mapping composite surfaces made of materials with different wetting behavior. Mechanical properties of the samples such as Young's modulus can be mapped on the nanoscale using PeakForce QNM tapping technique [7]. In the tapping mode DerjaguineMullereToporov (DMT) model is explored for the unloading of the probe in the contact with the substrate for each data point. Comparing to Hertzian model, the DMT model takes into account adhesive forces [16] and describes elastic contact deformation of a ball (tip apex) and a plane surface [17]. However, if multicomponent thin films are scanned, the sample thickness and sample composition should be considered [8,18]. The DMT model is not applied in cases when the relatively soft polymer material is confined by a hard substrate underneath or aside of the structural features. This may introduce an error in the calculation of modulus. Thus, probing mechanical properties of composite surfaces enables imaging with nanomechanical contrast. In this work we used nanomechanical contrast between a Nafion ionomer and Pt with the deformation channel. By keeping maximum loading force constant, we distinguish ionomer from Pt by higher deformation of the polymer comparing to the hard substrate. A local surface conductivity of the mapping sample is measured when a bias voltage is applied between a conducting tip and the sample, and the electrical current is measured. Specifically, from a currentetime plot during the PeakForce TUNA Tapping oscillation cycle three characteristics are collected: peak current, cycleaverage current, and contact-average current. Peak current is the current when maximum force is applied to the probe. Cycleaverage current are current averaged for the entire oscillation cycle when the tip touches the surface and when it is off the surface. Contact-averaged current is the current averaged for the period when the tip is in contact with the sample as judged from the forceeseparation curve. When the tip is brought in direct contact with a conductive substrate, usually the maximum current is observed [11]. Tip-sample contact area impacts the magnitude of the current, i.e. when the sample is deformed, or multiple contact between probe and rough sample occurs. However, if tip touches dielectric polymer no electrical current is detected. Offline analysis of the current maps is used to calculate statistics of the electrical properties of different regions, the spatial distribution of the properties. The data are used for study of the correlation of mechanical, topographic and electrical properties [11]. The described above mechanisms of the tip-substrate interactions create the background for the analysis of composite surfaces studied with AFM probes using different modes of the probe-substrate interactions on flat substrates. The interpretation of the AFM data becomes much less conclusive on rough surfaces when the tip-substrate contact area varies with dimensions of the topographical features on the surface [19,20]. Very small topographical features may not affect the probe-substrate interactions while structures with dimensions comparable with the probe size could much stronger contribute to the variations in probe-substrate interactions. Surface topographical structures are typically irregular
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
and very by size so that modeling of the probe-substrate interactions is very difficult problem [21,22]. The methodology for mapping of rough composite surfaces to a high extend relies on an empirical optimization of the AFM studies that are adjusted for specific samples of the materials. In this article, we are interested in the mapping of NafionePt fuel cell electrodes to analyze distribution of the Nafion ionomer on the electrode surface. The fuel cell electrode is an interesting example of application of AFM mapping methods for the study of complex surfaces that combine a chemical and topographical inhomogeneity. The complex structure of Nafion ionomer adds to the complexity of the composite electrode. Methods of thin Nafion film preparation and post-deposition treatment play an important role in the film's structure. Thin, less than 55 nm, films prepared by deposition from aqueous and waterethanol dispersions yield a hydrophilic top layer [23]. The structure of rod-shaped Nafion aggregates with sulfonic side-groups exposed in water and hidden fluorinated backbone is transferred to the structure of the film when hydrophobic domains are hidden inside, and the top layer is enriched with hydrophilic groups. However, annealing at 150 C results in switching of the structure when hydrophobic domains are exposed at the thin films hydrophobic interface [24]. Nafion ionomer is typically annealed during the fabrication of actual fuel cell electrodes. Therefore, Nafion has hydrophobic “skin” which in terms of adhesion measurements in air at ambient humidity will result in a low pull-off force comparing to more hydrophilic Pt. This difference in wetting properties was explored in this work to map composited surfaces of the fuel cell electrodes decorated with Pt and Nafion materials. The adhesion mapping mode was combined with topographical, mechanical and electrical mapping of the materials. The experiments were initially conducted for the model substrates with a low roughness prepared by the deposition of Nafion and Pt clusters on the surface of Siwafers. The experiments were followed by mapping of real working electrodes of fuel cells provided by General Motors Cor. 2. Materials and methods Nafion D2020 with equivalent weight (EW) 1000 (purchased from E. I. du Pont de Nemours and Company, Wilmington, DE) dispersed in n-propanol:water (64%:35%) solvent mixture was used in the studies. The stock solution was diluted to the desired concentration with n-propanol. The samples of the dissolved Nafion were incubated for 30e60 min and used the same day. Model Pt substrates were prepared by vapor deposition of densely packed Pt clusters (5e10 nm in size) on a doped silicon wafer. Model Pt substrate was rinsed with ethanol, dried in N2 flow and then plasma treated for 5e10 min. The plasma treatment improved the contrast between hydrophilic Pt (measured water contact angle 44.1 ) and hydrophobic Nafion films (contact angle > 95 , ref. [25]). SEM images of the provided by General Motors Cor. Pt “whisker” electrode decal sample are shown in Fig. 2. The decal has a saw-like profile on a larger scale with hills and trenches. On the smaller scale surface is covered with organic dye whiskers. A thin layer of Pt is vapor-deposited on whiskers. The platinum deposition is followed by deposition of a thin layer of Nafion ionomer. The laminate whisker electrode is transferred on a PFSA membrane [26,27]. In the study, a blank Pt-coated whiskers, five Pt-coated whisker samples loaded with different amounts of Nafion, and three samples with Nafion adsorption promoter additive (less than 5% of oleylamine, OA) [28] were used. 2.1. Atomic force microscopy Bruker AFM Multimode MM8 was used to measure adhesion,
3
deformation and conductivity maps of the prepared samples. Images were taken using PeakForce QNM (Quantitative Nanomechanical Measurements) for nanomechanical and adhesion maps [9]. While scanning with a silicon tip, sinusoidal forceedistance curves were measured for each pixel of the image. The tipsample interaction force is directly controlled using a continuous feedback loop. The curves are analyzed in real-time to generate surface maps. Adhesion maps are calculated from the pull-off force peak of the retraction curves. Deformation maps are generated directly from the forceedistance curve and include both the cantilever and sample deformation. Elastic modulus map is calculated using DerjaguineMullereToporov (DMT) or Sneddon model [9]. All the measurements were performed under ambient conditions at room temperature and at a relative humidity (RH) of 50e55%. Electrical current maps were generated using the PeakForceTUNA mode (Tunneling Atomic Force Microscopy) with a platinum coated tip [11]. During the measurements, the tip touches the surface and the average current is detected at each tip-sample contact. The average current is obtained using a lock-in amplifier over different time intervals as the TUNA-current (the average steady-state current) and the peak current, which is measured at the peak force. For measurements that included adhesion and deformation maps, ScanAssyst-Air probe (Bruker, spring constant 0.4 N/n, a silicon oxide tip) were used. For conductivity map measurements, PFTUNA probe (Bruker, spring constant 0.4 N/m, tip surface coating is Platinum/Iridium) were used. Humidity and temperature were maintained constant for the measurements. To minimize the effect of tip apex geometry, the same calibrated probe was used for one series of experiments. Contaminations attached to the tip can drastically change the probe-sample interaction. Setting contact forces as low as possible allowed maintaining the same probe clean during the series of measurements. A reference sample was always scanned prior and after scanning of the studied samples to confirm consistency of the data. PDMS substrate was used as a reference after the each adhesion measurement of Nafion or Pt layers. If the mean value of the pull-off force for the PDMS sample was changed because of contaminations or damage of the tip, the experiment was repeated with another tip. 2.2. Preparation and characterization of thin Nafion films on model surfaces Nafion thin films on model Pt substrates were prepared by spin coating. Adsorption and dip coating deposition methods resulted in non-uniform Nafion coatings on PteSi substrates. A precise control over humidity is required to avoid dewetting phenomena [29e31]. It was found that spin coating provided good reproducibility for the preparation of thin ionomer films. Prior to Nafion deposition, thickness of Pt layer (dPt) on Si substrate was measured. The PteSi substrate was scratched with a needle removing the Pt layer, and the thickness of the Pt layer was estimated using AFM imaging of the topography of the sample in the area of the scratch (Fig. 3a). AFM measurements showed that the thickness of Pt after plasma cleaning is dPt ¼ 10 ± 0.5 nm. The thickness was measured in three different locations along the scratch for each sample. The stock solution of non-autoclaved Nafion was further diluted with n-propanol to different concentrations and used in all spin coating experiments. Humidity of the air was 55 ± 5% (controlled with a room dehumidifier) and the temperature was 25 C. A droplet of 25 mL Nafion solution was spin coated at a speed of 2000 rpm for 60 s. The spin coating experiments were conducted with various concentrations of Nafion 0.5, 0.25, 0.125, 0.075, 0.05
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
4
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Fig. 2. SEM micro-graphs of the Pt “whisker” electrode decal at two scales (a and b). Hills are the top of the triangular profile, trenches are located between tingles. Length of whiskers varies between 200 and 500 nm.
Fig. 3. Topography image of the SiePt step and SiePt þ Nafion (0.125% spin coated) after the scratch with a steel needle. (a) Profile of the step (Pt is removed from the left part of the sample). (b) Fitted height of the film. (c) Profile of the step (Pt and Nafion are removed from the left part of the sample). (d) Fitted height of the film. (e) Thickness of Nafion layers on PteSi substrates deposited by spin coating vs. concentration of Nafion solution (RH 55%, 2000 rpm).
and 0.01 wt%. After spin coating samples were annealed in vacuum oven at 150 C. After the spin coating deposition of Nafion, the second scratch was made on the sample. This scratch removes the Nafion and Pt layers. The thickness of Nafion þ Pt layer dNfþPt was measured with AFM in the scratch area (Fig. 3c). The thickness of the deposited Nafion layer dNf was calculated by subtracting the thickness of Pt layer dPt from the thickness of Pt þ Nafion layer dNfþPt. A linear dependence of Nafion thickness on concentration of Nafion was observed at concentrations of Nafion less than 0.3% wt (Fig. 3e) after Hall et al. [32]. A series of samples with different thicknesses of Nafion in the range from 50 nm down to a monolayer and submonolayer were prepared. 3. Results and discussion 3.1. Adhesion mode for model substrates The entire cycle of the adhesion contrast mapping of the model Pt and Nafion coated substrates is shown in Fig. 4 that included mapping of the silica wafer with deposited Pt clusters (Fig. 4a), the silica wafer coated with Nafion (Fig. 4b) and a reference PDMS sample (Fig. 4c). The pull off force distribution for the PDMS sample is plotted in Fig. 4d (dotted green line (in the web version)). The mean value for this distribution is 4.23 nN. The mapping of the Pt and Nafion surfaces with the same probe was considered as acceptable if the same mean value 4.2 ± 0.1 nN was obtained for the
probing of the PDMS sample after the mapping experiments on the Pt and Nafion surfaces. The distributions of the pull-off forces for the probed materials were fitted with Gaussian distribution function. Fittings constants are summarized in Table 1. The mean value of the AFM tip-Nafion adhesion force is 2.55 nN for the selected probe, and it is independent on the thickness of the Nafion film. The distribution peak for Pt is 3 nN for the same tip. Therefore, Nafion and Pt can be discriminated by the difference in the adhesion force. In the second series of the experiments, a thin Nafion layer was deposited by spin coating from diluted Nafion solution (0.012%) to produce a submonolayer coating that similar to the thickness on Nafion films in the fuel cell electrodes. The scanning with PeakForce QNM in the adhesion and deformation modes provided a sharp contrast between Pt and Nafion materials (Fig. 5bec). Fig. 5d shows the distribution of adhesion force for the sample where Pt surface is partially covered with Nafion. Adhesion forces that are characteristic for both Pt and Nafion can be discriminated in the plot (Fig. 5 d). The difference in the probability density provides evidence for partial coverage of the surface with Nafion (15.54% of the sample surface, Fig. 5c). It is noteworthy that Pt and Nafion are discriminated at the lateral resolution less than 100 nm. 3.2. Deformation mode on model substrates Essential contrast between Nafion and Pt was achieved in the deformation mode (Fig. 6a, b and c). The deformation for Nafion
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
5
Fig. 4. AMF map of adhesion for (a) Nafion, (b) Platinum and (c) PDMS samples. (d) Distribution function of adhesion force in the scanned areas. Labeling is explained in the inset (see color web version).
structures is in the range of 3.2e3.8 nm and deformation for Pt clusters is constant at 2.6 nm. Even better contrast was achieved for the deformation channel for imaging in a liquid. With appropriate solvent selection, the ionomer swells and becomes much softer hence responses with a higher deformation and lower mechanical modulus. This mode is, however, less attractive for the actual electrodes because of their complex structure on multiple scales. Deformation and modulus measurements are accurate for flat substrates when geometry of the interacting materials is well characterized. 3.3. Electrical conductivity mode for model substrates PeakForce TUNA was used for a conductivity mapping of the samples. AFM probes for this mode are conductive, and when the voltage is applied between the tip and the sample, the current is measured at the each point. The contrast between Nafion and Pt originates from very different electrical conductivity of the materials: dielectric Nafion and metallic conductor Pt. The map for the conductivity of the sample with Nafion clusters is shown in Fig. 6d as measured peak currents at 1 V applied bias. Black spots are structures of Nafion (non-conductive) located on the Pt substrate (conductive). Fig. 6e shows combined topographic and conductivity maps of the sample. The conductivity map of the sample is imposed on the topography map of the sample. The height of the graph represents the topography of the sample and color represents the conductivity. As it can be seen for the Nafion “mountains” the conductivity is zero which reveals the location of non-conductive Nafion structures. 3.4. Mapping of fuel cell electrodes Samples of Nafion deposited on Pt whiskers electrodes were scanned to record topography images (typically 7 by 7 mm scan size) with adhesion maps on electrode's hills and in trenches (see
Fig. 2 and description in Materials and Methods section). Then adhesion force distribution for each individual whisker was analyzed. Average values for 10e15 whiskers were counted for each individual sample (Fig. 7aeb). The results clearly demonstrate the tendency for a decrease of adhesion force with increase of deposited Nafion amount (Fig. 7c). Adhesion force for a flat reference Nafion sample was found to be 2.6 ± 0.3 nN. Adhesion force for Pt on tops of blank whiskers without Nafion is 3.35 nN on the hills and 3.3 nN in the trenches (vs. 3.02 nN for the reference flat substrate). It obvious that the roughness of the whisker electrodes contributes to the deviations from the flat substrates. It is seen from Fig. 7 that the overall adhesion force decreases with Nafion thickness slightly stronger in the trenches, where possibly surface concentration of Nafion is higher as compared with that in the hills. The adhesion force was found to be less affected by the specific location on the electrode for samples with the high Nafion thickness prepared with additive OA. Using a simple additive model, it was estimated for the samples with no OA additive the coverage of whisker tops on the hills by Nafion is 47%, and in the trenches 53%. For the samples with OA additive, those values are 66% and 63%, respectively. PeakForce-TUNA was used to image (Fig. 8) three samples: blank whisker sample, coated with Nafion and coated with Nafion and OA. In the experiment bias voltage 20 mV was applied between the samples and the conductive probe. The image scan size is 5 mm with a resolution of 512 512. The conductivity maps show that Table 1 Characteristics of Gaussian distribution fit for Nafion, Pt and PDMS samples. Sample/adhesion force
Mean value (nN)
Standard deviation (nN)
PDMS Nafion (d ¼ 44.7 ± 4.80 nm) Nafion (d ¼ 13.3 ± 2.00 nm) Nafion (d ¼ 2.7 ± 0.80 nm) Pt
4.23 2.57 2.65 2.57 3.02
0.08 0.14 0.21 0.16 0.13
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
6
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Fig. 5. (a) Topography image of the Nafion structures spin coated from 0.12% solution onto the PteSi substrate, (b) deformation and (c) adhesion maps of the sample; (d) distribution of adhesion force for the sample. The unidirectional elongations of the structures are attributed to a non-ideal shape of the AFM tip apex (see color web version).
Fig. 6. Topography image of the Nafion structures spin coated from 0.012% solution onto PteSi substrate; (b) deformation map of the sample and (c) deformation profile values along the line from (b); (d) conductivity map of the sample and (e) conductivity overlay onto the topography 3D render (see color web version).
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
7
Fig. 7. A characteristic example of images for adhesion mapping of Pt whisker samples. Topography (a) and adhesion (b) recorded in the air on the hill of the corrugation. Average adhesion force for hills and trenches as a function of Nafion thickness for samples prepared without and with the adhesion promoter OA (c) (see color web version).
conductivity decreases in the series of the samples: blank > Nafion > Nafion þ OA. The average conductivities for the samples are 28.0 nA for blank, 21.4 nA for Nafion and 17.8 nA for Nafion þ OA samples. Distribution modes (most common value) are 62 nA for blank, and 0 nA for both Nafion coated samples. If we consider 28 nA to be the maximum conductivity for an uncovered sample, the estimated coverage is a percentage of overall current on the mapped area. The estimated coverages of the electrode surface by Nafion are 24% and 36% for the Nafion and NafionþOA samples respectively. These results are in good agreement with the data obtained with the adhesion mode for the hills (28% and 47% for the Nafion and NafionþOA samples, respectively).
4. Conclusions The AFM mapping experiments of the composite surface made of Pt and Nafion materials in the adhesion mode of PeakForce QNM provided very good contrast for the materials on the model flat substrates and very rough working fuel cell electrodes. Using model substrates, it was demonstrated that the adhesion force for Nafion films is independent on the film thickness up to 50 nm thick films. The latter was used for the analysis of Nafion distribution on the electrode surfaces with different loadings of Nafion. It is shown here that Pt and Nafion coated surface areas can be discriminated quantitatively at the best resolution of less than
Fig. 8. Histogram of sample conductivity (a) and 3D topography render (b, c, d) with conductivity maps overlay for blank whiskers, Nafion coated and NafionþOA coated electrodes. Brighter regions show greater conductivity (see color web version).
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
8
O. Trotsenko et al. / Polymer xxx (2015) 1e8
50 nm with the adhesion map by comparing adhesion force distributions with those for the reference materials obtained using the same AFM tip. The resolution can be potentially improved to sub 10 nm range with sharp probes. Deformation and TUNA modes demonstrated good contrast for Pt and Nafion and can be used to compliment adhesion measurements. Specifically for the fuel cell electrode surface, the developed methods provides the analysis of the ionomer distribution over the electrode surface, thus, helps to optimize the ink composition to decrease transport loses and concentration of defects. References [1] H. Hanawa, K. Kunimatsu, M. Watanabe, H. Uchida, In situ ATR-FTIR analysis of the structure of nafionePt/C and nafionePt3Co/C interfaces in fuel cell, J. Phys. Chem. C 116 (40) (2012) 21401e21406. [2] R.F. Egerton, P. Li, M. Malac, Radiation damage in the TEM and SEM, Micron 35 (6) (2004) 399e409. [3] S. Yakovlev, K.H. Downing, Visualization of clusters in polymer electrolyte membranes by electron microscopy, Phys. Chem. Chem. Phys. 15 (4) (2013) 1052e1064. [4] D.A. Cullen, R. Koestner, R.S. Kukreja, Z.Y. Liu, S. Minko, O. Trotsenko, A. Tokarev, L. Guetaz, H.M. Meyer, C.M. Parish, K.L. More, Imaging and microanalysis of thin ionomer layers by scanning transmission electron microscopy, J. Electrochem. Soc. 161 (10) (2014) F1111eF1117. € rgel, R. Costa, L. Carle , I. Galm, N. Can ~ as, B. Pascucci, [5] R. Hiesgen, S. So K.A. Friedrich, AFM as an analysis tool for high-capacity sulfur cathodes for LieS batteries, Beilstein J. Nanotechnol. 4 (2013) 611e624. [6] R. Hiesgen, S. Helmly, I. Galm, T. Morawietz, M. Handl, K.A. Friedrich, Microscopic analysis of current and mechanical properties of nafion® studied by atomic force microscopy, Membranes 2 (4) (2012) 783e803. [7] M. Sababi, J. Kettle, H. Rautkoski, P.M. Claesson, E. Thormann, Structural and nanomechanical properties of paperboard coatings studied by peak force tapping atomic force microscopy, ACS Appl. Mater. Interfaces 4 (10) (2012) 5534e5541. [8] J.M. Antunes, J.V. Fernandes, N.A. Sakharova, M.C. Oliveira, L.F. Menezes, On the determination of the Young's modulus of thin films using indentation tests, Int. J. Solids Struct. 44 (25e26) (2007) 8313e8334. [9] B. Pittenger, N. Erina, C. Su, Quantitative Mechanical Property Mapping at the Nanoscale with PeakForce QNM, 2010. [10] S.M. Morsi, A. Pakzad, A. Amin, R.S. Yassar, P.A. Heiden, Chemical and nanomechanical analysis of rice husk modified by ATRP-grafted oligomer, J. Colloid Interface Sci. 360 (2) (2011) 377e385. [11] C. Li, S. Minne, B. Pittenger, A. Mednick, M. Guide, T.Q. Nguyen, Simultaneous Electrical and Mechanical Property Mapping at the Nanoscale with PeakForce TUNA, 2011. [12] R. Hiesgen, S. Helmly, T. Morawietz, X.-Z. Yuan, H. Wang, K.A. Friedrich, Atomic force microscopy studies of conductive nanostructures in solid polymer electrolytes, Electrochim. Acta 110 (2013) 292e305.
[13] D.L. Sedin, K.L. Rowlen, Adhesion forces measured by atomic force microscopy in humid air, Anal. Chem. 72 (10) (2000) 2183e2189. [14] J.N. Israelachvili, Intermolecular and Surface Forces: Revised Third Edition, Academic Press, 2011. [15] F.M. Orr, L.E. Scriven, A.P. Rivas, Pendular rings between solids: meniscus properties and capillary force, J. Fluid Mech. 67 (04) (1975) 723e742. [16] B.V. Derjaguin, V.M. Muller, Y.P. Toporov, Effect of contact deformations on the adhesion of particles, J. Colloid Interface Sci. 53 (2) (1975) 314e326. [17] T.J. Young, M.A. Monclus, T.L. Burnett, W.R. Broughton, S.L. Ogin, P.A. Smith, The use of the PeakForce (TM) quantitative nanomechanical mapping AFMbased method for high-resolution Young's modulus measurement of polymers, Meas. Sci. Technol. 22 (12) (2011) 125703. [18] K. Sweers, K. van der Werf, M. Bennink, V. Subramaniam, Nanomechanical properties of alpha-synuclein amyloid fibrils: a comparative study by nanoindentation, harmonic force microscopy, and peakforce QNM, Nanoscale Res. Lett. 6 (1) (2011) 270. € tzinger, W. Peukert, Particle adhesion force distributions on rough [19] M. Go surfaces, Langmuir 20 (13) (2004) 5298e5303. [20] C. Creton, L. Leibler, How does tack depend on time of contact and contact pressure? J. Polym. Sci. Part B Polym. Phys. 34 (3) (1996) 545e554. [21] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between nanoscale rough surfaces: II. Measurement and comparison with theory, J. Colloid Interface Sci. 232 (1) (2000) 17e24. [22] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between nanoscale rough surfaces: I. Role of asperity geometry, J. Colloid Interface Sci. 232 (1) (2000) 10e16. [23] D.K. Paul, K. Karan, A. Docoslis, J.B. Giorgi, J. Pearce, Characteristics of selfassembled ultrathin nafion films, Macromolecules 46 (9) (2013) 3461e3475. [24] D.K. Paul, K. Karan, Conductivity and wettability changes of ultrathin nafion films subjected to thermal annealing and liquid water exposure, J. Phys. Chem. C 118 (4) (2014) 1828e1835. [25] D. Paul, A. Fraser, J. Pearce, K. Karan, Understanding the ionomer structure and the proton conduction mechanism in PEFC catalyst layer: adsorbed nafion on model substrate, ECS Trans. 41 (2011) 1393e1406. [26] D.P. Wilkinson, J. Zhang, R. Hui, J. Fergus, X. Li, Proton Exchange Membrane Fuel Cells: Materials Properties and Performance, CRC Press, 2009. [27] C. Ji, S. Bhargava, M. Dioguardi, Preparation of Nanostructured Thin Catalytic Layer-based Electrode Ink, Gm Global Technology Operations, Inc, 2010. [28] R.J. Koestner, H.M. Viterise, I. Kozhinova, Hydrophobic Onium Salt Addition to Fuel Cell Electrode Inks for Improved Manufacturability, GM Global Technology Operations LLC, 2013. [29] R.M. Walters, A. Taubert, J.-S. Kim, K.I. Winey, R.J. Composto, Surface segregation of counterions in ionomer films, Macromolecules 41 (23) (2008) 9299e9305. [30] M.A. Modestino, A. Kusoglu, A. Hexemer, A.Z. Weber, R.A. Segalman, Controlling nafion structure and properties via wetting interactions, Macromolecules 45 (11) (2012) 4681e4688. [31] Y. Feng, A. Karim, R.A. Weiss, J.F. Douglas, C.C. Han, Control of polystyrene film dewetting through sulfonation and metal complexation, Macromolecules 31 (2) (1998) 484e493. [32] D.B. Hall, P. Underhill, J.M. Torkelson, Spin coating of thin and ultrathin polymer films, Polym. Eng. Sci. 38 (12) (1998) 2039e2045.
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
Contents lists available at ScienceDirect
Polymer journal homepage: www.elsevier.com/locate/polymer
Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes Oleksandr Trotsenko a, Roland Koestner b, Yuri Roiter c, 1, Alexander Tokarev a, Sergiy Minko a, * a b c
Nanostructured Materials Lab, University of Georgia, Athens, GA, 30602, USA GM Powertrain Engineering, Pontiac, MI, 48340, USA Department of Chemistry, Clarkson University, Potsdam, NY, 13699, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 July 2015 Received in revised form 3 November 2015 Accepted 9 November 2015 Available online xxx
Optimizing Nafion loading and surface distribution of Nafion in the fuel cell electrode is critical for the fuel cell performance for minimizing ohmic and mass transport overpotentials. An atomic force microscopy method is used here for a qualitative and a quantitative discrimination between the ionomer and Pt in the fuel cell electrode. This work describes a methodology for the analysis of complex composite surface of fuel cell electrodes and discrimination of different materials on the electrode surface. The reported methodology could be extended for imaging composite rough surfaces when contrast is based on mechanical properties, adhesion and electrical conductivity. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Nafion Adhesion Fuel cell electrodes AFM PeakForce QNM
1. Introduction Commonly used methods for mapping composite surfaces on the nanoscale include energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) associated with electron microscopy methods. However, probing materials with high energy electrons could be distractive for the material structures especially made of organic substances. For example, fluorinated organic materials including fluoropolymers are sensitive to electron beam radiation [1]. Damage caused by interactions of the electron beam with the polymer includes bond breaking, free radical formation and crosslinking, and eventually formation of amorphous carbon [2]. This damage induces structural and chemical changes, as was demonstrated for polymer electrolyte fuel cell (PEFC) membranes [3]. At the same time, high doses are required for EDS or EELS to obtain a spectroscopic map of the surface with a high resolution. A number of strategies were developed [1,3,4] to minimize beam induced damages including cooling samples and selection of special substrates. However, electron microscopy
* Corresponding author. E-mail address: [email protected] (S. Minko). 1 Current address: The Procter & Gamble Company, Cincinnati, Ohio, 45224, USA.
techniques remain destructive yet to the fluoropolymer samples and could change their structure and chemical composition. This problem has attracted recent attention in the light of the development of fluoropolymer, specifically Nafion based, electrodes for hydrogen fuel cells. The PEFC electrodes are complex composite materials made of carbon black nanoparticles decorated with Pt nanoclusters coated with a Nafion ionomer. The tiny details of the structure of the composite electrode are critically important for understanding of the mechanisms of mass transport to the electrode surface. The latter problem is currently considered as a major source of energy losses associate with fuel cell cathodes. Thus, less damaging methods for mapping of the ionomer on the surface of fuel cell electrodes are highly demanded. Atomic force microscopy (AFM) is a powerful and versatile tool to study interfaces and surfaces [5,6]. For the last decades, many AFM based methods have been developed for estimation of mechanical, electrical, magnetic and chemical properties of materials surfaces at nano- and microscopic levels [7e10]. Recently developed software packages enabled simultaneous acquiring combinations of physical properties of the mapped surfaces. For example, adhesion characteristics, mechanical response, and electrical properties can be measure while mapping with Bruker PeakForce QNM and TUNA [11,12] extensions.
http://dx.doi.org/10.1016/j.polymer.2015.11.021 0032-3861/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
2
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Adhesion forces measured with AFM in the air originate from chemical and physical interactions of the probe with the sample. The latter includes van der Waals forces and capillary force generated by the humidity condensed on the probe tip and the sample. The capillary force strongly alternates the tipesurface interactions and could result in inconclusive results of AFM mapping in the humid atmosphere. However, even in the presence of condensed water a chemical contrast was reported based on pulloff forces [13]. It was demonstrated that adhesion measurements at relative humidity (RH) above 50% are not essentially dependent on humidity [13]. The chemical contrast originates from the fact that the pull-off force depends on the water contact angle on the surface of the sample. Hydrophobic surfaces demonstrate lower experimental pull-off force in the air than hydrophilic surfaces. According to Sedin and Rowlen [13], the measured force (Fmeas) includes four contributing forces: the tipesample interaction in vapor (Fstv), the tip-sample interaction in liquid (Fstw), and the capillary force (Fcap) which is a sum of the surface tension (Ft) and the force caused by pressure difference (Fp). The terms Fstv and Fstw may be calculated using Derjaguin approximation for separation of a spherical tip (1) with an apex radius Rt and a flat sample (2) in the presence of the third phase (3, vapor or liquid) [14]:
Fstv=w ¼ 2pRt ðg13 þ g23 g12 Þ where g13 is the tip-vapor and g12 is the tip-water interfacial tension, and g23 is the water-vapor interfacial tension. For both hydrophilic tip and sample, the capillary force dominates over Fstw [15] and:
Fmeas zFstv þ
Ft þ Fp ½ðRHRH 0 Þ=m
1þe
where RH is the relative humidity, RH' is the transition point relative humidity when the capillary forces prevail, m e is the slope of the transition. Ft and Fp depend on contact angles of water (Fig. 1):
Ft ¼ 2pRt gw sinðjÞsinðj þ q1 Þ Fp ¼ 2pHgw R2t sin2 ðjÞ where Rt is the radius of curvature of the tip, gw is the surface tension of water, j is the filling angle, q1 is the contact angle for water on the AFM tip, q2 is the contact angle for water on the flat substrate, and H is the local mean curvature as defined for a circular
Fig. 1. Schematic of a capillary bridge formed between the AFM tip and the sample with hydrophilic and hydrophobic surface properties. Rt is the radius of curvature for the AFM tip, j is the fill angle, r1 and r2 are the principal radii of curvature of the meniscus, q1 is the contact angle for water on the AFM tip, and q2 is the contact angle for water on the sample surface.
approximation of the meniscus (Fig. 1)
1 1 r1 r2 1 cosðq1 þ jÞ þ cosðq2 Þ sinðq1 þ jÞ þ ¼ 2Rt 1 cosðjÞ sinðjÞ
H¼
Both types of interactions Fstv/w and Fcap are surface specific. Therefore, the measured pull-off force depends on contact angle for water on the surface. The force is greater for a hydrophilic surface than for a hydrophobic surface. This simple analysis explains the origin for the chemical contrast when mapping composite surfaces made of materials with different wetting behavior. Mechanical properties of the samples such as Young's modulus can be mapped on the nanoscale using PeakForce QNM tapping technique [7]. In the tapping mode DerjaguineMullereToporov (DMT) model is explored for the unloading of the probe in the contact with the substrate for each data point. Comparing to Hertzian model, the DMT model takes into account adhesive forces [16] and describes elastic contact deformation of a ball (tip apex) and a plane surface [17]. However, if multicomponent thin films are scanned, the sample thickness and sample composition should be considered [8,18]. The DMT model is not applied in cases when the relatively soft polymer material is confined by a hard substrate underneath or aside of the structural features. This may introduce an error in the calculation of modulus. Thus, probing mechanical properties of composite surfaces enables imaging with nanomechanical contrast. In this work we used nanomechanical contrast between a Nafion ionomer and Pt with the deformation channel. By keeping maximum loading force constant, we distinguish ionomer from Pt by higher deformation of the polymer comparing to the hard substrate. A local surface conductivity of the mapping sample is measured when a bias voltage is applied between a conducting tip and the sample, and the electrical current is measured. Specifically, from a currentetime plot during the PeakForce TUNA Tapping oscillation cycle three characteristics are collected: peak current, cycleaverage current, and contact-average current. Peak current is the current when maximum force is applied to the probe. Cycleaverage current are current averaged for the entire oscillation cycle when the tip touches the surface and when it is off the surface. Contact-averaged current is the current averaged for the period when the tip is in contact with the sample as judged from the forceeseparation curve. When the tip is brought in direct contact with a conductive substrate, usually the maximum current is observed [11]. Tip-sample contact area impacts the magnitude of the current, i.e. when the sample is deformed, or multiple contact between probe and rough sample occurs. However, if tip touches dielectric polymer no electrical current is detected. Offline analysis of the current maps is used to calculate statistics of the electrical properties of different regions, the spatial distribution of the properties. The data are used for study of the correlation of mechanical, topographic and electrical properties [11]. The described above mechanisms of the tip-substrate interactions create the background for the analysis of composite surfaces studied with AFM probes using different modes of the probe-substrate interactions on flat substrates. The interpretation of the AFM data becomes much less conclusive on rough surfaces when the tip-substrate contact area varies with dimensions of the topographical features on the surface [19,20]. Very small topographical features may not affect the probe-substrate interactions while structures with dimensions comparable with the probe size could much stronger contribute to the variations in probe-substrate interactions. Surface topographical structures are typically irregular
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
and very by size so that modeling of the probe-substrate interactions is very difficult problem [21,22]. The methodology for mapping of rough composite surfaces to a high extend relies on an empirical optimization of the AFM studies that are adjusted for specific samples of the materials. In this article, we are interested in the mapping of NafionePt fuel cell electrodes to analyze distribution of the Nafion ionomer on the electrode surface. The fuel cell electrode is an interesting example of application of AFM mapping methods for the study of complex surfaces that combine a chemical and topographical inhomogeneity. The complex structure of Nafion ionomer adds to the complexity of the composite electrode. Methods of thin Nafion film preparation and post-deposition treatment play an important role in the film's structure. Thin, less than 55 nm, films prepared by deposition from aqueous and waterethanol dispersions yield a hydrophilic top layer [23]. The structure of rod-shaped Nafion aggregates with sulfonic side-groups exposed in water and hidden fluorinated backbone is transferred to the structure of the film when hydrophobic domains are hidden inside, and the top layer is enriched with hydrophilic groups. However, annealing at 150 C results in switching of the structure when hydrophobic domains are exposed at the thin films hydrophobic interface [24]. Nafion ionomer is typically annealed during the fabrication of actual fuel cell electrodes. Therefore, Nafion has hydrophobic “skin” which in terms of adhesion measurements in air at ambient humidity will result in a low pull-off force comparing to more hydrophilic Pt. This difference in wetting properties was explored in this work to map composited surfaces of the fuel cell electrodes decorated with Pt and Nafion materials. The adhesion mapping mode was combined with topographical, mechanical and electrical mapping of the materials. The experiments were initially conducted for the model substrates with a low roughness prepared by the deposition of Nafion and Pt clusters on the surface of Siwafers. The experiments were followed by mapping of real working electrodes of fuel cells provided by General Motors Cor. 2. Materials and methods Nafion D2020 with equivalent weight (EW) 1000 (purchased from E. I. du Pont de Nemours and Company, Wilmington, DE) dispersed in n-propanol:water (64%:35%) solvent mixture was used in the studies. The stock solution was diluted to the desired concentration with n-propanol. The samples of the dissolved Nafion were incubated for 30e60 min and used the same day. Model Pt substrates were prepared by vapor deposition of densely packed Pt clusters (5e10 nm in size) on a doped silicon wafer. Model Pt substrate was rinsed with ethanol, dried in N2 flow and then plasma treated for 5e10 min. The plasma treatment improved the contrast between hydrophilic Pt (measured water contact angle 44.1 ) and hydrophobic Nafion films (contact angle > 95 , ref. [25]). SEM images of the provided by General Motors Cor. Pt “whisker” electrode decal sample are shown in Fig. 2. The decal has a saw-like profile on a larger scale with hills and trenches. On the smaller scale surface is covered with organic dye whiskers. A thin layer of Pt is vapor-deposited on whiskers. The platinum deposition is followed by deposition of a thin layer of Nafion ionomer. The laminate whisker electrode is transferred on a PFSA membrane [26,27]. In the study, a blank Pt-coated whiskers, five Pt-coated whisker samples loaded with different amounts of Nafion, and three samples with Nafion adsorption promoter additive (less than 5% of oleylamine, OA) [28] were used. 2.1. Atomic force microscopy Bruker AFM Multimode MM8 was used to measure adhesion,
3
deformation and conductivity maps of the prepared samples. Images were taken using PeakForce QNM (Quantitative Nanomechanical Measurements) for nanomechanical and adhesion maps [9]. While scanning with a silicon tip, sinusoidal forceedistance curves were measured for each pixel of the image. The tipsample interaction force is directly controlled using a continuous feedback loop. The curves are analyzed in real-time to generate surface maps. Adhesion maps are calculated from the pull-off force peak of the retraction curves. Deformation maps are generated directly from the forceedistance curve and include both the cantilever and sample deformation. Elastic modulus map is calculated using DerjaguineMullereToporov (DMT) or Sneddon model [9]. All the measurements were performed under ambient conditions at room temperature and at a relative humidity (RH) of 50e55%. Electrical current maps were generated using the PeakForceTUNA mode (Tunneling Atomic Force Microscopy) with a platinum coated tip [11]. During the measurements, the tip touches the surface and the average current is detected at each tip-sample contact. The average current is obtained using a lock-in amplifier over different time intervals as the TUNA-current (the average steady-state current) and the peak current, which is measured at the peak force. For measurements that included adhesion and deformation maps, ScanAssyst-Air probe (Bruker, spring constant 0.4 N/n, a silicon oxide tip) were used. For conductivity map measurements, PFTUNA probe (Bruker, spring constant 0.4 N/m, tip surface coating is Platinum/Iridium) were used. Humidity and temperature were maintained constant for the measurements. To minimize the effect of tip apex geometry, the same calibrated probe was used for one series of experiments. Contaminations attached to the tip can drastically change the probe-sample interaction. Setting contact forces as low as possible allowed maintaining the same probe clean during the series of measurements. A reference sample was always scanned prior and after scanning of the studied samples to confirm consistency of the data. PDMS substrate was used as a reference after the each adhesion measurement of Nafion or Pt layers. If the mean value of the pull-off force for the PDMS sample was changed because of contaminations or damage of the tip, the experiment was repeated with another tip. 2.2. Preparation and characterization of thin Nafion films on model surfaces Nafion thin films on model Pt substrates were prepared by spin coating. Adsorption and dip coating deposition methods resulted in non-uniform Nafion coatings on PteSi substrates. A precise control over humidity is required to avoid dewetting phenomena [29e31]. It was found that spin coating provided good reproducibility for the preparation of thin ionomer films. Prior to Nafion deposition, thickness of Pt layer (dPt) on Si substrate was measured. The PteSi substrate was scratched with a needle removing the Pt layer, and the thickness of the Pt layer was estimated using AFM imaging of the topography of the sample in the area of the scratch (Fig. 3a). AFM measurements showed that the thickness of Pt after plasma cleaning is dPt ¼ 10 ± 0.5 nm. The thickness was measured in three different locations along the scratch for each sample. The stock solution of non-autoclaved Nafion was further diluted with n-propanol to different concentrations and used in all spin coating experiments. Humidity of the air was 55 ± 5% (controlled with a room dehumidifier) and the temperature was 25 C. A droplet of 25 mL Nafion solution was spin coated at a speed of 2000 rpm for 60 s. The spin coating experiments were conducted with various concentrations of Nafion 0.5, 0.25, 0.125, 0.075, 0.05
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
4
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Fig. 2. SEM micro-graphs of the Pt “whisker” electrode decal at two scales (a and b). Hills are the top of the triangular profile, trenches are located between tingles. Length of whiskers varies between 200 and 500 nm.
Fig. 3. Topography image of the SiePt step and SiePt þ Nafion (0.125% spin coated) after the scratch with a steel needle. (a) Profile of the step (Pt is removed from the left part of the sample). (b) Fitted height of the film. (c) Profile of the step (Pt and Nafion are removed from the left part of the sample). (d) Fitted height of the film. (e) Thickness of Nafion layers on PteSi substrates deposited by spin coating vs. concentration of Nafion solution (RH 55%, 2000 rpm).
and 0.01 wt%. After spin coating samples were annealed in vacuum oven at 150 C. After the spin coating deposition of Nafion, the second scratch was made on the sample. This scratch removes the Nafion and Pt layers. The thickness of Nafion þ Pt layer dNfþPt was measured with AFM in the scratch area (Fig. 3c). The thickness of the deposited Nafion layer dNf was calculated by subtracting the thickness of Pt layer dPt from the thickness of Pt þ Nafion layer dNfþPt. A linear dependence of Nafion thickness on concentration of Nafion was observed at concentrations of Nafion less than 0.3% wt (Fig. 3e) after Hall et al. [32]. A series of samples with different thicknesses of Nafion in the range from 50 nm down to a monolayer and submonolayer were prepared. 3. Results and discussion 3.1. Adhesion mode for model substrates The entire cycle of the adhesion contrast mapping of the model Pt and Nafion coated substrates is shown in Fig. 4 that included mapping of the silica wafer with deposited Pt clusters (Fig. 4a), the silica wafer coated with Nafion (Fig. 4b) and a reference PDMS sample (Fig. 4c). The pull off force distribution for the PDMS sample is plotted in Fig. 4d (dotted green line (in the web version)). The mean value for this distribution is 4.23 nN. The mapping of the Pt and Nafion surfaces with the same probe was considered as acceptable if the same mean value 4.2 ± 0.1 nN was obtained for the
probing of the PDMS sample after the mapping experiments on the Pt and Nafion surfaces. The distributions of the pull-off forces for the probed materials were fitted with Gaussian distribution function. Fittings constants are summarized in Table 1. The mean value of the AFM tip-Nafion adhesion force is 2.55 nN for the selected probe, and it is independent on the thickness of the Nafion film. The distribution peak for Pt is 3 nN for the same tip. Therefore, Nafion and Pt can be discriminated by the difference in the adhesion force. In the second series of the experiments, a thin Nafion layer was deposited by spin coating from diluted Nafion solution (0.012%) to produce a submonolayer coating that similar to the thickness on Nafion films in the fuel cell electrodes. The scanning with PeakForce QNM in the adhesion and deformation modes provided a sharp contrast between Pt and Nafion materials (Fig. 5bec). Fig. 5d shows the distribution of adhesion force for the sample where Pt surface is partially covered with Nafion. Adhesion forces that are characteristic for both Pt and Nafion can be discriminated in the plot (Fig. 5 d). The difference in the probability density provides evidence for partial coverage of the surface with Nafion (15.54% of the sample surface, Fig. 5c). It is noteworthy that Pt and Nafion are discriminated at the lateral resolution less than 100 nm. 3.2. Deformation mode on model substrates Essential contrast between Nafion and Pt was achieved in the deformation mode (Fig. 6a, b and c). The deformation for Nafion
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
5
Fig. 4. AMF map of adhesion for (a) Nafion, (b) Platinum and (c) PDMS samples. (d) Distribution function of adhesion force in the scanned areas. Labeling is explained in the inset (see color web version).
structures is in the range of 3.2e3.8 nm and deformation for Pt clusters is constant at 2.6 nm. Even better contrast was achieved for the deformation channel for imaging in a liquid. With appropriate solvent selection, the ionomer swells and becomes much softer hence responses with a higher deformation and lower mechanical modulus. This mode is, however, less attractive for the actual electrodes because of their complex structure on multiple scales. Deformation and modulus measurements are accurate for flat substrates when geometry of the interacting materials is well characterized. 3.3. Electrical conductivity mode for model substrates PeakForce TUNA was used for a conductivity mapping of the samples. AFM probes for this mode are conductive, and when the voltage is applied between the tip and the sample, the current is measured at the each point. The contrast between Nafion and Pt originates from very different electrical conductivity of the materials: dielectric Nafion and metallic conductor Pt. The map for the conductivity of the sample with Nafion clusters is shown in Fig. 6d as measured peak currents at 1 V applied bias. Black spots are structures of Nafion (non-conductive) located on the Pt substrate (conductive). Fig. 6e shows combined topographic and conductivity maps of the sample. The conductivity map of the sample is imposed on the topography map of the sample. The height of the graph represents the topography of the sample and color represents the conductivity. As it can be seen for the Nafion “mountains” the conductivity is zero which reveals the location of non-conductive Nafion structures. 3.4. Mapping of fuel cell electrodes Samples of Nafion deposited on Pt whiskers electrodes were scanned to record topography images (typically 7 by 7 mm scan size) with adhesion maps on electrode's hills and in trenches (see
Fig. 2 and description in Materials and Methods section). Then adhesion force distribution for each individual whisker was analyzed. Average values for 10e15 whiskers were counted for each individual sample (Fig. 7aeb). The results clearly demonstrate the tendency for a decrease of adhesion force with increase of deposited Nafion amount (Fig. 7c). Adhesion force for a flat reference Nafion sample was found to be 2.6 ± 0.3 nN. Adhesion force for Pt on tops of blank whiskers without Nafion is 3.35 nN on the hills and 3.3 nN in the trenches (vs. 3.02 nN for the reference flat substrate). It obvious that the roughness of the whisker electrodes contributes to the deviations from the flat substrates. It is seen from Fig. 7 that the overall adhesion force decreases with Nafion thickness slightly stronger in the trenches, where possibly surface concentration of Nafion is higher as compared with that in the hills. The adhesion force was found to be less affected by the specific location on the electrode for samples with the high Nafion thickness prepared with additive OA. Using a simple additive model, it was estimated for the samples with no OA additive the coverage of whisker tops on the hills by Nafion is 47%, and in the trenches 53%. For the samples with OA additive, those values are 66% and 63%, respectively. PeakForce-TUNA was used to image (Fig. 8) three samples: blank whisker sample, coated with Nafion and coated with Nafion and OA. In the experiment bias voltage 20 mV was applied between the samples and the conductive probe. The image scan size is 5 mm with a resolution of 512 512. The conductivity maps show that Table 1 Characteristics of Gaussian distribution fit for Nafion, Pt and PDMS samples. Sample/adhesion force
Mean value (nN)
Standard deviation (nN)
PDMS Nafion (d ¼ 44.7 ± 4.80 nm) Nafion (d ¼ 13.3 ± 2.00 nm) Nafion (d ¼ 2.7 ± 0.80 nm) Pt
4.23 2.57 2.65 2.57 3.02
0.08 0.14 0.21 0.16 0.13
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
6
O. Trotsenko et al. / Polymer xxx (2015) 1e8
Fig. 5. (a) Topography image of the Nafion structures spin coated from 0.12% solution onto the PteSi substrate, (b) deformation and (c) adhesion maps of the sample; (d) distribution of adhesion force for the sample. The unidirectional elongations of the structures are attributed to a non-ideal shape of the AFM tip apex (see color web version).
Fig. 6. Topography image of the Nafion structures spin coated from 0.012% solution onto PteSi substrate; (b) deformation map of the sample and (c) deformation profile values along the line from (b); (d) conductivity map of the sample and (e) conductivity overlay onto the topography 3D render (see color web version).
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
O. Trotsenko et al. / Polymer xxx (2015) 1e8
7
Fig. 7. A characteristic example of images for adhesion mapping of Pt whisker samples. Topography (a) and adhesion (b) recorded in the air on the hill of the corrugation. Average adhesion force for hills and trenches as a function of Nafion thickness for samples prepared without and with the adhesion promoter OA (c) (see color web version).
conductivity decreases in the series of the samples: blank > Nafion > Nafion þ OA. The average conductivities for the samples are 28.0 nA for blank, 21.4 nA for Nafion and 17.8 nA for Nafion þ OA samples. Distribution modes (most common value) are 62 nA for blank, and 0 nA for both Nafion coated samples. If we consider 28 nA to be the maximum conductivity for an uncovered sample, the estimated coverage is a percentage of overall current on the mapped area. The estimated coverages of the electrode surface by Nafion are 24% and 36% for the Nafion and NafionþOA samples respectively. These results are in good agreement with the data obtained with the adhesion mode for the hills (28% and 47% for the Nafion and NafionþOA samples, respectively).
4. Conclusions The AFM mapping experiments of the composite surface made of Pt and Nafion materials in the adhesion mode of PeakForce QNM provided very good contrast for the materials on the model flat substrates and very rough working fuel cell electrodes. Using model substrates, it was demonstrated that the adhesion force for Nafion films is independent on the film thickness up to 50 nm thick films. The latter was used for the analysis of Nafion distribution on the electrode surfaces with different loadings of Nafion. It is shown here that Pt and Nafion coated surface areas can be discriminated quantitatively at the best resolution of less than
Fig. 8. Histogram of sample conductivity (a) and 3D topography render (b, c, d) with conductivity maps overlay for blank whiskers, Nafion coated and NafionþOA coated electrodes. Brighter regions show greater conductivity (see color web version).
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021
8
O. Trotsenko et al. / Polymer xxx (2015) 1e8
50 nm with the adhesion map by comparing adhesion force distributions with those for the reference materials obtained using the same AFM tip. The resolution can be potentially improved to sub 10 nm range with sharp probes. Deformation and TUNA modes demonstrated good contrast for Pt and Nafion and can be used to compliment adhesion measurements. Specifically for the fuel cell electrode surface, the developed methods provides the analysis of the ionomer distribution over the electrode surface, thus, helps to optimize the ink composition to decrease transport loses and concentration of defects. References [1] H. Hanawa, K. Kunimatsu, M. Watanabe, H. Uchida, In situ ATR-FTIR analysis of the structure of nafionePt/C and nafionePt3Co/C interfaces in fuel cell, J. Phys. Chem. C 116 (40) (2012) 21401e21406. [2] R.F. Egerton, P. Li, M. Malac, Radiation damage in the TEM and SEM, Micron 35 (6) (2004) 399e409. [3] S. Yakovlev, K.H. Downing, Visualization of clusters in polymer electrolyte membranes by electron microscopy, Phys. Chem. Chem. Phys. 15 (4) (2013) 1052e1064. [4] D.A. Cullen, R. Koestner, R.S. Kukreja, Z.Y. Liu, S. Minko, O. Trotsenko, A. Tokarev, L. Guetaz, H.M. Meyer, C.M. Parish, K.L. More, Imaging and microanalysis of thin ionomer layers by scanning transmission electron microscopy, J. Electrochem. Soc. 161 (10) (2014) F1111eF1117. € rgel, R. Costa, L. Carle , I. Galm, N. Can ~ as, B. Pascucci, [5] R. Hiesgen, S. So K.A. Friedrich, AFM as an analysis tool for high-capacity sulfur cathodes for LieS batteries, Beilstein J. Nanotechnol. 4 (2013) 611e624. [6] R. Hiesgen, S. Helmly, I. Galm, T. Morawietz, M. Handl, K.A. Friedrich, Microscopic analysis of current and mechanical properties of nafion® studied by atomic force microscopy, Membranes 2 (4) (2012) 783e803. [7] M. Sababi, J. Kettle, H. Rautkoski, P.M. Claesson, E. Thormann, Structural and nanomechanical properties of paperboard coatings studied by peak force tapping atomic force microscopy, ACS Appl. Mater. Interfaces 4 (10) (2012) 5534e5541. [8] J.M. Antunes, J.V. Fernandes, N.A. Sakharova, M.C. Oliveira, L.F. Menezes, On the determination of the Young's modulus of thin films using indentation tests, Int. J. Solids Struct. 44 (25e26) (2007) 8313e8334. [9] B. Pittenger, N. Erina, C. Su, Quantitative Mechanical Property Mapping at the Nanoscale with PeakForce QNM, 2010. [10] S.M. Morsi, A. Pakzad, A. Amin, R.S. Yassar, P.A. Heiden, Chemical and nanomechanical analysis of rice husk modified by ATRP-grafted oligomer, J. Colloid Interface Sci. 360 (2) (2011) 377e385. [11] C. Li, S. Minne, B. Pittenger, A. Mednick, M. Guide, T.Q. Nguyen, Simultaneous Electrical and Mechanical Property Mapping at the Nanoscale with PeakForce TUNA, 2011. [12] R. Hiesgen, S. Helmly, T. Morawietz, X.-Z. Yuan, H. Wang, K.A. Friedrich, Atomic force microscopy studies of conductive nanostructures in solid polymer electrolytes, Electrochim. Acta 110 (2013) 292e305.
[13] D.L. Sedin, K.L. Rowlen, Adhesion forces measured by atomic force microscopy in humid air, Anal. Chem. 72 (10) (2000) 2183e2189. [14] J.N. Israelachvili, Intermolecular and Surface Forces: Revised Third Edition, Academic Press, 2011. [15] F.M. Orr, L.E. Scriven, A.P. Rivas, Pendular rings between solids: meniscus properties and capillary force, J. Fluid Mech. 67 (04) (1975) 723e742. [16] B.V. Derjaguin, V.M. Muller, Y.P. Toporov, Effect of contact deformations on the adhesion of particles, J. Colloid Interface Sci. 53 (2) (1975) 314e326. [17] T.J. Young, M.A. Monclus, T.L. Burnett, W.R. Broughton, S.L. Ogin, P.A. Smith, The use of the PeakForce (TM) quantitative nanomechanical mapping AFMbased method for high-resolution Young's modulus measurement of polymers, Meas. Sci. Technol. 22 (12) (2011) 125703. [18] K. Sweers, K. van der Werf, M. Bennink, V. Subramaniam, Nanomechanical properties of alpha-synuclein amyloid fibrils: a comparative study by nanoindentation, harmonic force microscopy, and peakforce QNM, Nanoscale Res. Lett. 6 (1) (2011) 270. € tzinger, W. Peukert, Particle adhesion force distributions on rough [19] M. Go surfaces, Langmuir 20 (13) (2004) 5298e5303. [20] C. Creton, L. Leibler, How does tack depend on time of contact and contact pressure? J. Polym. Sci. Part B Polym. Phys. 34 (3) (1996) 545e554. [21] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between nanoscale rough surfaces: II. Measurement and comparison with theory, J. Colloid Interface Sci. 232 (1) (2000) 17e24. [22] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between nanoscale rough surfaces: I. Role of asperity geometry, J. Colloid Interface Sci. 232 (1) (2000) 10e16. [23] D.K. Paul, K. Karan, A. Docoslis, J.B. Giorgi, J. Pearce, Characteristics of selfassembled ultrathin nafion films, Macromolecules 46 (9) (2013) 3461e3475. [24] D.K. Paul, K. Karan, Conductivity and wettability changes of ultrathin nafion films subjected to thermal annealing and liquid water exposure, J. Phys. Chem. C 118 (4) (2014) 1828e1835. [25] D. Paul, A. Fraser, J. Pearce, K. Karan, Understanding the ionomer structure and the proton conduction mechanism in PEFC catalyst layer: adsorbed nafion on model substrate, ECS Trans. 41 (2011) 1393e1406. [26] D.P. Wilkinson, J. Zhang, R. Hui, J. Fergus, X. Li, Proton Exchange Membrane Fuel Cells: Materials Properties and Performance, CRC Press, 2009. [27] C. Ji, S. Bhargava, M. Dioguardi, Preparation of Nanostructured Thin Catalytic Layer-based Electrode Ink, Gm Global Technology Operations, Inc, 2010. [28] R.J. Koestner, H.M. Viterise, I. Kozhinova, Hydrophobic Onium Salt Addition to Fuel Cell Electrode Inks for Improved Manufacturability, GM Global Technology Operations LLC, 2013. [29] R.M. Walters, A. Taubert, J.-S. Kim, K.I. Winey, R.J. Composto, Surface segregation of counterions in ionomer films, Macromolecules 41 (23) (2008) 9299e9305. [30] M.A. Modestino, A. Kusoglu, A. Hexemer, A.Z. Weber, R.A. Segalman, Controlling nafion structure and properties via wetting interactions, Macromolecules 45 (11) (2012) 4681e4688. [31] Y. Feng, A. Karim, R.A. Weiss, J.F. Douglas, C.C. Han, Control of polystyrene film dewetting through sulfonation and metal complexation, Macromolecules 31 (2) (1998) 484e493. [32] D.B. Hall, P. Underhill, J.M. Torkelson, Spin coating of thin and ultrathin polymer films, Polym. Eng. Sci. 38 (12) (1998) 2039e2045.
Please cite this article in press as: O. Trotsenko, et al., Probing rough composite surfaces with atomic force microscopy: Nafion ionomer in fuel cell electrodes, Polymer (2015), http://dx.doi.org/10.1016/j.polymer.2015.11.021