Characterization of Photovoltaics Using Atomic Force Microscopy

With growing energy needs worldwide, it is essential to develop sustainable photovoltaic (PV) technologies. These are based on materials which have the property of producing electricity from light energy. Solar cell technology is currently hung up on the cost reduction issue which hinders its widespread commercial use.

This is because lowering solar energy costs requires more cost-effective manufacturing practices¹, higher efficiency of power conversion and longer lifespans for solar devices. Each of these are dependent upon better characterization methods and in particular, the ability to achieve enhanced spatial resolution. This is even more so given the growing use of materials that are characterized on the micro and nanoscale, such as polycrystals in perovskite films, organic semiconductors which contain bulk heterojunction networks, and light-trapping layers that are nanotextured.

The atomic force microscope (AFM) is an ideal technology to complement other imaging methods², as well as whole-device analytics, because it can achieve spatial resolution to nanoscale breadths.  In addition, it can measure the structure as well as the functional data, which better brings out the correlation between the structure, the characteristics, the processing and the performance of the device, as Figure 1 shows.

This article discusses the ability of AFMs today to analyze the properties of perovskites and organic conductors, the two exciting semiconductor materials of today. AFMs are also of great use in studying others, such as organic semiconductors (Si, CdTe) and chalcopyrites (CIGSSe, or CuInSe₂), as well as tandem systems that employ more than one absorbing material, by analogy.

Figure 1. Visualizing nanoscale photoresponse in MAPbI₃  Solar-cell performance metrics such as short-circuit current I_SC are usually measured at the device scale, but characterization on the nanoscale can elucidate the crucial role of microstructure. The image shows short-circuit current I_SC overlaid on topography for a film of methylammonium lead triiodide (CH₃NH₃PbI₃ or MAPbI₃ ) under ~0.07 W/cm² illumination. It was obtained by first acquiring images of current with photoconductive AFM (pcAFM) at bias voltages ranging from 0 to +1 V. The images were then combined to form an I-V curve for each pixel location, from which values of I_SC were determined. Scan size 3 μm; acquired on the MFP‑3D‑BIO AFM. Adapted from Ref. 3.

Perovskites

The use of hybrid organic-inorganic perovskites as the building material for solar cells is an appealing advance because it allows great increases in the power conversion efficiency. Namely, at a level of over 22% in a span of just seven years^{1, 4}. This material also brings down the cost of solar cell production using simpler and less expensive techniques for solution processing, such as spin coating. Research today targets the measurement of fundamental characteristics as well as long-term stability of this material. Both of these areas are significantly aided by the use of AFM.⁴

Grain Structure of Perovskites

For both basic and practical reasons, it is important to know the microstructure of this material. For example, it could help understand why the grain size of perovskite crystals modulates the PV response, as well as resolving bigger issues in manufacture, such as the manner in which perovskite crystallizes from the precursor state.

As Figure 2 illustrates, AFM is helpful by giving a quantitative 3D picture of the height of the surface, or its topography. These images show various characteristics of the film, like uniformity and coverage, which in turn allows surface metrics (such as roughness) to be quickly determined for different films for rapid comparison purposes.

Topography images with AFM are obtained using tapping or contact modes, and the vertical resolution is typically subnanometric. Some of the newest AFMs can achieve picometric scales of vertical resolution, so that crystals and molecules can be imaged at lattice scale. These devices are also developed with a host of automated features to speed up the experimental setup and make data acquisition smoother.

When perovskites are exposed to the environment, oxidation or other reactions may lead to irreversible microstructural changes, as well as alterations in other properties. To prevent these, the AFM experiments should be carried out in a purified environment of inert gas within custom cells. These also regulate the humidity of ambient or inert gases for the experiment. If even stricter control is needed, the whole AFM may be placed in a glovebox to seal it from the atmosphere, as Figure 5 shows.

Figure 2. Altering grain structure Solution-processing techniques can yield dense perovskite films with uniform surface coverage. However, these films are typically very fine grained, leading to increased grain-boundary losses and hence reduced photoconversion efficiencies. In this work, a guanidinium thiocyanate (GUTS) precursor treatment was developed to increase grain size. These topography images for (left) an untreated MAPbI₃ film and (right) a film post-treated with GUTS / isopropanol solution (4 mg / ml, GUTS-4) show that treatment increased the average grain size from nanometers to micrometers. In addition, the power conversion efficiency of solar cells prepared with GUTS-4-treated films increased by ~2% over that of cells with untreated films. Scan size 5 μm. Adapted from Ref. 5.

Measurement of Electrical and Functional Responses

Photoelectrical data must be acquired in order to gain an insight into the many PV modes of action. This measurement must be performed on a micro or nanoscale because of the polycrystalline structure of the perovskite films. This is where the AFM really helps to analyze charge transport, trapping and recombination with other related processes, because it can perform high-resolution electrical mapping. The power of these techniques is boosted if an AFM that can illuminate the sample is used.

Local current can also be mapped using conductive AFM (CAFM). If this is carried out using illumination, it is called photoconductive AFM (pcAFM). Both of these techniques use a conducting tip to detect the flow of current from the sample under the DC bias voltage applied to it. Any local changes in the photoconductivity, light-induced variations in carrier mobility, and other aspects of this behavior are captured by scanning in contact or using a fast-force-mapping mode.

Signal artifacts are a possibility, to avoid which the AFM detection laser is switched off while pcAFM measurements are underway. The use of different parameters like bias voltage, intensity of illumination, wavelength and polarization can be adjusted to improve the data acquired.

These techniques also provide current-voltage (I-V) curves at nanoscale resolution. A position is defined by the user and the tip is placed on it. Subsequently, the current is measured while ramping the bias voltage. The I-V curves help understand how charge is generated and injected, as well as contact resistance, the results produced by annealing and other variables in the process as shown in Figure 3.

The use of CAFM and pcAFM require a stretch of AFM capabilities, because these measurements must be very sensitive and require low noise. This is regardless of the fact that the currents may be anywhere from picoamperes to microamperes, which is six orders of magnitude. Again, experiments aimed at quantification of the contact between the tip and the sample need the cantilever spring constant to be calibrated so that the force applied can be found.

Figure 3. Investigating ion migration at grain boundaries The mechanisms for many intriguing effects in perovskites such as current hysteresis and thermoelectricity are poorly understood. The image represents KPFM surface potential overlaid on topography for a polycrystalline MAPbI₃ film. Correlating surface potential with crystallographic orientation obtained by transmission electron microscopy (not shown) revealed that boundaries between grains with large surface potential differences (e.g., △) had higher angles than those with small potential differences (e.g., ⬜ and ◯). Local I-V curves acquired with CAFM showed strong dark-current hysteresis at the high-angle grain boundaries but very little hysteresis at low-angle boundaries. (Blue and red arrows indicate increasing and decreasing voltage, respectively.) The results indicated that grain-boundary migration was much faster than, and dominated over, migration within grains. Scan size 2 μm; acquired on the MFP-3D AFM. Adapted from Ref. 6.

Figure 4. Correlating local optical and nanoelectrical properties Understanding the origins of spatial heterogeneity in perovskite properties will help attain higher efficiency values. In this work, a methylammonium lead tribromide (CH₃NH₃PbBr₃ or MAPbBr₃ ) film was deposited on glass / indium tin oxide / poly(3,4-ethylenedioxythiophene) polystyrene (glass / ITO / PEDOT:PSS). Maps of local relative photoluminescence (PL) intensity were created by scanning a 488-nm laser beam across the bottom of a sample mounted on the AFM stage. Images of injected current acquired with CAFM (here, at +3.2 V bias) showed behavior anti-correlated with PL intensity. The dashed, dotted, and solid curves indicate regions with dim, intermediate, and bright PL response but high, intermediate, and low injected current, respectively, despite similar topography. Moreover, FM-KPFM surface potential images lacked any correlated behavior. Comparison to results for a MAPbBr₃ film on bare glass suggest that the heterogeneities arose from effects at the electrode-film interface, not within the film. Scan size 7 μm; acquired on the MFP-3D AFM. Adapted from Ref. 7.

Unlike these modes, KPFM uses the contact potential difference between the sample and the tip as shown in Figures 3 and 4. It has the important advantage of being capable of quantifying work function, which is a major cause for variations in potential in many PV systems. Using KPFM to achieve imaging of work function at nanoscale leads to very minute information on band bending, the density of the dopant, and light-induced variations.

KPFM is typically carried out using a dual-pass approach, with amplitude modulation (AM), like EFM. This is not mandatory, because a single-pass frequency-modulated mode of operation is also possible. The use of FM-KPFM is associated with higher spatial resolution and yields more information by probing the hyper-harmonic effect produced by the cantilever.

Using a multimodal approach and correlating the results also helps to gain a deeper insight into PV materials. Here many AFM modes are used with or without other tools for characterization of the material. This includes scanning electron microscopy (SEM) and transmission electron microscopy (TEM), photoluminescence (PL), and Raman spectroscopy, to acquire complementary data. One instance is shown in Figure 3 where KPFM, CAFM, and TEM are used on the same sample, and Figure 4 which involves the use of CAFM, KPFM and PL.  

Figure 5. Detecting ferroelasticity Topographic imaging (left) of a MAPbI₃ (CH₃NH₃PbI₃) film prepared by solvent annealing revealed micrometer-sized crystalline grains with a terraced structure. The corresponding image of vertical PFM amplitude (right) was acquired at 300 kHz (near resonance) with +2.5 V AC bias. Domains of regularly spaced stripes not present in the topography were observed, with 90° changes in orientation between adjacent domains. Sections across the color-coded lines in the PFM image show that the stripe periodicity varied and ranged from approximately 100 to 350 nm. The results suggest the film was ferroelastic, with a domain structure highly dependent on film texture and thus the specific preparation route. Scan size 7 μm. Acquired on the MFP-3D AFM in a glovebox with nitrogen. Adapted from Ref. 8.

Engineering Interfacial Layers

The simplest form of a solar cell includes a perovskite layer (which acts as the absorber) in between two sandwiching electrodes. Improved designs use more layers to enhance the performance. The AFM is useful when it comes to examining these layers one by one or together. It is possible to change the view of the solar cell,  either the plan view when the tip of the conducting AFM is used as the top contact, or the cross-section view, when the characteristics of material behavior is analyzed across and at the interfaces themselves.

Using nanoscale imaging to reveal the topography of interfacial layers allows a good understanding of the roughness at the surface, which is a key parameter affecting adhesion between the layers. It also shows some features of the morphology, such as phase segregation and the dispersion of organic films.

It is also useful to use electrical modes, such as CAFM and pcAFM when the uniformity of conduction is to be evaluated, for instance, or to find out where charge trapping or recombination of charges is occurring. It is especially useful to perform characterization using KPFM because it is very sensitive to the potential at the surface contact, and has a work function.

The choice of interfacial layers is often determined by the need for carriers to have an easier route to migrate away from the absorber towards the electrode. Thus they are designed to increase the energy level alignment at individual interfaces. The feedback provided by KPFM is very useful since it supplies images of spatial variations in the work function and band bending, as seen in Figure 6.

Figure 6. Improving stability with interlayers Effective use of electron transport layers (ETLs) requires better control of their properties. These maps of surface potential were acquired on a MAPbI₃ (CH₃NH₃PbI₃) film on NiOx before and after addition of phenyl-C₇₁-butyric acid methyl ester (PC₇₀BM) and rhodamine 101 (Rh) layers. The Rh layer significantly reduced spatial variations in potential by passivating defects at the perovskite grain boundaries. The graph of induced surface photovoltage (i.e., difference in surface potential from light to dark) shows the additional layers decreased surface potential, reducing band bending at the ETL /cathode interface. The results helped explain measurements of increased efficiency and stability for devices with Rh layers. Scan size 1 μm; acquired on the MFP-3D AFM in dual-pass KPFM mode. Adapted from Ref. 9.

Organic Semiconductors

Another new PV material is composed of polymers with small organic molecules which comprise the foundation for the organic solar cell. These have great potential because of their use of more environmentally friendly and easily obtained materials which are manufactured at lower costs and by simpler methods, such as solution processing and vapor deposition. Power conversion efficiencies of over 10% are already known, making this technology fit for commercial adoption. The biggest obstacle now is therefore prolonging the lifetime of the solar cell from years to decades. This depends primarily on understanding how light, heat and other factors of the environment affect the performance. This is where AFM measurement is very useful, as it can assess the local structure and other characteristics.

Mapping the Morphology of BHJ

The structure of an organic solar cell is built around a bulk heterojunction (BHJ) as the light absorber, which contains both donor and acceptor materials in a nanoscale network structure achieved by self-assembly. The efficiency of this cell is affected mainly by the specific connectivity and phase segregation of the network.

It is difficult to predict what kind of structure will result from any selected route of processing. Another issue is the potential change in morphology that can occur over time due to various effects of aging. It is therefore crucial to analyze the properties of the morphology of BHJ films at micro- and nanoscale. An option often used is electron microscopy, but this involves damage to the sample while processing it for adequate image contrast.

AFM imaging of the topography of a sample shows the size of various BHJ components as well as their dispersion. It also helps to analyze how process variables affect the rate of solvent evaporation and annealing, as Figure 7 shows. The tapping mode is generally used for topographic imaging on organic substrates, as this uses very small sideways and vertical tip-sample contact force.

This reduces the potential for damage to the sample to the minimum, while enabling greater spatial resolution because of the reduced tip-sample area in contact. The use of very small cantilevers is characteristic of the new generation AFMs which perform rapid scanning. These are capable of resolving and controlling a force in the sub-piconewton range, which is important when analyzing delicate polymers which are easy to deform.

It is also possible to detect mechanical properties with AFM modes that help to study the morphology of BHJ. One example is the use of phase imaging in the tapping mode, which allows blend components to be seen in greater contrast and achieves higher resolution of finer details of the structure. It is also possible to map the elastic modulus using force curve techniques, and thus to understand phase separation and dispersion better, as in Figure 8.

Other than these, there are nanomechanical modes suitable for fast and qualitative imaging, and also for quantitative maps of the response to elastic and viscous forces. One particular example is the use of the recent bimodal tapping methods, like AM-FM mode, to achieve rapid but high-resolution mapping.

Figure 7. Tuning performance via fluorination Substituting fluorine for hydrogen in the conjugated polymer backbone can enhance efficiency and durability. Here, systematic studies of this effect were performed on four narrow-band-gap polymers: PF-0 without fluorine, PF-1a and PF-1b with intermediate fluorine and different regioselectivity, and PF-2 with the most fluorine (see Ref. 12). Solution-processed films of polymer / PC₇₀BM blends were created with varying amounts of the solvent additive DIO. These topography images for the PF-1a blend indicate that low amounts of DIO increased phase separation and thus improved power conversion efficiency, but higher amounts yielded sub-optimum morphology. The graph reveals that root-mean-squared roughness generally increased with fluorine content in all four blends, likely due to enhanced aggregation. Scan size 5 μm; acquired on the MFP-3D AFM in tapping mode. Adapted from Ref. 12.

Figure 8. Evaluating molecular weight effects – Films were created containing blends of diketo-pyrrolopyrrole-thienothiophene polymer (PDPP4T-TT) and phenyl-C₆₁-butyric acid methyl ester (PCBM) with varying polymer chain number average molecular weights. Maps of Young’s modulus acquired with force curve techniques allowed the separate BHJ phases to be identified, with lower and higher modulus values corresponding to PDPP4T-TT and PCBM, respectively. (Insets show the corresponding tapping-mode topography.) Large PCBM domains observed for the film with intermediate molecular weight indicated a PCBM-rich surface created by vertical segregation during spin casting. The result could explain the unusually low values of series resistance measured on a transistor made with this film. In contrast, the phases appeared well intermixed in the other films, which yielded transistors with higher series resistance. Scan size 3 μm; acquired on the Cypher AFM. Adapted from Ref. 13.

Here, the sample was a layer of poly(3-hexylthiophene) and phenyl- C₆₁-butyric acid methyl ester (P3HT:PCBM) bulk heterojunction annealed on an indium tin oxide (ITO) substrate. The sample was imaged at −1 V bias using the ORCA holder. During current imaging, the 530-nm illumination source was turned on and off while increasing the intensity in 1% increments (full power ~0.9 W/cm²). The vertical section through the image reveals the dependence of measured current on intensity and demonstrates high sensitivity to small changes in intensity.

Imaging Nanoscale Photoresponse

It is important to study the areas of charge injection, transport, trapping, and recombination in organic semiconductors, to improve the efficiency as well as to decrease the lowering of performance. This kind of data can be obtained using AFM imaging of the light response at nanoscale, helping to understand how this occurs and the exact location of each process within the BHJ.

The use of CAFM and pcAFM to acquire images from organic semiconductors helps visualize nanoscale photocurrent and networks of charge transport in the donor-acceptor blend. Using these modes it is possible to analyze the part played by anisotropy of structure, light intensity and other parameters in the process of converting light to electric charge, as Figure 9 shows.

The issue with organic semiconductors is their soft and flimsy nature, rendering them damage-prone when exposed to the typical lateral forces experienced with the traditional CAFM and pcAFM performed in contact mode. Another problem is the variation in measured current due to sample and tip wear, suffered during contact-mode scanning, which can make it difficult to interpret the image.

The development of rapid current mapping modes is designed to obviate such issues. Here, the cantilever moves vertically over a continuous sinusoidal curve with simultaneous lateral scanning. The current is measured while force curves in a high-speed array are also acquired. If illumination is also included, the topography can be correlated with the current measurements to determine the relationship between the local structure and the property. In addition, many forms of data analysis are presented by these methods, if complete current vs. time curves and deflection vs. time curves are saved.

The use of EFM and KPFM also has several advantages in helping map the electrical characteristics of organic semiconductors. EFM allows local variations to be mapped in capacitance gradients, while KPFM is useful for mapping surface potential, thus allowing improvements in the performance and long-term stability of the device. These modes do not require contact with the sample, so that there are minimal energy-barrier effects to overcome due to the work function of the tip. The measurements, in short, actually show the open-circuit response of the system.

When dual-pass scanning is used, both EFM and KPFM image acquisition takes a few minutes which makes them ideal for analyzing processes which do not take place very fast. Rapid processes are better examined using other electrical modes like FM-EFM or cantilever ringdown imaging. Such processes include charge injection and carrier diffusion over milliseconds to seconds, as well as photochemical breakdown. The latter is examined using electrical modes as above, to acquire images of the local variations in the dissipation of power and charge trapping.

Figure 9. Exploring photocurrent heterogeneity in P3HT:PCBM The pcAFM current image at +1 V bias reveals domains of higher and lower conductivity in a blend of poly(3-hexylthiophene) (P3HT) and PCBM. I-V curves for the color-coded positions in the image were acquired in the dark and while illuminated (~0.09 W/cm², 530 nm). In both cases, current increased with voltage below −0.3 V bias and then transitioned to much higher resistance at further positive bias. In some positions the amount of current flow depended on the illumination condition (e.g., black and blue circles), while it was always high in others (e.g., green circle). Hysteresis between the forward and reverse bias directions was also observed, indicative of sample capacitive charging. Scan size 1 μm. Acquired on the MFP-3D Infinity AFM with PV option and ORCA holder.

There are still other techniques, such as time-resolved EFM or heterodyne KPFM which allow aspects such as the lifetime of local charge carriers, charging rates in response to light, and thermal annealing, to be studied dynamically in perovskites and organic semiconductors. These are added options on commercially available AFMs rather than being standard. However, they show how sophisticated an open software platform can be, because all Asylum AFMs have open control architecture. This does away with all limits on changing the routines of image acquisition and analysis. One example is the synchronization of changes in illumination with the measurements taken, while another is the amalgamation of commands to form a single automated batch protocol.

Interlayer Optimization

Organic solar cells usually have more layers added in to enable functions such as charge extraction and reception, and regulated surface recombination. The optimization of the layer design requires nanoscale data which is obtained by AFM analysis. One instance is the assessment of BHJ morphological changes due to additional interlayers, which can be done by topographic imaging. Such additional layers may affect the efficiency with which carrier recombination occurs. EFM and KPFM can also be used to acquire images across interfaces so that a better design can be built, which achieves a greater electric field-energy level alignment between the light absorber and the electrodes.

Another advantage of using more layers is the better stability of the device in the long term. For instance, device geometries may be turned upside down or fully enclosed. When AFM testing is done to evaluate device lifetime and stability, it may be desired or even crucial to be able to regulate the environmental factors as well. This allows the testing to be done while the exposed device is enclosed in inert gas, or under increased humidity, or humidity levels which simulate real conditions, as shown in Figure 10. Another benefit is the ability to control temperature using specialized stages to achieve accurate and consistent temperatures of up to a few hundred degrees.

Figure 10. Characterizing humidity-dependent effects P-type metal oxides could serve as effective hole extraction layers in organic solar cells, but the impact of environmental conditions on their electrical properties is not fully understood. In this work, KPFM imaging of a polycrystalline NiOx film revealed nanoscale spatial variations in surface potential that depended on relative humidity (RH). The average value of surface potential decreased with increasing RH, while the average size of topographical features increased. This behavior is consistent with charge screening due to water adsorption on the film surface. The observed spatial irregularities in surface potential most likely arose from uneven chemisorption by exposed crystallites of different orientations. Scan size 1 μm; acquired on the Cypher AFM. Data courtesy of the Center for Nanophase Materials Sciences, Oak Ridge National Laboratory; adapted from Ref. 18.

Environmental Control for PV Materials that are Sensitive to Oxygen and Water

There are several PV materials which should not be exposed to oxygen or water in the surroundings without irreversible damage due to surface reactions. To prevent degradation of the sample, or just thin down the layer of water to improve the accuracy of the electrical measurements, full isolation from the environment comes with Turnkey Glovebox Solutions on both the Cypher series of AFMs and the MFP-3D.

Another possibility to achieve environmental isolation is the Closed Fluid Cell designed for MFP-3D and the Liquid Perfusion Cantilever Holder for Cypher ES AFMs. These units are supplied in an external glovebox. They contain the sample and the cantilever. The inlet ports are sealed. The whole unit is kept on the AFM so that measurements can be taken without exposing the sample to the atmosphere.

Conclusions

PV materials are being used to offer technologies which may potentially help resolve the energy crisis looming globally. The most promising among the new technologies is the use of perovskites and organic semiconductors. These advances may bring abundant, inexpensive sustainable energy within the reach of this generation, but this needs much more research to fully characterize the PV materials used.

The AFMs of today have a large range of modes which help to visualize the material structure at nanoscale, as well as the functional response. This is achieved under both dark and variable illumination conditions. When this is considered in conjunction with the better spatial resolution, more rapid imaging, and better environmental control possible, AFMs are seen to be a crucial part of PV research.

References and Further Reading

1. A.  Polman, M.  Knight, E.  C.  Garnett, B.  Ehrler, and W.  C.  Sinke, Science352, aad4424 (2016).
2. E.  M.  Tennyson, J.  M.  Howard and M.  S.  Leite, ACS Energy Lett.2, 1825 (2017).
3. Y.  Kutes, Y.  Zhou, J.  L.  Bosse, J.  Steffes, N.  P.  Padture, and B.  D.  Huey, Nano Lett. 16, 3434 (2016).
4. Li, B.  Huang, E.  N.  Esfahani, L.  Wei, J.  Yao, J.  Zhao, and W.  Chen, npj Quantum Materials2, 56 (2017).
5. N.  D.  Pham, V.  T.  Tiong, D.  Yao, W.  Martens, A.  Guerrero, J.  Bisquert, and H.  Wang, Nano Energy 41, 476 (2017).
6. Y.  Shao, Y.  Fang, T.  Li, Q.  Wang, Q.  Dong, Y.  Deng, Y.  Yuan, H.  Wei, M.  Wang, A.  Gruverman, J.  Shield, and J.  Huang, Energy Environ.  Sci.9, 1752 (2016).
7. D.  Moerman, G.  E.  Eperon, J.  T.  Precht, and D.  S.  Ginger, Chem.  Mater.29, 5484 (2017).
8. I.  M.  Hermes, S.  A.  Bretschneider, V.  W.  Bergmann, D.  Li, A.  Klasen, J.  Mars, W.  Tremel, F.  Laquai, H.-J.  Butt, M.  Mezger, R.  Berger, B. J.  Rodriguez, and S.  A.  L.  Weber, J.  Phys.  Chem.  C120, 5724  (2016).
9. 9. J.  Ciro, S.  Mesa, J.  I.  Uribe, M.  A.  Mejia-Escobar, D.  Ramirez, J.  F.  Montoya, R.  Betancur, H.-S.  Yoo, N.-G.  Park, and F.  Jaramillo, Nanoscale 9, 9440 (2017).
10. J.  R.  O’Dea, L.  M.  Brown, N.  Hoepker, J.  A.  Marohn, and S.  Sadewasser, MRS Bull.37, 642 (2012).
11. M.  Pfannmoeller, W.  Kowalsky, and R.  R.  Schroeder, Energy Environ.  Sci.6, 2871 (2013).
12. J.  Yuan, M.  J.  Ford, Y.  Zhang, H.  Dong, Z.  Li, Y.  Li, T.-Q.  Nguyen, G.  Bazan, and W.  Ma, Chem.  Mater.29, 1758 (2017).
13. A.  Gasperini, X.  A.  Jeanbourquin, and K.  Sivula, J.  Polym.  Sci., Part B:  Polym.  Phys.54, 2245 (2016).
14. M.  Kocun, A.  Labuda, W.  Meinhold, I.  Revenko, and R.  Proksch, ACS Nano 11, 10097 (2017).
15. R.  Giridharagopal, P.  A.  Cox, and D.  S.  Ginger, Acc.  Chem.  Res.49, 1769 (2016).
16. J.  L.  Garrett, E.  M.  Tennyson, M.  Hu, J.  Huang, J.  N.  Munday, and M.  S.  Leite, Nano Lett.17, 2554 (2017).
17. T.-H. Lai, S.-W.  Tsang, J.  R.  Manders, S.  Chen, and F.  So, Mater.  Today 16, 424 (2013).
18. C.  B.  Jacobs, A.  V.  Ievlev, L.  F.  Collins, E.  S.  Muckley, P.  C.  Joshi, I.  N.  Ivanov, J.  Photonics Energy 6, 038001 (2016).

This information has been sourced, reviewed and adapted from materials provided by Asylum Research - An Oxford Instruments Company.

For more information on this source, please visit Asylum Research - An Oxford Instruments Company.