Aerosol‐Jet‐Printed Encapsulation of Organic Photovoltaics

Organic electronic devices (OEDs) are prone to oxygen‐ and water‐induced degradation and therefore need to be encapsulated with barrier materials. In this work, an aerosol jet (AJ)‐printing process is developed to coat perhydropolysilazane (PHPS) directly onto OEDs by adapting the print setup and systematically optimizing the process parameters. Furthermore, a novel curing process that converts PHPS to silica barrier layers is developed by combining damp heat (DH) exposure with subsequent vacuum–UV irradiation. This two‐step treatment is shown to be considerably faster and gentler than the state‐of‐the‐art curing processes and also yields a quantitatively higher conversion. Both the printing and the conversion process are fully compatible with OED devices, which is demonstrated by a damage‐free direct encapsulation of organic solar cells. The encapsulated cells show a significant reduction of degradation in DH conditions (65 °C/85% r.h.), maintaining >95% of their initial performance for >100 h. Complementary electroluminescence measurements reveal that the AJ‐printed barrier layers effectively prevent lateral water ingress into the devices. Herein, the proof of principle is provided that AJ printing can be used to print barrier layers directly onto OEDs and is thus an industrially highly relevant technology to precisely encapsulate such devices even on 3D objects.

DOI: 10.1002/adem.202300322 Organic electronic devices (OEDs) are prone to oxygen-and water-induced degradation and therefore need to be encapsulated with barrier materials. In this work, an aerosol jet (AJ)-printing process is developed to coat perhydropolysilazane (PHPS) directly onto OEDs by adapting the print setup and systematically optimizing the process parameters. Furthermore, a novel curing process that converts PHPS to silica barrier layers is developed by combining damp heat (DH) exposure with subsequent vacuum-UV irradiation. This two-step treatment is shown to be considerably faster and gentler than the state-of-the-art curing processes and also yields a quantitatively higher conversion. Both the printing and the conversion process are fully compatible with OED devices, which is demonstrated by a damage-free direct encapsulation of organic solar cells. The encapsulated cells show a significant reduction of degradation in DH conditions (65°C/85% r.h.), maintaining >95% of their initial performance for >100 h. Complementary electroluminescence measurements reveal that the AJ-printed barrier layers effectively prevent lateral water ingress into the devices. Herein, the proof of principle is provided that AJ printing can be used to print barrier layers directly onto OEDs and is thus an industrially highly relevant technology to precisely encapsulate such devices even on 3D objects. materials based on nano-clays [19] or glass flakes, [20] perhydropolysilazane (PHPS) has been used as printed encapsulant due to its high capability of blocking the ingress of moisture and oxygen. [10,[21][22][23][24] PHPS is a nontoxic and low-cost material that is also widely used as barrier coating for passivating steel or silicon surfaces (including photovoltaics). [25][26][27][28] The backbone of this inorganic polymer class consists of silicon and nitrogen atoms (SiNH 2 -NH) and thus can be converted to silica by exposure to water vapor or reactive oxygen species. [29] Hence, this material is suitable to form dense, homogenous, and defect-free silica films from solution-based processes. [30] In this article, we investigate the potential of AJ-printed PHPS layers for encapsulating OSCs. AJ printing is a noncontact, maskfree printing method that utilizes an AJ to deposit uniform layers of precise geometries onto different substrates. [31][32][33][34][35] This method is frequently used for printed electronics and has the advantage of being able to print onto a wide range of substrates, including flexible ones, and curved or even 3D surfaces. [33,34,36,37] The high lateral resolution of this technique will allow the selective encapsulation of those parts of OEDs, which are sensitive to environmental factors, without affecting other parts of the circuit, especially electrical contacts.
Here, we develop and optimize an AJ-printing process for depositing PHPS layers. We evaluate the performance and stability of both, the barrier layers themselves as well as the encapsulated OSCs devices. The developments are precisely described, which include adjustments to the printing setup and optimization of process parameters to obtain the best-possible topology of printed barrier layers and complete coverage of the solar cells upon encapsulation. We further elaborate a novel process to convert PHPS to silica with extraordinarily high conversion rates in very short times without harming the underlying OSC. Overall, for the first time, the results of this work present an industrially highly relevant process to directly encapsulate OSCs by AJ printing, which can also be applied to 3D surfaces.

AJ Print Setup
For the sake of printing PHPS onto arbitrary (including 3D) objects with high lateral resolution and good thickness control, we chose AJ printing and adapted a commercially available AJ printer to our needs (Figure 1a).
The base AJ setup with a marathon sprint deposition head from Optomec, Inc. consists of a pneumatic atomizer, a solution reservoir, a sheath gas and exhaust hose, a virtual impactor, and a deposition head with a nozzle (300 μm opening). For printing of finely defined structures, a shutter can be attached. This assembly is mounted into a 5-axis kinematic system from Neotech AMT, which allows printing in three dimensions, that is, onto 3D objects.
Some special modifications are made to the setup to enable printing of PHPS. Instead of the softer and cheaper ethylene propylene diene monomer rubber o-rings, the chemically more stable fluorine rubber o-rings are used to give the gas stream pathway a better seal, which reduces instabilities and hence allows for a more homogenous print. Another issue with PHPS is that it immediately starts to precipitate as soon as some of the solvents evaporate, resulting in nozzle clogging and blocking of the gas flow from the atomizer to the nozzle. To prevent this, a bubbler filled with di-n-butyl ether (the solvent of the PHPS formulation) is added to the flow path, which ensures saturation of the gas stream with the solvent, to avoid any clogging of the system. Finally, accumulation of large droplets of PHPS on the nozzle is observed for longer printing durations, which is likely due to condensation of solvent in the hose connecting the virtual impactor and the print head. Such droplets may lead to nozzle clogging and inhomogeneous printing, which is why a droplet catcher is added to the printing setup. The substrate is placed on a plate that can rotate in all directions, which allows www.advancedsciencenews.com www.aem-journal.com for 3D-printing, and the addition of a shutter allows selectively interrupted printing (Figure 1b).

Influence of PHPS Solution Formulation and Printing Parameters
Since defects and inhomogeneities in the PHPS layer will have an effect on its barrier properties, the quality of the printed layers needs to be examined and optimized by adapting the PHPS solution formulation as well as the printing parameters. One major issue that limits the homogeneity of PHPS prints with an AJ is the "coffee ring effect" (CRE). The CRE is the result of inhomogeneous evaporation over the area of droplets on a surface. Higher evaporation rates at the edges of the droplets result in a solute dragging flow of solvent toward them, which creates a ring-like shape of the dried droplet that is typically observed from coffee stain. [38] Consequently, an AJ-printed line will exhibit increased thickness toward its edges, as shown in Figure 2a,b.
To quantify the CRE, we introduce a "CRE index" ψ that represents the ratio of the highest thickness of the edges to the lowest thickness between them, that is, ψ = h max /h min . This allows for quantitative comparability over a large variety of printing parameters. In this work, ψ was determined by scanning the profiles of a printed line at 100 positions with a confocal microscope and taking the average value.
While areas with insufficient layer thickness will directly limit the barrier properties of the PHPS layer, areas with excessive layer thickness may evoke crack formation [9] or prevent complete transformation of the PHPS layer to SiO 2 due to incomplete penetration of the curing agents into the printed line, which in turn impairs the barrier properties negatively as well. Consequently, uniform coating over the entire print area is necessary for achieving good barrier layers. Therefore, the influence of different printing parameters on ψ are investigated and optimized to reduce it to unity.

Printing of Single Lines
The basic optimization of printing parameters is performed by printing and examining individual lines of PHPS onto glass substrates. There are three gas flow rates of importance for AJ printing (see Figure 3): the atomizer flow rate (AFR), or carrier gas flow rate, determines the amount of aerosol that is transferred from the atomizer to the deposition head. The exhaust flow rate (EFR) is generated by the virtual impactor and determines the particle size distribution in the aerosol. The sheath flow rate (SFR) affects the focus of the gas stream that is directed to the substrate. [39] The first step of optimization is to adjust AFR and EFR for a PHPS solution that was diluted 1:1 with di-n-butyl ether as proposed by Channa et al. [10] for blade coating. The threshold for material deposition is an AFR of 700 cm 3 min À1 , if the sheath gas flow is turned off (i.e., SFR = 0). Dependent on the AFR, there is a threshold for the EFR, above which no material is deposited onto the substrate. The correlation of EFR and AFR is measured and plotted in Figure 4a.
The AFR for all following printing experiments is set to 1000 cm 3 min À1 , at which the upper limit for printing is an EFR of 500 cm 3 min À1 . It is noteworthy that the upper limit of the EFR can vary by around 10% for each printing procedure, because it is highly dependent on a good sealing as well as ink properties (e.g., viscosity). By reducing the applied EFR by 20 cm 3 min À1 below its upper limit, the risk of potential interruptions of material deposition due to process fluctuations is effectively reduced (see Figure S1, Supporting Information, for a more detailed correlation of EFR and line properties). In addition, such slightly lower EFR is also found to reduce the CRE index ψ of printed lines in some cases, which can be seen in Figure 4b. This plot also shows the dependence of ψ on the SFR for different EFRs. The lowest CRE index is observed for  values of ERF of 20 and 30 cm 3 min À1 below the upper limit and for values of SFR of 60 cm 3 min À1 , while lower and higher SFRs increase ψ, and hence decrease the homogeneity of the printed lines. This may be explained by the fact that, on one hand, a decreasing SFR generally leads to an increased volume of deposited material (see Figure S2, Supporting Information), which prolongs the drying time of the deposited droplets and in turn enhances the CRE. [40,41] On the other hand, the increase of ψ with increasing SFR at values beyond 60 cm 3 min À1 may relate to the deposited material being blown to the line edges by the increased intensity of the sheath gas flow, as previously reported by Tait et al. [42] Finally, the feed rate (FR) of the substrate also influences the appearance of printed lines (see Figure 4c). The FR is an equivalent for the printing speed and equals the distance covered per minute (at constant material deposition rate), that is, the higher the FR, the less material is deposited on the substrate. Consequently, lower FRs correlate with a higher volume of deposited material and thus longer drying times, which favors the CRE, as described before (see Figure S3, S4, Supporting Information, for a more detailed correlation of FR and line properties). In contrast, too high FRs will result in rough surfaces or even incomplete coverage of the intended areas. The results of our optimization show that the best trade-off, that is, the lowest CRE index, is achieved using an FR of 500 mm min À1 .
Finally, since concentrated PHPS solutions start to chemically react under ambient conditions (light, air, and humidity) already within a couple of minutes and in turn lead to clogging of the nozzle during printing, the PHPS solution is diluted with di-n-butyl ether. Consequently, higher dilution rates allow for longer printing durations. However, increasing the dilution also results in thinner and more inhomogeneous prints, which is shown in Figure 4d. For dilution ratios higher than 1:1, the line quality is found to decrease drastically, becoming too thin and inhomogeneous for the intended application. As a result, the best trade-off between process stability and quality of the layer is a dilution ratio of 1:1, which allows for process durations of at least 1 h.
At this point, it is to note that we optimized the process parameters regarding the functionality of the PHPS as a barrier layer, that is, aiming for the highest homogeneity possible, since too thick and/or uneven layers become rigid and may form cracks (as shown in Figure S5, Supporting Information) and thus lose their barrier properties, as already reported in the literature. [10] With our optimized parameter set that reaches the lowest CRE, the line width of the AJ-printed PHPS line is %1 mm. While it would be possible to further increase or decrease the line width, this would be at the cost of homogeneity (see Figure S6, Supporting Information), and for our application the individual line width is not a key parameter. www.advancedsciencenews.com www.aem-journal.com

Printing of Areas
For printing a 2D area with PHPS, several lines are printed next to each other with a certain degree of overlapping. This degree of overlapping is defined by the "stepover," a value that represents the overlap between adjacent lines. For 100% stepover, the lines are edge to edge (i.e., no overlap), and for a stepover of 50%, the distance between the center of two lines is half the line width. Due to the CRE, this results in an alternating hill-valley topography (see Figure 5a). The lower the stepover is, the thicker the resulting layers are, but also the higher ψ gets, as can be seen in Figure 5b. Both too thick PHPS layers as well as too high ψ values increase the probability of crack formation in the layer (as mentioned in the previous chapter), which impairs the barrier functionality. The lowest CRE is found for a stepover of 50%, while too high stepover values increase the probability of holes in the layer, which also leads to bad barrier properties.
Printing two of such layers on top of each other (without intermittent curing) results in no significant improvement of the CRE index (see Figure 5c), instead, it only results in a proportionally higher roughness of the PHPS layer.
To conclude, based on the previously described optimizations, we managed to reduce the CRE to a minimum with ψ values down to 1.9 and 3.4 for single lines and areas, respectively. The respective printing parameters are summarized in Table 1 and are used for the following experiments. The UV/vis spectrum of such PHPS layers shows a transmission of %100% over the whole spectral range (see Figure S7, Supporting Information), that is, no optical losses are expected if photovoltaic devices are irradiated through PHPS barrier layers.
It is to note that it would certainly be desirable to further decrease the CRE value close to 1, which was however not possible with our setup, since especially the low concentration of the PHPS ink that is needed not to clog the nozzle in our setup unfortunately leads to an increased roughness of the layer (see Figure 4d). Consequently, an optimization of the printing ink in terms of different solvent systems would be one approach to further decrease the roughness of the layers. Furthermore, there are other aspects that could be investigated to potentially further reduce the topographical variation, for example, to investigate the effect of different nozzle diameters or geometries (e.g., flat nozzle). A wider nozzle enables printing of wider line profiles at an appropriate material deposition rate, which could potentially improve the overall surface homogeneity, as fewer lines are required to cover the desired area. Finally, a more extensive investigation of the interaction between the different printing parameters by means of design of experiment could provide more insight into the process window and could yield potential improvement of the achievable layer homogeneity.

Encapsulation of Solar Cells
After printing PHPS, the layer needs to be cured (i.e., converted to silicon dioxide) to provide proper barrier properties. This can be done either by exposure to damp heat (DH) or by vacuum-UV (V-UV) irradiation with an excimer lamp) under oxygencontaining environments. [43] In our case, DH curing was performed at 65°C/85% relative humidity and V-UV irradiation was done with a xenon excimer lamp (λ emission (Xe 2 *) = 172 nm) in air with a lamp-sample distance of 5 mm. The conversion can be verified and quantified by infrared spectroscopy. Upon conversion, the peaks corresponding to stretching vibrations of N-H (3400 cm À1 ), Si-H (2150 cm À1 ), and Si-N (830 cm À1 ) decrease and eventually disappear, while new peaks of Si-O at 450 cm À1 (bending vibration) and 1050 cm À1 (stretching vibration) appear and increase in the spectrum. [43,44] This is shown in Figure 6a for a PHPS film being cured by V-UV light for 1 min, which shows no significant conversion to SiO 2 yet, and for 5 min, where conversion to SiO 2 is clearly detectable. The ratio of the peaks at 1050 and 830 cm À1 can therefore be Figure 5. a) Confocal microscope image of an AJ-printed PHPS pad, b) comparison of height profiles of cured PHPS layers printed with 30% (blue) and 50% (red) stepover, and c) comparison of the height profiles double-(red) and single-layer (red) pads for a stepover of 50%. www.advancedsciencenews.com www.aem-journal.com used to determine the quality of the curing process. According to Channa et al., a peak ratio of two and higher indicates sufficient curing for the purpose of encapsulation. For a sufficiently high conversion ratio of PHPS to silica by V-UV curing only, it is necessary to expose the layers for at least 10 min to V-UV (see Figure 6b). However, OSCs of the layer stack glass/indium tin oxide (ITO)/ZnO (N10)/PM6:Y6:PC 61 BM/MoO 3 /Ag being exposed to such conditions, suffer severe damage, especially when being unprotected, but also when bearing a printed layer of PHPS on top, which can be seen in Figure S8a, Supporting Information. In areas where no PHPS is printed, strong oxidation of the silver top electrode is observed, but also some parts that were covered with PHPS exhibit small-area damages after V-UV curing. The respective power conversion efficiency (PCE) of the cells shows a relative decrease of 70.9% and 17.8% for the bare cells and the PHPS-protected ones, respectively. As can be seen in Figure 6b, this peak ratio threshold is surpassed after 10 min of V-UV curing (black data), whereas with DH curing the peak ratio remains below two even after 60 min of curing (grey data). Consequently, the conversion rate of PHPS is lower under DH compared to V-UV irradiation. This is in accordance with the literature and the reason, why in most cases, V-UV curing is the preferred curing method. [43] However, if the encapsulation is printed directly onto an OED, the highly energetic V-UV irradiation and the formed reactive oxygen species may potentially damage the subjacent OED, a problem that has already been discussed in the literature. [22,45] To reduce the resulting damage, we employ a two-step process that combines the two curing methods. The first step consists in initial curing under DH conditions, which is expected to be less harmful to the OED. The second curing step is performed by V-UV irradiation, which is known to provide more complete conversion. The silica layer that is created during the first curing step is intended to prevent potential photooxidation during the V-UV exposure of the second process step and thus reduce damage to the device. Surprisingly, with this two-step process, we find that much higher peak ratios than for the two individual methods can be achieved in even shorter time (see colored curves in Figure 6b). [10] For example, the combination of 25 min DH exposure and subsequent 5 min of V-UV irradiation result in a peak ratio of 7.5.
The reason for this extreme increase in conversion ratio is not fully understood yet. Possibly, humidity permeates further into uncured PHPS before reacting with the polymer than the reactive oxygen species formed by V-UV irradiation, thus creating silanol groups throughout the layer. To complete the conversion to silica, it is necessary to apply heat that causes condensation of the silanol groups. [46] While complete condensation is not achieved at 65°C, the following V-UV irradiation has been reported to effectively support the conversion. [43] When using our newly developed two-step curing procedure for converting PHPS to silica, the devices need to be exposed to V-UV for only 90 s (after 60 min DH pre-curing) to be fully cured. In this case, no visible damage of devices with a PHPS layer on top is observed (see Figure S8b, Supporting Information). The gentleness of the two-step process is further confirmed by current-voltage (I-V) measurements after each process step. Table 2 shows the average photovoltaic performance of all devices, before encapsulation, after application of a PHPS layer by AJ printing, after pre-curing in DH for 60 min, and after subsequent V-UV curing for 60 min. Remarkably, the performance of the devices was not impaired over the complete process. This means that the two-step curing process allows for high-conversion ratios of ≥2 and at the same time is not harmful to the OSCs.

Measurement and Optimization of Barrier Properties
To investigate the AJ-printed layers regarding their protection against water vapor transmission, encapsulated OSCs of the layer stack glass/ITO/ZnO (N10)/PM6:Y6:PC 61 BM/Poly(3,4-ethylenedioxythiophene) (PEDOT):F/Ag are subjected to a degradation test in DH (65°C/85% relative humidity). Under these conditions, OSCs comprising PEDOT-based hole-transport layers are known to undergo severe degradation due to the hygroscopicity of the PEDOT, featured by the increase of series resistance, and thus can be used as a probe for water ingress into the devices. [47] For AJ-encapsulated solar cells, the two-step encapsulation process described in Chapter 2.2.1 is used, while unencapsulated cells serve as reference. The initial performances (before degradation) of the unencapsulated and AJ-encapsulated solar cells are listed in Table 2. Figure 7a depicts the evolution of device performance during the course of degradation inside the climate chamber. After 137 h, the unencapsulated cells have lost 81% of their initial PCE, while the AJ-encapsulated cells still maintain almost their full initial performance (only 4% loss). The respective current density-voltage (J-V) curves of the cells (Figure 7b,c) indicate water ingress as the culprit for the deterioration of the unencapsulated cell. [47] While the open-circuit voltage (V OC ) remains stable, the fill factor (FF) significantly drops due to an increased series resistance and the commencing formation of a second diode ("S-shape"). For the unencapsulated devices, the FF drops to 51% of its initial value, whereas the FF of the encapsulated devices only drops slightly (see Figure S9, Supporting Information). These results clearly show that the AJ-printed PHPS barriers significantly slow down the water vapor transmission into the solar cells.
To retrace the way of water ingress into the devices upon DH degradation, electroluminescence (EL) images of fresh and degraded devices are taken. Each substrate contains six individual solar cells, two of which have not been encapsulated and four have been encapsulated with PHPS by AJ-printing prior to DH exposure, as shown in the device layout in Figure 8a. Figure 8b shows the EL image of the solar cells before degradation. All active cell areas light up homogeneously, and with the same intensity, independently of whether being encapsulated or not, which is in-line with the equal initial device efficiency of all cells. Upon degradation, all cells lose in overall EL intensity (see Figure 8c), which is in accordance with the observed drop of current injection at the bias of þ1.5 V that is applied for EL imaging (Figure 7b,c). However, the AJ-encapsulated devices (bottom four cells) still show a homogeneous EL signal over the whole active layer with hardly any "dark spots." On the contrary, the unencapsulated devices (two top cells) show large inactive areas or even complete device failure, which is attributed to local ingress of water. [47] Since the dark spots are not homogenously distributed over the cell area, as can be seen in the top left cell, but rather are all connected to one or two edges of the cell, it is very likely that the degradation of the unencapsulated devices is due to lateral water ingress underneath the Ag top electrode. This behavior is effectively suppressed by the AJ encapsulation of the devices.  www.advancedsciencenews.com www.aem-journal.com

Conclusions
A process for encapsulating OSCs with PHPS by means of AJ printing has been developed. This process can be used to protect printed electronics on both 2D as well as 3D surfaces from water and oxygen. Our optimized PHPS solution formulation and printing parameters enable stable printing for at least 1 h and provide homogeneous prints. Furthermore, the newly developed two-step curing process for the conversion of PHPS to silicon dioxide by combining DH exposure and subsequent V-UV irradiation has been shown to transform the PHPS faster and more quantitatively in comparison to both individual curing processes and is thus of high interest for industrial application. In addition, both the printing and the conversion process are fully compatible with OEDs, which enables their encapsulation by printing the barrier layer directly on top of the device without causing any performance losses. Applying this approach to a production environment, OEDs can be encapsulated in-line by using the same printing techniques as used for the device fabrication itself, which saves time and cost for additional materials and machinery (e.g., for encapsulation by lamination). Finally, utilizing AJ printing for applying the barrier allows for a selective and precise encapsulation of water-and oxygen-sensitive OEDs on any kind of substrate, including 3D objects with curved surfaces.
OSC Fabrication: Prestructured glass/ITO substrates were cleaned in an ultrasonic bath in IPA and blown dry. Subsequently, a layer of zinc oxide (N-10, Avantama) was coated onto the substrate with a doctor blade (coating speed: 5 mm s À1 ). Before annealing, the ZnO at 120°C for 30 min in air excess material was removed from the bottom contact areas. PM6:Y6:PC 61 BM (0.9:1:0.1) was used as active layer material. The coating parameters were adjusted, so that the optical density of the films was around 0.5-0.6. The layer was annealed for 10 min at 110°C inside a nitrogen-filled glove box. PEDOT:F was blade-coated on top with 10 mm s À1 without subsequent annealing. The top electrode consisted of 246 nm thick evaporated silver. The resulting cell area was 0.1 cm 2 .
Characterization Methods: Fourier-transform infrared spectroscopy spectra of PHPS samples were recorded with an Alpha II system from Bruker in attenuated total reflection (ATR) mode operating the OPUS/ IR 7.2 software. The profiles of printed encapsulations were measured with a Keyence VK-X3050-3D laser scan microscope. Accelerated lifetime tests to determine the effect of AJ-printed PHPS barriers on the lifetime of OSC devices were performed under well-controlled conditions. The encapsulation was carried out as described in Chapter 2.2.2. The substrates with encapsulated and unencapsulated OSCs were positioned in a climate chamber (ESPEC LHL-144), with the preset condition of 65°C at 85% relative humidity following the standardized DH test protocol for organic photovoltaics (ISOS-D-3). [48] C-V characteristics and PCEs of the solar cells were measured during the ageing experiments by an LOT solar simulator (Class AAA) at 100 mW cm À2 . For this purpose, the solar cells were taken out of the climate chambers and put back after the measurement. To analyze degradation by light, the encapsulated P3HT films were placed under continuous irradiation in ambient air in the chamber of a SUNTEST XXLþ sun simulator (Atlas Materials Testing Technology GmbH) with daylight filter. The light source was a Xenon lamp with an irradiation intensity set to 60 W m À2 in the range of 300-400 nm. The chamber temperature was kept at 40°C, while the black-body temperature was 65°C. Changes in the UV-Vis spectra during degradation were investigated using a Shimadzu UV-1800 spectrometer. Crack formation was observed with an Olympus MX51 optical microscope. The EL images of degraded and undegraded cells were taken with a Ninox VIS-SWIR 640 camera from Raptor Photonics with an air-cooled extended indiumgallium-arsenide detector (InGaAs) upon excitation with a Keysight B2901A Precision Source/Measure Unit from Keysight Technologies, Inc.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.