SYSTEMS HAVING POLYMERIC FIBERS WITH METALLIC NANOPARTICLES THEREON AND METHODS OF FABRICATION

Abstract

Systems and methods that entail polymeric fibers produced via an electrospinning process, and metallic nanostructures adhered to surfaces of the polymeric fibers via an electroless deposition process. Suitable materials for the polymeric fibers and metallic nanostructures include polyacrylonitrile (PAN) fibers and copper nanostructures, respectively.

Description

BACKGROUND OF THE INVENTION

The present invention generally relates to functional material systems and devices and to methods of producing the same. The invention particularly relates to systems that include polymeric fibers having metallic nanostructures adhered thereto.

Certain metallic nanostructures possess striking magnetic, thermal, electronic, and/or surface properties markedly distinct from those of their bulk counterparts. However, the efficient and extended utility of these properties can be attenuated by spontaneous aggregation typically encountered in these nanostructures, which is a structural response to reduce their surface energy.

Polymers have been used to hinder aggregation and bestow stability on metallic nanostructures. The directed assembly of metallic nanostructures on microscopic and submicroscopic sized features often serves two purposes: mitigate structural coalescence between individual nanoparticles; and confine the growth of these metallic species on appropriate substrates so as to exploit distinctive properties that these nanomaterials possess for various functional applications.

Often, polymer matrices either serve as a medium for dispersion of pre-formed metallic nanoparticles, or are utilized as a micro-reactor for in-situ synthesis. In both cases, nanoparticle stability is often achieved through steric hindrance from the polymer chains. However, a corollary of these processes is that a substantial number of the nanoparticles may be permanently embedded in the polymer and so as to be occluded from the functional environment. This renders the material design less efficient for applications in which the nanoparticle must make contact with fluids or gases, and may lead to either additional costs or an unexpected change in the properties of the fibers with embedded particles (e.g., such as an increase in stiffness).

Attempts have been made to immobilize metallic nanostructures on electrospun nanofibers and microfibers. Leveraging the relative ease of polymer chemistry modification, a high-surface area to volume ratio as well as process flexibility, electrospun fibers may serve as matrices for encapsulation of functional organic and inorganic materials and as templates for directed growth confinement of nanostructures. However, the contrast between the intrinsic strong cohesive energy of metallic species and the weak van der Waals forces that hold together polymer macromolecules strongly influences interfacial interaction, creating weak adhesion in most metal-polymer systems. For applications in which electrospun structures are subjected to mechanical strain (e.g., when being used as water filters where flow causes nonwoven membranes to flex) poorly adhered nanoparticles can cause the structure to lose efficacy and lead to potential release of nanoparticles into the environment.

In view of the above, it can be appreciated that there are certain problems, shortcomings or disadvantages associated with the prior art, and that it would be desirable if systems and/or methods were available for creating polymeric systems, such as systems including electrospun polymer fibers, having well-adhered, surface-confined nanoparticles thereon that were capable of at least partly overcoming or avoiding these problems, shortcomings or disadvantages.

BRIEF DESCRIPTION OF THE INVENTION

The present invention provides a method suitable for creating systems including polymer fibers having well-adhered, surface-confined metallic nanoparticles thereon and systems formed thereby.

According to one aspect of the invention, a system is provided that includes polymeric fibers produced via an electrospinning process, and metallic nanostructures adhered to surfaces of the polymeric fibers via an electroless deposition process. Suitable materials for the polymeric fibers and metallic nanostructures include polyacrylonitrile (PAN) fibers and copper nanostructures, respectively.

According to another aspect of the invention, a method is provided that includes producing polymeric fibers via an electrospinning process, and producing metallic nanostructures on and adhered to surfaces of the polymeric fibers via an electroless deposition process. Suitable materials for the polymeric fibers and metallic nanostructures include polyacrylonitrile (PAN) fibers and copper nanostructures, respectively.

Technical effects of the system and method described above preferably include the ability to provide metallic nanostructures that are well-adhered to polymeric fibers and substantially confined to surfaces thereof thereby allowing for the production of devices that exploit distinctive properties of the nanoparticles for various functional applications.

Other aspects and advantages of this invention will be appreciated from the following detailed description.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1A through 1C schematically represent a parallel electrode configuration for aligned nanofiber processing (FIG. 1A), a drum configuration for aligned microfiber processing (FIG. 1B), and a process of electroless deposition for nanocube growth (FIG. 1C).

FIGS. 2A through 2C relate to as-spun aligned polyacrylonitrile (PAN) fibers. FIG. 2A includes an image of PAN nanofibers with an inset showing the nanofibers at a higher magnification. FIG. 2B includes an image of microfibers with an inset showing the microfibers at a higher magnification. FIG. 2C represents orientation distribution of both the nanofibers and the microfibers. The size of the scale bars in the insets of FIGS. 2A and 2B are 1 μm.

FIGS. 3A through 3D include images showing dispersed positions of nanocubes on nanofibers (FIG. 3A), a higher magnification image of a well-defined nanocube on a nanofiber showing cube face anchored to the fiber surface (FIG. 3B), dispersed distribution of nanocubes on microfibers (FIG. 3C), and a higher magnification image of nanocubes anchored to the surface of a single microfiber (FIG. 3D).

FIG. 4 includes SEM and EDS mapping of copper nanocubes on aligned nanofibers.

FIGS. 5A through 5E include images showing a representative nanocube on a single nanofiber at zero strain (FIG. 5A), a representative incipient neck region showing nanocubes in vicinity of surface crack (FIG. 5B), a nanocube anchored to an underlying neck region at ε=7% (FIG. 5C), representative adhered nanocubes on a neck region at ε=11% (FIG. 5D), and representative adhered nanocubes on a neck region at ε=14% (FIG. 5E).

FIGS. 6A through 6D include images that show a PAN microfiber at zero strain (FIG. 6A), (b) a PAN microfiber at ε=7% (FIG. 6B), (c) a PAN microfiber with adhered nanocubes thereon in a path of an advancing transverse crack at ε=11% (FIG. 6C), and a PAN microfiber with adhered nanocubes thereon at ε=14% (FIG. 6D).

FIG. 7 includes three plots of Raman spectra of PAN fibers that include pristine fibers (plot a), nanofibers after copper nanocube deposition (plot b), and microfibers after copper nanocube deposition (plot c).

FIG. 8 includes an image showing an approximate asymmetric interfacial arrested crack between a copper nanocube and a necked PAN nanofiber at ε=11%.

FIG. 9 schematically represents asymmetric debonding of a cube “island” on a compliant or soft substrate under the assumptions of plain strain conditions.

FIGS. 10A through 10C schematically represent an idealized 2D representation of equilibrium shape of a single cubic copper nanostructure: as a free crystal according to the Gibbs-Wulff theory (FIG. 10A), on a planar substrate as proposed in the Gibbs-Wulff-Kaischew shape theory (FIG. 10B), and on a curved fiber surface (FIG. 10C).

FIGS. 11A through 11H include images showing representative cubic copper nanostructures exhibiting different levels of substrate-influenced truncation on the nanofibers. Scale bars in the images are 300 nm.

FIG. 12 includes a table disclosing dimensions of truncated cubes and corresponding adhesion energies.

FIG. 13 schematically represents palladium and silver seeding processes on electrospun PAN fibers for electroless copper deposition.

FIGS. 14A through 14C include an SEM image showing as-spun PAN fibers (FIG. 14A), and SEM image showing surface topography of an individual fiber (FIG. 14B), and a histogram of fiber diameter distribution (FIG. 14C). Average fiber diameter was 1.017±0.08 μm.

FIGS. 15A through 15D include an SEM image showing high density Ag-seeding obtained from 0.1 M AgNO₃ in an SMR bath (FIG. 15A), an SEM image showing low density Ag-seeding obtained from 0.01 M AgNO₃ in an SMR bath (FIG. 15B), an SEM image showing fine palladium particles on the surface of the PAN fibers (FIG. 15C), and a plot representing Ag-seeding density and particle size versus AgNO₃ concentration in an SMR bath (FIG. 15D). Scale bars are 1 μm.

FIG. 16 includes an array of SEM images that in combination represent evolution of copper nanoparticles on seeded fibers. Scale bars are 1 μm.

FIGS. 17A through 17C include high magnification images of microstructures of copper evolution for PAN—Ag—Cu-15 (H.D.; FIG. 17A), PAN—Pd—Cu-15 (FIG. 17B), and PAN—Ag—Cu-15 (L.D.; FIG. 17C). Scale bars are 500 nm.

FIGS. 18A through 18F include bright field TEM images of a latitudinal cross-section of PAN-Ag (FIG. 18A), a latitudinal cross-section of PAN—Ag—Cu-15 (FIG. 18B), a longitudinal cross-section of PAN—Ag—Cu-15 (FIG. 18C), a latitudinal cross-section of PAN-Pd (FIG. 18D), a latitudinal cross-section of PAN—Pd—Cu-15 (FIG. 18E), and a longitudinal cross-section of PAN—Pd—Cu-15 (FIG. 18F). Scale bars are 200 nm.

FIGS. 19A and 19B include graphs representing x-ray diffraction patterns for copper evolution on H.D. Ag seeded fibers with an inset representing conflated peaks of {111} copper planes and {200} silver planes (FIG. 19A) and Pd-seeded fibers (FIG. 19B).

FIGS. 20A and 20B include plots that represent Raman shifts for pristine PAN, PAN—Ag—Cu-15 (L.D.), PAN—Ag—Cu-15 (H.D.), and PAN—Pd—Cu-15 (FIG. 20A), and decomposition profiles of the metallized PAN fibers for TGA analysis (FIG. 20B; top) and derivative curves with an inset that highlights relative shifts in temperatures at maximum weight loss for the samples (FIG. 20B; bottom).

FIG. 21 includes a table disclosing relative compositions of the constituent materials in the metallized fibers for PAN—Ag—Cu-15 (H.D.) and PAN—Pd—Cu-15.

DETAILED DESCRIPTION OF THE INVENTION

Disclosed herein are systems comprising well-adhered metallic nanostructures on polymeric fibers and methods for producing such systems. The fibers may be produced using various electrospinning techniques. The electrospun fibers may be pretreated with seed crystals of select transition or noble metals to facilitate subsequent nucleation and growth of the metallic nanostructures thereon. The nanostructures may then be produced on the fibers via electroless deposition from aqueous solutions. Although various materials may be used for the polymeric fibers and the metallic nanostructures thereon, the following description will focus on specific embodiments that include copper nanoparticles on polyacrylonitrile (PAN) fibers. However, the methods disclosed herein may be applicable to other materials.

PAN is a hydrophilic polymer that possesses good thermal and chemical stability, and fibers derived therefore are currently used, for example, in various medical and industrial applications. In addition, active polar nitrile groups in the PAN molecule aid the introduction of additional functional groups that support metal (seed) adsorption during electroless deposition. Nominally, the low reactivity of copper precludes the formation of strong adhesion on untreated polymeric surfaces, therefore the fibers may be subjected to a chemical treatment based on alkaline hydrolysis to improve adhesion of the nanostructures thereto.

Suitable but nonlimiting materials for the seed crystals include silver or palladium. Silver and palladium possess affinity for hydrogen, and therefore offer practical exchange current densities for hydrogen gas evolution, promoting rapid anodic reaction kinetics during electroless deposition. For the electroless deposition process, the seeded fibers may be immersed in an aqueous solution that includes, for example, a copper salt, disodium ethylene diamine tetraacetate (Na₂H₂EDTA), and formaldehyde (HCHO; reducing agent) wherein the solution has a pH of about 12.3 to 12.5.

Nonlimiting embodiments of the invention will now be described in reference to experimental investigations leading up to the invention.

In a first series of experimental investigations, systems were produced that included distributed copper nanocubes on aligned polyacrylonitrile (PAN) nanofibers and microfibers. These systems were then tested to analyze the potential of the method and to determine adhesive properties of the nanocubes to the fibers.

For this first series of investigations, nano-scale aligned PAN fibers were electrospun from a 10 wt. % PAN (MW 150,000)/1 wt. % acetone (greater than 99.5%) solution in N,N-dimethylformamide (DMF; greater than 99.8%) using a vertical electrospinner configuration with parallel electrodes. The solution was dispensed through a 24-gauge blunt needle at a flow rate of 0.20 mL/h towards grounded electrodes in a configuration as demonstrated in FIG. 1A. The needle was held at 7.6 kV at a separation distance of 4.5 cm from the substrate. Deposition proceeded for 10 minutes at a time before careful transfer to circular nylon washers (ID: 14 mm; ED: 22 mm) which offered a robust support structure for the electrospun fiber mats. Aligned polyacrylonitrile (PAN) microfibers were prepared from a 12 wt. % PAN/1 wt. % acetone solution. Fibers were electrospun from a 24-gauge blunt needle held at a separation distance of 11 cm from a rotating drum collector, shown schematically in FIG. 1B. Solution flow rate was maintained at 0.30 mL/h for stable electrospinning, with the needle at +2.28 kV and the drum at −2.58 kV. The drum was rotated at 1800 rpm to align the fibers as they were electrospun. Deposition proceeded for 20 minutes before transfer of fiber mats to nylon washers.

The fibers were permanently affixed to the nylon washers with epoxy glue, and allowed to cure for 24 hours. To mitigate or eliminate handling-induced fiber deformation during the deposition procedure, samples were affixed to L-shaped strips wrapped with carbon adhesive tape that enabled easy transportation between baths. The deposition method is schematically represented in FIG. 1C. First, samples were cleaned in a 1.63 M solution of sodium carbonate (Na₂CO₃) for three minutes. Next, samples were immersed in 1M solution of sodium hydroxide (NaOH) at a temperature between 45 and 50° C. for fifteen minutes. The aligned fibers were seeded with silver species to catalyze subsequent copper deposition. 200 μl of ammonia solution (NH₄OH) was added to a 10 ml solution of 0.01 M AgNO₃ under constant stirring. Lastly, a 5 ml solution containing 10 wt. % glucose was added to the solution and stirred for one minute. To prevent possible photocatalytic reduction of silver, the reaction vessel was wrapped with aluminum foil. Samples were immersed in the silver baths for one minute, and subsequently rinsed with a copious amount of deionized water. There was no apparent change in color or translucence of the samples (due to low number density of fibers). A fresh silver bath was prepared for each sample. All seeding baths were operated at room temperature under quiescent conditions. The chemicals used were reagent grade.

For electroless copper deposition, 0.1 g of copper salt (CuSQ₄·5H₂O) and 0.6 g of disodium ethylene diamine tetraacetate (Na₂H₂EDTA) were dissolved and mixed in 20 ml of deionized water. 280 μl of formaldehyde (HCHO) was added into the solution. Drops of sodium hydroxide solution (0.92 g NaOH+20 ml H₂O) were pipetted into the solution to adjust the pH of the electroless bath to about 12.5. The silver-seeded fibers were immersed for 15 minutes at room temperature also under quiescent conditions. The clear blue color of the bath gradually turned pale green, which subsequently became deep green with attendant turbidity signifying homogenous precipitation of copper in the solution. Afterwards, samples were rinsed in deionized water and air-dried. A fresh bath was prepared for each sample.

The samples underwent multiple testing procedures. Uniaxial tensile tests were carried out on a load frame at a cross head speed of 0.5 mm/min under displacement control. The samples (i.e., the washers and fiber mats) were gripped by specialized serrated jaws to prevent slippage during tensile loading. Nominal gage length of the samples was 14 mm. Interfacial adhesion under distinct strain levels of 7%, 11%, and 14% strain was systematically investigated. After each test, samples were extracted and prepared for imaging.

Both electrospinning configurations for fiber alignment, that is, parallel electrode and rotating drum configurations yielded good directionality as shown in FIGS. 2Aa and 2B, for the nanofibers and microfibers, respectively. In addition, both fiber cross-sections were uniform along fiber length, with no observed occurrence of beads. Average fiber diameters and standard deviation for the nanofibers and microfibers were 336±46 nm, and 1.086±0.1031 μm, respectively. Topographically, the nanofibers had a smooth surface (inset of FIG. 2A). In contrast, the microfibers exhibited ridge-like surface relief, likely a result of rapid solvent evaporation (inset FIG. 2B). The orientation distribution for both sets of fibers are shown in FIG. 2C.

The distribution's sharp peak about the median angles (nominally slightly displaced from zero) for both microfibers and nanofibers was indicative of good alignment. Taking into consideration these angular offsets, about 97% of the microfibers were within ±10° of their major axial orientation, while 66% of the nanofibers were within ±10° of their major axial orientation. The greater deviation in alignment observed in the nanofibers can be attributed to a low flexural resistance which is a function of their thin cross-sections, inducing much greater angular offsets than seen in the microfibers. Global alignment was apparent but local curvatures in individual fibers impeded full rectilinear lengths as displayed in the microfibers.

Metastable electroless solutions offer flexibility in compositional control, operational parameters, morphology, and size modulation of nanocrystals in metal deposition procedures. In addition, the necessary immersion of substrates into the solution aids homogenous deposition of discrete nanoparticles or consolidated nanoparticle films for a non-planar substrate geometry as presented by electrospun polymer fibers. FIGS. 3A and 3C show the evolution of well-defined copper nanocubes on the nanofibers and microfibers, respectively. In addition to their slightly truncated edges or rounded corners, the nanocubes had a disperse distribution on the PAN fibers. Average edge lengths of the nanocubes on the nanofibers and microfibers were 137±39 nm and 124±27 nm, respectively. However, this difference was not statistically significant given a P-value>0.05. The cube edges not orthogonal to the fiber surface were used for the average edge length computation. High magnification elevation views (of nanocubes on nanofibers) and planar views (of nanocubes on microfibers) are shown in FIGS. 3B and 3D, respectively.

Although the nanocubes predominantly evolved with a cube face contacting the supporting PAN fibers, the intrinsic cylindrical shape of the fibers altered the original planar geometry of a cubic facet, introducing commensurate curvature. However, this effect was considerably less pronounced on the microfibers due to their greater cross-section in relation to the edge lengths of the cubes. Based on particle count, the nanocube morphology accounted for 60 to 65% of all grown nanostructures on the PAN fibers. The remaining structures were a mix of rose-petal, cauliflower and bulbous shapes that formed as a result of secondary nucleation on the facets of progenitor cubic crystals.

In terms of the thermodynamics of crystal shape and growth, the surface energy of low-index crystal facets typically dictates the resulting crystal habit or morphology. Consequently, based on the energetic sequence y⁽¹¹¹⁾<y^{100}<y^{110}, equilibrium shape of single crystals should either have a full octahedral shape to maximize the manifestation of {111} facets or have a truncated octahedral shape based on the coevolution of {111} and {100} facets. However, solution-based deposition processes enable control over final crystal morphology through selective stabilization of specific crystal facets with designated capping agents or surfactants. With respect to the investigations described herein, the cubic crystal habit, composed of {100} facets, was obtained without the use of exogenous capping agents. While primarily aiding the complexation of copper ions (Cu²⁺) in the electroless bath to mitigate spontaneous precipitation of these species in the bulk of the solution, the carboxylate functional groups present in the EDTA molecule have also been proposed to preferentially interact with the {100} facet, inhibiting growth in the <100> direction relative to the <111> direction, and effectively constraining the final crystal morphology to a cubic shape. Additionally, by virtue of the conformation of the copper nanocube interface to the substrate curvature, it is conceivable that cubic crystal evolution on the fibers was achieved through a concerted stabilization process that involved active surface species derived from prior chemical treatment.

Energy dispersive spectroscopy (EDS) was performed for elemental identification. FIG. 4 shows elemental maps derived from characteristic x-ray lines for Ag (Lα1), Cu (Kα1), and O (Kα1). The elemental maps show that the cubic structures are indeed composed of copper species. In addition, x-ray diffraction data from preliminary studies (not included) identified the nanoparticulate crystalline phase of pure copper. Furthermore, the short deposition time for the silver seeding process ensured the precipitation of sparse and extremely small catalytic silver seeds that were below the resolution limit of the SEM, but were identified in the EDS mapping.

Traditionally, when a metal film is supported by a polymer substrate and the resulting composite structure is subjected to a tensile load, the more compliant polymer material can suppress strain localization, a prelude to delamination, in the metal film ultimately facilitating a congruous deformation behavior. Alternatively, the compliant substrate can “shield” a stiff material by absorbing most of the deformation strain if they are attached as islands rather than conformal films. The copper nanocubes grown on the PAN fibers approximate the latter case. Additionally, electrospun fibers present a distinct geometry as well as special size-dependent micro mechanical deformation characteristics. Because of the expected discrepancy in induced strain coupled with the aforementioned “shielding” effect, integrity of interfacial adhesion was necessarily evaluated in regions with demonstrably high strain concentrations i.e., necks and the immediate vicinities of surface microcracks or tears.

FIGS. 5A through 5E show representative micrographs of copper nanocubes anchored to the surface of the nanofibers at the different strain levels. Firstly, the electrospun PAN nanofibers accommodated multiple neck regions at all induced strain levels, indicating an intermittent occurrence of surface instabilities. However, in contrast to macroscale deformation, extensive propagation of each necked region was restricted by adjacent necks. Distance between necks and neck amplitude, calculated as half the difference between average fiber diameter and fiber diameter at the neck, ranged from 90 to 280 nm and 70 to 300 nm, respectively. FIG. 5A shows a typical copper nanocube anchored to a nanofiber at zero applied strain. FIG. 5B provides a representation of incipient neck formation in the nanofibers including a radial discontinuity or microcrack precedes neck formation and elongation. Moreover, this occurs in close proximity to a nanocube, and it is not improbable that the nanocube is in a high strain field of the fiber substrate. But, this effect might be negated by possible relaxation of surface layer macromolecules aided by their greater chain mobility due to less kinematic hindrance from entanglements. FIG. 5C shows deformation at 7% strain where a copper nanocube is firmly anchored to a visibly necked region, with no apparent signs of delamination. Previous investigations on deformation of a single PAN nanofiber established the onset of plastic deformation to be between 5% and 10% engineering strain. At a higher strain of 11%, FIG. 5D, substantial reduction of the nanofiber cross-section (d of about 186 nm) was observed, coupled with periodic undulation as a result of multiple neck formation. Remarkably, the copper nanocubes maintained in contact, indicating good adhesion despite possibly enhanced interfacial shear from induced strain mismatch. Nevertheless, a closer inspection of the microstructure revealed the presence of an arrested interfacial crack that apparently propagated from a cube edge. Applied strains of 14%, FIG. 5E, did not necessarily translate to thinner nanofiber cross-sections which would have aided evaluation at even greater interfacial shear or at extended interfacial crack lengths. Instead, more necking regions were formed along the nanofiber length. Nonetheless, the copper nanocubes overall exhibited good adhesion, underscoring a robust interface.

Differences in polymer macromolecule configuration in PAN microfibers and nanofibers influences overall micromechanical deformation behavior or features with respect to stress/strain transmission during tensile loading. Chiefly, because their higher aspect ratio makes them more amenable to applied electrostatic drawing effects during processing, nanofibers possess an enhanced chain orientation or alignment along the fiber axis, to a greater extent than microfibers. FIGS. 6A through 6D show representative microstructures of copper nanocubes on microfibers at the different strain levels.

With respect to fiber deformation, the microfibers did not exhibit necking phenomena during tensile loading, in contrast to observations of the nanofibers. Rather, craze formation and attendant transverse tears (or microcracks) from strain accumulation were observed to have been sporadically distributed, approximating bulk deformation behavior. However, slight reductions in microfiber diameter may have been counteracted by elastic recovery prior to SEM imaging. It has been proposed that the greater lateral entanglements in bulk polymers inhibits chain ductility, causing local dilatations that ultimately transform into crazes. Consequently, plastic deformation in single microfibers was substantially reduced or restricted. Based on this limitation in plastic flow, it was theorized that interfacial shear, in comparison to that in the nanofibers, was significantly diminished. Notwithstanding, FIG. 6A shows a representative microstructure of a cube-supporting microfiber at zero strain. A typical microstructure at 7% applied strain is shown in FIG. 6B, where the propagation of a transverse microcrack through an interfacial area was evident. In theory, the presence or evolution of microcracks at an interface portend a weakening adhesion which may lead to delamination or debonding events. Empirically, crack formation proceeds from fiber surface, and then propagates through fiber thickness. At strains of 11%, pronounced interfacial microcrack was observed, with nanocube contact still preserved as represented in FIG. 6C. Remarkable adhesion of contiguous copper nanocubes at the edge of a fully developed microcrack “precipice” derived from an originally intact interfacial area was observed as represented in FIG. 6D, for an applied strain of 14%, signifying well adhered nanostructures and a robust interface.

Given the strong interfacial adhesion of the nanocubes on both fiber architectures, Raman spectroscopy was used to probe possible modifications in fiber surface chemistry after copper nanocube deposition. The obtained spectra are shown in FIG. 7.

The spectrum for pristine fibers is shown in plot (a) of FIG. 7, exhibiting a single sharp, intense peak at a wavenumber of 2240 cm⁻¹. This is also an IR active band, indicative of cyano functional groups. Spectra for both nanofibers and microfibers after deposition are shown in plots (b) and (c) of FIG. 7, respectively. Broad and partially conflated bands at 1350 cm⁻¹ and 1580 cm⁻¹ were observed, respectively attributed to D and G bands of carbonaceous materials. At the same time, the cyano-band was dramatically reduced in both structures. The G-band was indicative of graphitic (sp²) configuration on the fiber surface, suggesting a conversion of the initial linear or aliphatic PAN chains in the surface to cyclical structures. The O-band was indicative of defects or the incorporation of foreign atoms (or molecular entities) in the PAN macromolecules. Altogether, these chemical and structural changes implied the existence of strong chemical bonds at the nanocube-fiber interface, facilitating the strong adhesion observed in both nanofibers and microfibers. It is also plausible that the bottom-up synthesis approach for the cubic nanostructure formation improved chemical interaction between the treated fiber surface and evolving crystals during growth.

The elastic mismatch between a stiff material affixed on a flexible substrate imposes shear stress states at the interface that can induce the nucleation and propagation of cracks. FIG. 8 shows an elevation view of a partially debonded copper nanocube in the necked region with an arrested interfacial crack at applied nominal strains of 11%.

Based on the mechanics of interfacial fracture, a resultant crack length upon debonding can be used to approximate the adhesion energy following an analytical model for the estimation of interfacial fracture energy between stiff island structures and a soft substrate. A schematic for an asymmetric debonding event is depicted in FIG. 9, wherein a single debond crack propagates from one end of an island edge. Equation 1 expresses the accompanying strain energy release rate, G, as a function of the island width, L, the strain applied to the substrate, ε_app, and the interfacial crack length, a, E*^s and E*^f are the plane strain Young's modulus for the substrate and film, respectively.

$\begin{array}{cc}G=\frac{\pi }{16}{\left({\epsilon }_{\mathrm{app}}\right)}^2\left(L-2a\right){\left(\frac{1}{E_s^{*}}+\frac{1}{E_f^{*}}\right)}^{-1}& \left(1\right)\end{array}$

First, the energy release rate of a crack represents the total energy released during crack propagation per unit increase in crack size. In other words, the release rate represents the dissipation of elastic strain energy in a material or at an interface as a consequence of crack growth. Specifically, the marked distinction in the stiffness of the PAN nanofibers and copper nanocubes gives rise to a non-steady state condition wherein the energy release rate is dependent on the length of interfacial crack. As a result, at critical applied strains, the energy release rate firstly attains a maximum value at attendant crack length that are substantially smaller than characteristic island size or dimension. The energy release rate decreases with crack growth until it becomes lower than the interfacial fracture energy at which point delamination or debonding ceases. Because of this relationship between the crack length and energy release rate, an approximation of the adhesion energy can be made using equation 2.

An estimate of the adhesion energy was made based on microstructural evidence for nanocube delamination at applied global strain of 11% as provided by FIG. 8, and the following input parameters: nanocube length of 146±10 nm, reported PAN nanofiber modulus of 3 GPa, copper modulus of 117 GPa, Poisson ratio of 0.3 and a measured crack length, 2a=83±4 nm. The resulting adhesion energy was 0.48±0.04 J/m² where the standard deviation represents uncertainties resulting from cube and crack dimension measurements. The limitations of identifying individual particles which fit this geometry at any given strain made reproducible measurements of this method challenging.

Alternatively, an approximation of the adhesion energy can be obtained through analysis of resultant particle shapes taking into consideration that these shapes are derived from the equilibration of surface free energies with the immediate microenvironment, including the substrate. This method has the advantage of having many particles which can be evaluated under identical deposit ion conditions. Firstly, at a constant volume, the final particle shape should be derived from minimization of the total surface free energy. Accordingly, the Gibbs-Wulff theory states that for a free or isolated crystal at equilibrium, the distance of the center of each bounding facet to an arbitrary central point (Wulff point) within the crystal volume is proportional to the corresponding specific surface energy of that facet. In other words, γ_i/h_i=constant (where γ_i=specific surface free energy of a crystal face i, and h_i=distance of face center to the Wulff point, FIG. 10A). However, crystal surface energetics, and by extension, final equilibrium shape is modified when the crystal is in contact (i.e., deposited or grown) with a foreign substrate. Consequently, incorporation of the influence of the substrate into the Gibbs-Wulff shape theory was addressed in the unified Gibbs-Wulff-Kaischew theory. In brief, crystal shape on a substrate is effectively truncated through its thickness by a measure proportional to the specific adhesion or interfacial energy (P), depicted in FIG. 10B. Therefore, the adjusted proportionality given substrate-particle interaction is expressed in equation 2:

$\begin{array}{cc}\frac{{\gamma }_i}{h_i}=\frac{{\gamma }_j-\beta }{j_{j^{*}}}=\mathrm{constant}& \left(2\right)\end{array}$

where h_j*=distance of the Wulff point to a planar substrate-crystal interface j* taken to be parallel to a crystal plane j, which in turn holds a parallel relationship with the top equilibrium face i. γ_i is the specific surface energy of a face parallel to contact face j*. Consequently, as the adhesion energy increases, crystal truncation increases and vice versa. A useful analogy is the systematic truncation of spherical liquid droplets on solid substrates as the wetting behavior or adhesion increases, reflected by the contact angle in the classical Young's equation. This simple model provides a geometric framework for a quantitative approximation of the adhesion energy of a cubic crystal structure on. a substrate, where, due to its geometric simplicity, the cube distances h_i and h_j* are readily expressed as functions of measurable cube dimensions. However, in the context of the copper nanocubes on PAN fibers, the fiber curvature influences shape of the interface, creating a non-planar geometry, as shown schematically in FIG. 10C. An apparent implication of this curvature is that the contact interface is not strictly parallel to a j^th plane in the cubic nanostructures. Hence, given the relatively shallow curvatures, it was assumed that a proximate crystal plane was tangential to the apex of the interfacial curvature. Consequently, h_j* evaluated from this reference plane was taken as the effective crystal truncation.

For the computation of the adhesion energy, the approximate elevation profiles of copper nanocubes exhibiting distinct levels of truncation on the nanofibers are shown in FIG. 9. These microstructures help to achieve relative accuracy in dimension measurements of the cubic nanostructures.

Previous micrographs provided strong evidence to infer that the equilibrium shape of the copper nanostructures, if isolated or unattached, was a cubic structure. However, in a stricter sense, accurate structural derivation of the equilibrium shape of the free particle from the particle shape as modified by the substrate could only be made if it contained the Wulff point. Otherwise, the crystal shape in general was undefined. As a result, in the characterization of the copper nanostructure on a PAN fiber substrate, a dimension ratio, i.e., B/A, of 0.7 was designated to be indisputably indicative of a cubic structure under the reasonable assumption that the Wulff point was a center of inversion symmetry. In addition, the dimension of the non-orthogonal top equilibrium facet became the effective cube dimension since it remained unchanged as the substrate effect was limited to the through-thickness of the crystal. Based on these assumptions, and using the idealized schematic of the cube growth on the PAN fibers as shown in FIG. 10C, the distance h_j* was expressed as a function of cube dimensions as well as fiber radius. The surface free energy of {100} copper facets was taken as 1.783 J/m². Nanocubes on microfibers were excluded from the truncation analysis due to the ridge-like surface roughness that obstructed clear assessment and evaluation of the interfacial area.

The table of FIG. 12 shows the summary of analyses of copper nanocube shapes on nanofibers, and corresponding adhesion energies as predicted by the Gibbs-Wulff-Kaischew shape theory.

With respect to the energy release rate model for estimation of the adhesion energy, the peculiar substrate geometry of the fibers coupled with the stochastic nature of crystal nucleation precluded the acquisition of more approximate elevation views of cubes exhibiting debonding events as shown in FIG. 8. As a result, the value obtained could not be vetted by rigorous statistical analysis. In addition, while the global strain for the adhesion energy computation was utilized, neck formation and propagation in the nanofibers can considerably increase the local plastic strain rate, and as a corollary, neck strains can be markedly greater than applied strains. Under the assumption of negligible volume changes during deformation, and local strain approximation in the necked region based on the reduction in cross sectional area, adhesion energy of about 84 J/m² was obtained. This unusually high value erroneously subsumed the plastic work which was not accounted for by the model. As a result, the value obtained from the energy release model represent a first order lower bound approximation of the adhesion energy.

For the Gibbs-Wulff-Kaischew shape theory, inaccuracies in the measurement of copper nanocube dimensions represent a source of uncertainties in the adhesion energy quantification. Nevertheless, with the aforementioned factors as qualifications, the adhesion energy as predicted by the strain energy release model was in good agreement with values predicted for styrene-co-acrylonitrile systems and (001) copper facet as obtained with molecular dynamics (MD) simulations (0.51±0.02 J/m²). In addition, the average adhesion energy value as established using the Gibbs-Wulff-Kaischew shape theory was consistent with values obtained for gold delamination on polyimide substrate (about 1 J/m²) using a four-point bend testing, and for copper films with adhesion-promoting titanium interlayer and stressed chromium overlayers on polyimide substrates (about 1 J/m²) using in-situ tensile tests inside a scanning electron microscope. Altogether, these models gave useful approximations of the adhesion energy of copper nanostructures on PAN nanofibers.

In a second series of experimental investigations, systems were produced that included distributed copper nanoparticles on PAN nanofibers to form either continuous conformal coatings or isolated metallic nanoparticles on the surfaces of the fibers. These systems were then tested to analyze the parameters for control over nanoparticle density and the differences between systems formed with silver seed crystals and palladium seed crystals.

For this second series of experimental investigations, PAN powders (MW=150000 g/mol, ρ=1.18 g/cm³) were dissolved in dimethylformamide (DMF) to make a solution with 13 wt. % PAN concentration. Subsequently, 1 wt. % acetone was added to the resulting solution to mitigate or eliminate bead formation in the electrospinning process. The solution was stirred continuously for 24 hours at room temperature. The electrospinning solution was loaded into a 3 mL syringe with a 19 ga (ID=0.8126 mm) needle. The collector plate was aluminum foil situated at a distance of 15 cm from the syringe tip. A voltage generator supplied a constant DC voltage of 15 kV between the collector plate and needle tip creating an electric field strength of 100 kV/m. Flow rate was maintained at 0.34 mL/hr by a syringe pump. Total collection time was eight hours. Finally, the non-woven mats were peeled off from the aluminum substrate, resulting in a mat with a nominal thickness of 30 μm and a total planar dimension of 13 cm×13 cm.

For silver seeding, rectangular coupons with dimension 15×30 mm² were cut from the mat for deposition. To improve handling, ensure planar configuration during immersion in the various baths and mitigate the collapsing effect of surface tension of the aqueous solution after bath emersion, fiber mats were affixed to strips of carbon adhesive tapes. A schematic of the deposition process is shown in FIG. 13. The fiber mats were treated in a 1.63 M solution of sodium carbonate (Na₂CO₃) for three minutes. Next, Alkaline hydrolysis of the samples was achieved in 1 M solution of sodium hydroxide (NaOH) at a temperature between 45 and 50° C. for 15 minutes. For the seeding, two different concentrations of silver precursor (AgNO₃), 0.1 M and 0.01 M, were systematically investigated. A 10 ml solution of both salt concentrations was prepared: to each bath, 200 μl of ammonia solution (NH₄OH) was first added under constant stirring, causing the transient appearance of a brownish color in the original clear solution for the bath with concentrations of 0.1 M AgNO₃, and no perceptible color changes for 0.01 M AgNO₃ concentrations.

Lastly, a 5 ml solution containing 10 wt. % glucose was added to the solution and stirred for one minute. Thereafter, samples were immersed in the silver baths for one minute and then rinsed thoroughly. A progressive brownish yellow discoloration of the sample was observed during immersion in the 0.1 M AgNO₃ baths, while samples in the 0.01 M AgNO₃ bath seemingly maintained the white color of the PAN fibers but acquired a very slight yellow hue after rinsing and drying. All seeding baths were operated at room temperature under quiescent conditions. A fresh silver bath was prepared for each sample. The reaction pathway has been suggested to proceed as follows:

$\begin{array}{cc}{\mathrm{Ag}}_{\left(\mathrm{aq}\right)}^{+}+2{\mathrm{NH}}_{3\left(\mathrm{aq}\right)}\leftrightarrows {\mathrm{Ag}\left(\mathrm{NH}{3}_2\right)}_{\left(\mathrm{aq}\right)}^{+}& \left(3\right)\\ {2\left[{\mathrm{Ag}\left({\mathrm{NH}}_3\right)}^2\right]}^{+}+R-\mathrm{CHO}\left(\mathrm{glucose}\right)+H_{2}O\leftrightarrows 2{\mathrm{Ag}}_{\left(s\right)}^0\left(\mathrm{seeds}\right)+2{\mathrm{NH}}_3+R-\mathrm{COOH}+2H^{+}& \left(4\right)\end{array}$

In this scheme, the silver ions react with NH₃ to form a two-coordinate ammine complex as represented in reaction 3, manifested experimentally as the aforementioned brownish color change. These species are subsequently reduced by glucose to silver seeds as represented in reaction 4.

For palladium seeding, a two-step activation protocol was preceded by sample pre-cleaning and alkaline hydrolysis procedures noted in the previous section. The sensitization bath was a mixture of 5 g of SnCl₂·2H₂O, 8 ml of HCl (AR 37%) and 40 ml of deionized water, and the activation bath was 0.1 g of PdCl₂, 5 ml of HCl (AR 37%) and 20 ml of water. In preparing these solutions, both metal salts were first dissolved in the stated volumes of hydrochloric acid under mild stirring. For the sensitization solution, the stated volume of deionized water was added after the mixture of the tin chloride and hydrochloric acid became transparent, signifying complete dissolution. For the activation solution, addition of deionized water occurred after a brownish turbid bath was observed. The mats were sequentially immersed in these baths for three minutes each. After the activation step, the mats acquired a greenish coloration that gradually turned milky brown upon extended rinsing. Prior to immersion in the electroless copper bath, the fiber mats were treated in an acceleration bath consisting of 1M HCl for three minutes to aid the dissolution of excess tin or tin chlorohydroxide from the surface. Mats were rinsed between each step, and the bath operating conditions were similar to the silver seeding processes. A fresh bath was prepared for each sample. The electrochemical reaction pathway is one of galvanic displacement, wherein the tin ions adsorbed on the fiber surface reduce palladium ions to the metal on introduction in the activation bath and are subsequently replaced by them.

$\begin{array}{cc}{\mathrm{Pd}}^{2+}+{\mathrm{Sn}}^{2+}\to {\mathrm{Sn}}^{4+}+{\mathrm{Pd}}^0& \left(5\right)\end{array}$

For electroless copper deposition, an initial solution containing 0.1 g of copper salt (CuSO₄·5H₂O), 0.6 g of disodium ethylene diamine tetraacetate (Na₂H₂EDTA) and 20 ml of deionized water was prepared. Subsequently, 280 μl of formaldehyde (HCHO) was added into the solution as a reducing agent. Sodium hydroxide droplets from a 1.15 M solution were pipetted into the solution to adjust the pH of the metastable bath to 12.3-12.4. A simplified scheme for the anodic and cathodic half reactions as well as overall the electrochemical reaction for which formaldehyde (HCHO) is the main reducing agent is expressed as:

$\begin{array}{cc}{\mathrm{Cu}}^{2+}+2e\to \mathrm{Cu}& \left(6\right)\\ \mathrm{HCHO}+2{\mathrm{OH}}^{-}\to {\mathrm{HCOO}}^{-}+H_{2}O+1/2H_{2}O+e& \left(7\right)\\ {\mathrm{Cu}}^{2+}+2\mathrm{HCHO}+4{\mathrm{OH}}^{-}\to \mathrm{Cu}+2{\mathrm{HCOO}}^{-}+2H_{2}O+H_2& \left(8\right)\end{array}$

It is noteworthy that formaldehyde is susceptible to disproportionation reactions that give rise to the formation of formic acid and methanol, and the rate of this reaction increases with increasing pH. This may potentially reduce the rate of precipitation in the electroless copper bath. Both silver and palladium seeded fibers were immersed for 5, 10 and 15 minutes, respectively, (i.e., evolution of copper nanoparticles was time-resolved). In general, the bath color change sequence was from clear blue to pale green and lastly to deep green with evident turbidity signifying homogenous precipitation. However, this color change sequence proceeded in a much more rapid manner in the case of the Pd-seeded fibers as compared to the Ag-seeded fibers, accompanied by a bulk solution precipitation of Cu nanoparticles. Afterwards, samples were rinsed in deionized water and air-dried. Henceforth, the cu-deposited seeded fibers will be referred to using the format PAN—X—Cu—Y where X is the seed catalyst and Y denotes immersion time.

An intricate interplay of electrostatic fields, polymer rheology, solution dielectric properties and surface phenomena, electrospinning provided a facile technique for the production of submicron to micron sized fibers. In brief, semi-continuous polymer jets were electrostatically drawn from the needle tip, developing a series of in-flight bending and whipping instabilities that successively create thinner jet cross-sections until they impinge on the collector plate as solid fibers after solvent evaporation. FIG. 14A shows the random, layered arrangement of the as-spun PAN fibers exhibiting complex topology. The fibers had uniform cross-sections with an average diameter of 1.017±0.08 μm. Furthermore, the individual fiber surfaces were relatively smooth as represented in FIG. 14B, exhibiting slight surface asperities that could be attributed to artifacts of the rapid in-flight solvent evaporation.

Silver ions were firstly chelated in the silver mirror reaction, and subsequently reduced to silver atoms/seeds, whereas the direct reduction of palladium ions to palladium atoms/seeds was the electrochemical reaction in the two-step process. In the presence of the PAN fibers, these redox reactions took place preferentially on the fiber surface. Importantly, by virtue of the mat wicking resulting from capillary effects, the working solution was able to reasonably penetrate the complex mat architecture presented by tortuous fibers and random layering. FIGS. 15A through 15C show the microstructure of the seeded fibers. A discrete distribution of deposited seeds was evident for both Ag seeding protocols as represented in FIGS. 15A and 15B, as well as for Pd seeding as represented in FIG. 15C. This system of seed distribution was the consequence of the combined effects of surface sites amenable to nucleation and their stochastic bifurcation into cathodic and anodic sites as is characteristic of electroless processes. All seed morphologies approximated a spherical shape. However, for Ag seeding a marked difference in seed density was observed. For the fibers immersed in an SMR bath with 0.1 M AgNO₃, a high density of fine Ag seeds in close proximity could be observed; while fibers treated in the SMR bath with 0.01 M AgNO₃ showed comparatively finer Ag seeds that were sparsely distributed leaving much of the fiber surface exposed. Furthermore, apparent seed density increased from 23±10 particles/μm² to 47±6 particles/μm², and particle size increased from 29±7 nm to 60±18 nm with an increase in the concentration of the silver salt in the SMR bath. In essence, modulating the silver precursor concentration while keeping invariant the chelation and reducing agent concentration led to fewer apparent nucleation and a smaller seed size on the PAN fiber surface. Henceforth, Ag-seeded samples are referred to as H.D. and L.D. for the high and low seed density cases, respectively. It was also deduced that the incubation time was independent of the seeding density.

In contrast, the Pd seed crystals were ultra-fine, having an average particle size of 5.6 nm (determined from XRD patterns, under the assumption that the crystallite size was about equal to particle size). This ultra-fine nature precluded a quantification of the apparent Pd seeding density, but it was qualitatively estimated to be in excess of seed densities of the Ag-seeded fibers. This was in accordance with general observations of Pd seed densities on tin-sensitized planar substrates. In the two-step activation procedure, hydrochloric acid in the sensitization bath inhibited the hydrolysis of tin (II) chloride into insoluble hydroxotin (II) chloride (Sn(OH)Cl), facilitating the formation of tin chloro-complexes that have been posited as the active reducing species. Consequently, the hydrochloric acid concentration strongly modulated the evolution of catalytic palladium seeds upon activation, controlling particle size and promoting high seeding densities. For instance, for sensitization baths with 30 mL/L and 60 mL/L of HCl, Pd seeding densities of 640 and 1910 particles/μm² were observed, respectively, on planar TiN. The HCl concentration in the sensitization bath was 200 mL/L and it was inferred that the Pd seeding density was correspondingly high.

In regards to the reaction pathway for the activation of formaldehyde (HCHO) in an electroless copper deposition process, in essence, formaldehyde was hydrolyzed to methylene glycol (H₂C(OH)₂, but it takes the methyl diol anion form, H₂C(OH)O—, due to the high pH of the electroless copper solution. The negative charge of this molecule lowered the thermodynamic barrier to the abstraction of hydrogen from the carbon atom (in the C—H bond) necessary for initiation of the anodic oxidation process (reaction 7). The dehydrogenated product was terminally oxidized to formic acid (or formate anions in an alkaline bath) with a corresponding reduction of copper species. Analogously, as has been shown with saturated hydrocarbons, transition metals possess unique chemistries that facilitate the breakage of the C—H bond required for dehydrogenation. In close proximity, the antibonding molecular orbital, σ*_C—H, was populated by electrons from the occupied transition metal d orbital, inducing a weakening of these bonds. At the same time, metal-H interaction/bonding occurs with an attendant overlap of electron population. Based on these processes, Pd and Ag seed crystals facilitate the dehydrogenation of formaldehyde, allowing the nucleation and subsequent growth of copper nanoparticles on the surface of the electrospun PAN fibers. As an advantage, deposition was not terminated when these seed crystals were occluded by deposited copper species given that Cu⁰ also represented adsorption sites for methyl diol anion activation and oxidation. This afforded control over copper growth during deposition on the PAN fibers.

FIG. 16 includes an array of images that represent an evolution of copper nanoparticles on the Ag—and Pd-seeded fibers. The first observation of note was that both Ag seeding conditions have an incubation time of equal to or greater than 5 minutes before Cu begins to deposit, since the microstructures after 5 minutes Cu deposition, images (a) and (g) of FIG. 16, are qualitatively similar to that of their corresponding Ag-seeded fibers, FIGS. 15A and 15B. This delayed deposition of Cu on the Ag seed systems was likely due to time needed for the spontaneous polarization of the equilibrium potentials of the redox half reactions (reactions 6 and 7) to establish a compromise potential at which metal deposition takes place according to the mixed potential theory. Secondly, for high-density Ag-seeded fibers, as the deposition time increased, a progressive increase in Cu coverage from distinct particles/islands to full consolidated films was observed. At ten minutes deposition, PAN fibers were substantially covered with contiguous copper nodules, but distinct patches of the underlying surface were still visible. A fully conformal Cu nanoparticle film on the fibers was not evident until 15 minutes deposition with the sporadic occurrence of copper outgrowths as represented in image (c) of FIG. 16. The hierarchical assembly of Cu nanoparticles was clearly evident at 10 minutes deposition wherein single Cu nanoparticles aggregated into clusters of irregularly shaped nodules on a PAN fiber core.

This subsequently transformed into full conformal and compact copper nanoparticle coatings at 15 minutes deposition as represented in image (c) of FIG. 16. Moreover, the copper nanoparticle films contained intermittent voids or mesopores. One plausible conjecture was that the voids may be attributed to hydrogen gas evolution as predicted by reaction 8, which effectively acts as a barrier to densification in the deposition field. A more probable explanation may be that voids form as a result of geometric misfit between adjacent particles, large enough to be uncovered during the particle/nodule growth process.

The low density Ag-seeded samples maintained a discrete distribution of copper islands on the PAN fibers after 15 minutes of exposure to the copper electroless plating solution. The island size increased from 44±16 nm at 10 minute deposition to 82±23 nm at 15 minute deposition, but still had not formed a compact film. Since the initial particle density was 29 m⁻² (and average particle spacing of about 210 nm, for particle size of 29 nm) more time would be needed to create a compact film. Coverage was modulated by the density of catalytic silver seeds, with the Cu nanoparticles growing exclusively on the silver core seeds, and large areas of the fiber surface were devoid of copper nanoparticles. In addition, copper NP growth evolved to conform to the spherical morphology of the underlying Ag seeds.

The microstructural evolution of Cu nanoparticles on Pd-seeded fibers are shown in images (d), (e), and (f) of FIG. 16. In contrast to the Ag-seeded fibers, a near-complete copper coverage was observed after 5 minutes deposition, signifying that the mixed potential was attained rapidly on Pd seed systems, and this enhanced the kinetics of precipitation, at least before occlusion of the seeds by the first few monolayers of Cu. Thus, with the palladium seeds, the incubation time was much greater than 5 minutes. This was a similar observation to previous studies of the nascent stage of growth of copper nanostructures on palladium-seeded TaN substrates (seed size, 5-10 nm), that showed that continuous copper coverage was obtained after deposition times of 45 s⁻¹ min. The shorter incubation time may be explained in terms of the three orders of magnitude difference in measured exchange current densities for H₂ evolution on palladium with respect to silver. This value, which was 1.3×10⁻⁶ A/cm² for silver, and 1×10⁻³ A/cm² for palladium, would, in principle, correlate with the rate of the aforementioned dehydrogenation process, and was a function of the relatively weaker binding of hydrogen with silver in contrast to palladium (in the classic volcano plots of current density vs free energy of hydrogen adsorption, palladium situates close to the peak, and silver lies close to the lower right leg).

The deposit microstructures were qualitatively similar with increasing deposition time, with greater copper outgrowth at 10 minutes and 15 minutes, as represented in images (e) and (f) of FIG. 16, respectively. The copper nodules on the Pd seeded fibers were much finer than those observed on the high density Ag seeded fibers, and they aggregated into a rocklike, porous film structure. The finer nodules were likely a result of the comparatively higher density of Pd seeds, providing a greater number of nucleation centers that presumably conferred a diffusion-dominated growth process for the copper nodules. So, in summary, the Pd seeded fibers exhibited a higher seeding density and a lower incubation time, leading to the more rapid formation of a compact Cu coating.

FIGS. 17A through 17C show, in greater detail, the resultant microstructures for the three distinct seeding regimens after 15 minute Cu nanoparticle deposition. In particular, the comparison between conformal copper nanoparticle films obtained from Ag-seeding (H.D.) and Pd-seeding, FIGS. 17A and 17B, shows apparent differences in film compactness.

This indicates that hydrogen gas evolution was substantially greater for Pd-seeded fibers, at least in the early stages of deposition. The enhanced H₂ “bubbles” formation created void-filled deposits that served as the “irregular” templates for subsequent deposition. Accordingly, it was expected that the observed roughness should confer an increased surface area. For the low-density Ag-seeding, the nanoparticle islands were non-faceted, suggesting, on the overall, that the deposition process was faster than surface diffusion.

FIGS. 18A through 18F show TEM images of representative cross-sections of PAN-Ag, PAN-Pd, PAN—Ag—Cu-15 (H.D.) and PAN—Pd—Cu-15. Generally, the Ag- and Pd-seed crystals as well as deposited Cu nanoparticles conformed to the substrate geometry, and deposition/precipitation events were restricted to the PAN fiber surface. The cross-sections were consistent with their corresponding SEM images in FIGS. 15A-15C, 16, and 17A-17C, respectively. However, for PAN-Ag, FIG. 18A, finer Ag crystals on the order of 3-4 nm were interspersed between the island particles, previously unobservable in the SEM image of FIG. 15A. Although the majority of Pd seeds were bound to the fiber surface in the PAN-Pd samples, varying degrees of particle embedding (a few nm beneath the fibers surface) was observed. The PAN—Pd—Cu-15 samples, FIG. 18E, did not exhibit such embedding. The distinction in the copper nanoparticle film growth as facilitated by Ag seeding and Pd seeding was clearly observed in FIGS. 18B and 18E, respectively. The PAN—Ag—Cu-15 (H.D.) exhibited a highly compact film structure as is evidenced by the greater mass thickness contrast. On the other hand, PAN—Pd—Cu-15 exhibited loosely packed film deposits with attendant microporosity from inclusion of H₂ bubbles. Corresponding latitudinal cross-sections for PAN—Ag—Cu-15 (H.D.) and PAN—Pd—Cu-15 are shown in FIGS. 18C and 18F, respectively.

Phase confirmation and crystallite size measurements of the metallized fibers were evaluated using x-ray diffraction. FIGS. 19A and 19B show the obtained diffraction patterns of the H.D. Ag-seeded fibers and Pd-seeded fibers, wherein the patterns can be viewed as a superimposition of the individual patterns of PAN and the grown metallic species. For all seeding protocols, peak broadness was indicative of the nanoparticulate nature of the deposited metallic species. The broad peak between 200 and 350 was attributed to the attendant amorphicity of the electrospun PAN fibers. For the H.D. Ag seeding, FIG. 19A, the presence of silver crystals in the seeded fibers was evidenced by characteristic silver peaks at Bragg angles of 38.4°, 44.3° and 64.5° which correspond to the {111}, {200}, and {220} Ag planes respectively (JCPDS Card No. 4-783). For the Pd-seeded fibers, FIG. 19B, characteristic Pd peaks are observed at 40.1° and 46.2° for {111} and {200} Pd planes respectively (JCPDS Card No. 46-1043). However, there were apparent differences in the intensity of the {111} reflections despite identical seeding and measurement conditions. This could have been a result of sample and measurement variability with respect to possible differences in fiber mat packing, slight differences in probe volumes, etc.

Importantly, copper peaks at 43.5° and 50.4° for (111) and (200) Cu planes (JCPDS Card No. 4-0836) became apparent at high deposition times (i.e., at 10 minutes and 15 minutes for H.D. Ag-seeded fibers, and at 5 minutes onwards for the Pd-seeded fibers) corroborating the microstructural observation of a marked distinction in induction/incubation period for copper deposition for both seeding systems. In addition, for Ag-seeding, the evolution of copper peaks at 10 minutes was consistent with observations made with dynamic light scattering of time-resolved aqueous solution precipitation of Cu nanoparticles catalyzed by Ag nanoparticles. In accordance with the attenuation of the overall intensity of silver and copper reflections for L.D. seeded fibers, the higher index reflections were either obscured by background signals or absent. Average silver crystallite size was 12.75 nm and 4.51 nm for H.D and L.D. Ag seeding, respectively, signifying that the Ag particles in FIG. 15 are polycrystalline in nature. For both seeding chemistries, the average size of the Cu nanoparticles was about 9 nm.

Considering the imbuement of the electrospun mats with the active species of the seeding and electroless copper deposition processes, structural modification of the underlying PAN fiber was expected. Raman spectroscopy provides structural characterization of carbonaceous materials. The shifts can be split in two distinct regions: a first order region which spanned shifts between 1100-1800 cm⁻¹ and the second order region from 2200-3400 cm⁻¹. The pristine fibers showed a narrow, sharp peak in the second order region at 2240 cm⁻¹, characteristic of the cyano-functional group present in polyacrylonitrile, which was also an infrared (IR) active band, and a second-order band at 2932 cm⁻¹. Incipient, convolved high frequency bands at 1350 cm⁻¹ and 1580 cm⁻¹, respectively classified as the D and G bands were observed in the seeded fibers (i.e., PAN-Ag (H.D.) PAN-Ag (L.D.) and PAN-Pd). Upon copper nanoparticle deposition, the relative intensities of these bands became more distinctive with increasing deposition times, more so for the Pd-seeded fibers. While the observed evolution of these specific peaks was undoubtedly a result of chemical modifications, it should be noted that localized plasmon resonances of the copper nanoparticles upon laser excitation enhances Raman scattering. Furthermore, it was established that metallic nanoparticle density as well as inter-particle spacing both augment this enhancement. Hence, definitive comparisons of surface chemistry modifications were made using sample instances that have high copper nanoparticle density and distinguishable D and G bands, which corresponded to samples with 15 minute immersion times. The acquired Raman shifts for pristine PAN, PAN—Ag—Cu-15 (H.D.), PAN—Ag—Cu-15 (L.D.) and PAN—Pd—Cu-15 are shown in FIG. 20A.

The G-band originated from “in-plane” E_2g type vibrational mode of overlapping sp²-hybridized carbon atoms, while the D-band or defect band was an “out-of-plane” A_1g type vibrational mode generated by structural disorder or introduction of heteroatoms in the carbon-based structure. Also, the D band was attributed to the breathing mode of sp² atoms in ringed structures. In the context of PAN metallization, both bands were indicative of the possible cyclization of PAN molecules (i.e., the transformation of distinct C—N groups into conjugated systems of C═N bonds) forming a part of connected six-membered cyclical structures. FTIR spectra added evidence in support of this transformation. However, cyclization did not extend into the bulk of the PAN fibers, given that the cyano-bands were still visible in the infrared spectra of these samples (given a PAN refractive index of 1.5, for a mid-IR incident probe wavelength of 2240 cm⁻¹, the theoretical penetration depth was about 0.90 μm which was on the order of the fiber diameter). Additionally, the evidence of surface chain cyclization was further strengthened by observing the manner of evolution of the D and G bands in relation to the cyano band. A marked attenuation of the cyano band intensity was apparent in the Raman shifts of PAN—Ag—Cu-15 (H.D.), PAN—Ag—Cu-15 (L.D.) and PAN—Pd—Cu-15 in FIG. 20A, and this was accompanied by the appearance of D and G-bands, suggesting that the depletion (or more appropriately, conversion) of the cyano-functional group on the surface of the fiber and emergence of aromatized components may have been closely related. In parallel, the band intensity at 2932 cm¹, attributable to methylene stretching, was observed to be dramatically reduced for PAN—Ag—Cu-15 (H.D.) and PAN—Pd—Cu-15 samples. A commonly used metric for quantification of the degree of structural organization/disorganization is a ratio of intensities of the D and G bands (I_D/I_G). As disorganization increases, the ratio increases and vice versa. For PAN—Ag—Cu-15 (H.D.), PAN—Ag—Cu-15 (L.D.) and PAN—Pd—Cu-15, I_D/I_G were 1.25, 0.86 and 1.16, respectively. Accordingly, the order of decreasing disorganization was PAN—Ag—Cu-15 (H.D.)>PAN—Pd—Cu-15>PAN—Ag—Cu-15 (L.D.). Broadly, these ratios allowed for the inference of a high degree of surface disorganization (better termed as a reorganization in this instance) of the PAN fibers, and were roughly proportional to the seeding densities (and attendant Cu nanoparticle coverage) coverage in the respective nanocomposite systems.

Thermograms and DTG curves from the controlled pyrolysis of pure PAN, PAN—Ag—Cu-15 (H.D.), PAN-Ag (H.D.), PAN-Pd and PAN—Pd—Cu-15 are shown in FIG. 20B. In general, thermal decomposition of the samples proceeded in a multi-step manner. Furthermore, weight loss regimes can be grouped under three broad temperature ranges. From room until about 250° C., slight weight losses were observed, attributable to the emission of water molecules and other volatile products. Dramatic weight loss from 250° C. to 330° C. was commonly observed in PAN-containing materials, and was a result of evolution of ammonia gas and hydrogen cyanide accompanied by a thermally induced cyclization of the nitrile groups. The gradual weight loss between 330° C. and 800° C. was attributed to methane and hydrogen gas emissions. However, the chemical treatment processes may presumably alter slightly the constitution of these classic pyrolytic PAN products. A degradation in the thermal stability of PAN—Ag—Cu-15 (H.D.), PAN-Ag (H.D.), PAN and PAN—Pd—Cu-15 was observed in the DTG curves, FIG. 20B, wherein the differential in temperatures at maximum weight loss (peak temperatures) of these samples with respect to the pure PAN fibers was between 12° C. and 30° C. The values are shown in the table of FIG. 21. While the cyclization of the nitrile functional groups as evidenced by the Raman spectra in FIG. 20, has been postulated to enhance the thermal stability of PAN, this effect was nullified and reversed by the exceptional thermal conductivity/transport of the copper and silver nanoparticle films through an enhanced thermal coupling to the underlying PAN fibers during pyrolysis that manifested as a thermal stability degradation. This observation was in contrast to improvements in thermal stability of PAN fiber nanocomposites where the metallic nanoparticles were dispersed in the fiber matrix. However, the PAN-Pd sample exhibited an enhanced thermal stability and this may be attributed to the size-mediated degradation of the thermal conductivity of metallic nanoparticles which occurs when the characteristic dimensions of the body approach the mean free path of conducting electrons. In this case, the Pd particle size (about 5.6 nm) was much smaller than the typical mean free paths of metallic materials (on the order of tens of nanometers). Coupling this effect with the discrete nature of Pd seed distribution (FIG. 15C and FIG. 18D) may induce a less thermally responsive PAN fiber nanocomposite. A summary of the relative composition derived from analysis of the residual char is shown in the table of FIG. 21.

In spite of increased reaction kinetics resulting in very rapid copper deposition from solution as observed previously, the Pd-seeded fibers supported a comparatively smaller amount of copper nanoparticles than the Ag-seeded fibers as seen in the table of FIG. 21. Again, this stems from the hydrogen gas/bubble generation that undermines the formation of compact copper nanoparticle films, limiting the amount of material deposited per layer on the PAN fibers.

The experimental investigations described above supported the potential of the systems and methods disclosed herein and further provided substantial information relating to the properties of the resulting systems. In the first series of investigations, micromechanical deformation of the well-defined copper nanocubes synthesized on PAN nanofibers and microfibers via a solution-phase synthesis method determined a robust adhesion of copper nanocubes on the nanofibers and microfibers. Raman shifts provided strong evidence for chemisorption as the primary anchoring mechanism, and this was believed to be improved by the fact that crystal growth was based on a solution-based deposition protocol. Finally, the adhesion energy computation using the Gibbs-Wulff-Kaischew shape theory and energy release rate model gave useful first order approximation values of about 1 J/m² and a lower bound of 0.48 J/m², respectively.

In the second series of investigations, a comparative analysis of electroless deposition of copper nanoparticles on electrospun PAN fibers as mediated by catalytic seeds of palladium and silver to determined conditions that can create continuous conformal copper nanoparticle coatings or maintain individually isolated copper nanoparticles. Cyclization of polyacrylonitrile macromolecules as a consequence of the Pd and Ag seeding and electroless copper metallization processes allowed for precipitation/deposition. Each seeding route and chemistry conferred distinct features that may be beneficial for certain applications. Using the silver mirror reaction for silver seed deposition created a conformal and discrete distribution of copper nanoparticles on the PAN fibers that can be controlled by modulating the concentration of the precursor silver salt, with apparent densities of 23±10 particles/μm² and 47±6 particles/μm² and corresponding particle sizes of 29±7 nm and 60±18 nm obtained for concentrations of 0.01 M and 0.1 M AgNO₃, respectively. The Pd seed size (about 5.6 nm) appeared to have an even higher density of seeds. The seed density correlated well with the ratio of the D:G band intensities in Raman spectroscopy, implying that cyclization of the surface of the PAN during chemical processing was directly related to the seeding behavior and attendant Cu nanoparticle coverage. Time-resolved electroless deposition of copper nanoparticles helped to establish time thresholds or cut-offs for microscopic and diffraction scattering detection of particle formation, being much greater than 5 minutes on the Pd-seeded fibers and equal to or greater than 5 minutes on the Ag-seeded fibers. For the cases of conformal copper nanoparticle coatings, compact films were obtained for the Ag-seeded fibers, whereas for the Pd-seeded fibers the films were porous and rocky with attendant microporosity. TGA analysis showed a degradation in thermal stability of the metallized fibers when the metallic coating size was greater than the electron's mean free path, but a slight enhancement in thermal stability when PAN was coated with isolated Pd particles with a size smaller than the likely mean free path of the electrons.

While the invention has been described in terms of specific embodiments, it is apparent that other forms could be adopted by one skilled in the art. For example, the physical configuration of the fibers and/or nanoparticles thereon could differ from that shown, and materials and processes/methods other than those noted could be used. In addition, the invention encompasses additional embodiments in which one or more features or aspects of different disclosed embodiments may be combined. Therefore, the scope of the invention is to be limited only by the following claims.

Claims

1. A method comprising: producing polymeric fibers via an electrospinning process; andproducing metallic nanostructures on and adhered to surfaces of the polymeric fibers via an electroless deposition process.

2. The method of claim 1, wherein the metallic nanostructures define discrete distributed nanoparticles on the surfaces of the polymeric fibers.

3. The method of claim 1, wherein the metallic nanostructures define a conformal coating on the surfaces of the polymeric fibers.

4. The method of claim 1, further comprising performing an alkaline hydrolysis treatment on the polymeric fibers prior to the electroless deposition process.

5. The method of claim 1, further comprising pretreating the polymeric fibers with seed crystals of silver and/or palladium prior to the electroless deposition process.

6. The method of claim 1, further comprising pretreating the polymeric fibers with seed crystals of transition or noble metals prior to the electroless deposition process and controlling the resulting metallic nanostructure distribution on the polymeric fibers by controlling the density of the seed crystals.

7. The method of claim 1, wherein the polymeric fibers are polyacrylonitrile (PAN) fibers and the metallic nanostructures are copper nanostructures.