Quantifying Polymer Chain Orientation in Strong and Tough ... · 5/13/2019 · solvent evaporation during electrospinning27 that can act similarly to fast quenching in traditional

Quantifying Polymer Chain Orientation inStrong and Tough Nanofibers with LowCrystallinity: Toward Next GenerationNanostructured SuperfibersDimitry Papkov,†,‡ Nicolas Delpouve,§ Laurent Delbreilh,§ Steven Araujo,§ Taylor Stockdale,†

Sergey Mamedov,∥ Kaspars Maleckis,†,▲ Yan Zou,†,▽ Mohammad Nahid Andalib,† Eric Dargent,§

Vinayak P. Dravid,⊥ Martin V. Holt,# Christian Pellerin,¶ and Yuris A. Dzenis*,†,‡

†Department of Mechanical and Materials Engineering, University of NebraskaLincoln, Lincoln, Nebraska 68588-0526, UnitedStates‡Nebraska Center for Materials and Nanoscience, University of NebraskaLincoln, Lincoln, Nebraska 68588-0298, United States§Departement Systemes Desordonnes et Polymeres, Equipe Internationale de Recherche et de Caracterisation des Amorphes et desPolymeres, Normandie Univ, UNIROUEN, INSA ROUEN, CNRS, GPM, 76000 Rouen, France∥Division of HORIBA Instruments, Inc., HORIBA Scientific, 20 Knightsbridge Road, Piscataway, New Jersey 08854, United States⊥Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, United States#Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States¶Departement de chimie, Universite de Montreal, Montreal, QC H3C 3J7, Canada

*S Supporting Information

ABSTRACT: Advanced fibers revolutionized structural materials in thesecond half of the 20th century. However, all high-strength fibersdeveloped to date are brittle. Recently, pioneering simultaneous ultrahighstrength and toughness were discovered in fine (<250 nm) individualelectrospun polymer nanofibers (NFs). This highly desirable combinationof properties was attributed to high macromolecular chain alignmentcoupled with low crystallinity. Quantitative analysis of the degree ofpreferred chain orientation will be crucial for control of NF mechanicalproperties. However, quantification of supramolecular nanoarchitecture inNFs with low crystallinity in the ultrafine diameter range is highlychallenging. Here, we discuss the applicability of traditional as well as emerging methods for quantification of polymerchain orientation in nanoscale one-dimensional samples. Advantages and limitations of different techniques are criticallyevaluated on experimental examples. It is shown that straightforward application of some of the techniques to sub-wavelength-diameter NFs can lead to severe quantitative and even qualitative artifacts. Sources of such size-relatedartifacts, stemming from instrumental, materials, and geometric phenomena at the nanoscale, are analyzed on the exampleof polarized Raman method but are relevant to other spectroscopic techniques. A proposed modified, artifact-free methodis demonstrated. Outstanding issues and their proposed solutions are discussed. The results provide guidance for accuratenanofiber characterization to improve fundamental understanding and accelerate development of nanofibers and relatednanostructured materials produced by electrospinning or other methods. We expect that the discussion in this review willalso be useful to studies of many biological systems that exhibit nanofilamentary architectures and combinations of highstrength and toughness.KEYWORDS: continuous nanofibers, electrospinning, macromolecular orientation, low crystallinity, size effects in nanofibers,simultaneously strong and tough nanofibers, chain orientation quantification, characterization of subwavelength-diameter nanofibers,nanoscale-related artifacts

Nanomaterials and nanotechnology have the potentialto produce the next step change in materials researchand form the basis for the new generation of advanced

Received: November 15, 2018Accepted: April 30, 2019Published: April 30, 2019

Review

www.acsnano.orgCite This: ACS Nano XXXX, XXX, XXX−XXX

© XXXX American Chemical Society A DOI: 10.1021/acsnano.8b08725ACS Nano XXXX, XXX, XXX−XXX

Dow

nloa

ded

via

NO

RT

HW

EST

ER

N U

NIV

on

May

13,

201

9 at

19:

30:2

2 (U

TC

).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

www.acsnano.org

http://pubs.acs.org/action/showCitFormats?doi=10.1021/acsnano.8b08725

http://dx.doi.org/10.1021/acsnano.8b08725

fibers and composites. Since the discovery of carbon nanotubesand, more recently, graphene and graphene oxide and themeasurement of their extraordinary mechanical properties,intense research effort has been devoted to translating theseproperties to the macroscale. However, so far, the mechanicalproperties of nanocomposites, using these materials, have fallenwell below properties of existing advanced fiber-reinforcedcomposites, and multiple fundamental problems remain.1,2 Attheir core, these problems are associated with the discontinuousnature of these nanoparticles.3,4

Continuous nanofibers (NFs) represent an emerging class ofnanomaterials with critical advantages for structural andfunctional applications.5,6 Continuous NFs are expected topossess enhanced and unusual properties, unavailable in micron-sized fibers, while at the same time alleviating processingproblems associated with discontinuous nanomaterials pro-duced by bottom-up synthetic methods. Electrospinning is atechnique that produces continuous NFs by jetting polymersolutions in high electric fields. The process in its simplest formresults in random nonwoven nanofiber mats. However,techniques for aligned assemblies and individual nanofiberproduction are also available. A unique combination of nano-and macrodimensions in electrospun NFs and relative ease oftheir fabrication, handling, and processing into applicationsexplain rapidly growing interest in continuous NFs for a varietyof applications from tissue engineering7,8 to electronics andoptoelectronics.9−11 Ultrahigh electrospinning draw ratios of upto 5 orders of magnitude, unheard of in traditionalmanufacturing techniques, hold promise of extraordinarymechanical properties. However, to date, electrospun NFswere considered mechanically weak.

Recently, we have reported dramatic size effects in mechanicalproperties of individual electrospun polyacrylonitrile (PAN)NFs.12 Contrary to the classical strength/ductility trade-off instructural materials, nanofibers exhibited simultaneous increasesin strength, modulus, and toughness with the decrease in NFdiameter (see Figure 1). Major improvements were observed inthe ultrafine diameter range (<250 nm). Similar behavior waslater observed in other polymer systems, extending to diametersas low as a few tens of nanometers. Some biological systems withcomparable-sized substructures have also demonstrated combi-nation of high toughness and strength. Such a combination ofsimultaneously high mechanical properties is extremely rare butvery desirable in virtually any engineering application.13 It canreduce the need for overdesign resulting in unnecessary largefactors of safety in current structures. Materials withsimultaneously high strength and toughness are considered tobe “the holy grail” of structural materials research.13 However,despite several important advances, no definitive progress hasbeen yet achieved in this area, and all structural materials that arein use today suffer from a classical strength−toughness trade-off.In conventional advanced polymer fibers, specialized

manufacturing techniques, such as liquid-crystalline and gelspinning, have been developed to achieve high polymer chainalignment, leading to extraordinary strength and modulus.However, polymer chain alignment in these fibers isaccompanied by high crystallinity, leading to linearly elasticdeformation with low strain at failure and toughness. Strain atfailure of all current advanced polymer fibers does not exceed 5%(some developmental fibers, such as carbon nanotube andgraphene-based fibers, as well as biological fibers, such as spidersilk, can exhibit higher strains to failure, but they are generally

Figure 1. Simultaneous increases in strength, modulus, and toughness of individual electrospun PAN NFs. Size effects in (a) modulus; (b) truestrength, and (c) toughness (as defined by the area under the stress/strain curve). (d) Comparison of specific strength and energy to failure inNFs and typical commercial and developmental fibers and materials.20−26 Adapted from ref 12. Copyright 2013 American Chemical Society.The shaded area in panels (a−c) corresponds to the diameter region of highest interest.

ACS Nano Review

DOI: 10.1021/acsnano.8b08725ACS Nano XXXX, XXX, XXX−XXX

B


not classified as advanced fibers due to their relatively lowstrength).In contrast, low and reducing crystallinity was verified

experimentally in electrospun PAN NFs with decreasing NFdiameters.12 Similarly, low crystallinity was observed in otherstudies of PAN14,15 and other electrospun polymer sys-tems.16−19 Low NF crystallinity resulted in persistent elasto-plastic deformation with high (tens to a hundred percent) strainsat failure and ultrahigh toughness. The observed low anddecreasing crystallinity is unusual, as fibers of smaller diametersare expected to have improved polymer chain alignment. This, inturn, should lead to increased polymer crystallinity.One possible reason for this unusual combination of high

degree of chain alignment and low crystallinity is ultrafastsolvent evaporation during electrospinning27 that can actsimilarly to fast quenching in traditional manufacturingprocesses and retard crystallization. The exact range ofcrystallinities in electrospun nanofibers will depend on thepolymer system and nanomanufacturing conditions. Forexample, crystallinities in ultrafine electrospun PAN andNylon-6 NFs produced in our lab were as low as 40 and 50%(compare this to typical 75−95% crystallinities of conventionalhigh-strength polymer fibers). This resulting low crystallinitycan lead to retention of high NF deformability, beyond the yieldpoint, and to ultrahigh toughness. In summary, the observedsimultaneous high strength and modulus are thus attributed tohigh preferred polymer chain orientation. On the other hand,high toughness is the result of low crystallinity (see schematic ofthe structural model of electrospun NFs in Figure 2). Thus, the

discovered properties stem from an unusual structure, resultingfrom inherent and unique peculiarities of the electrospinningprocess, unachievable thus far in conventional manufacturingtechnologies that rely on orders of magnitude slower solventremoval from the 100−1000 times thicker polymer jets.Alignment of polymer chains is expected to improve with the

reduction of NF diameter due to increased draw ratios andconfinement effects. Increases in NF modulus (see Figure 1a)must correlate with improved polymer chain alignment. Directexperimental quantification of orientation is critical for develop-ment of materials and optimization of properties (see schematicin Figure 2). In the past, in-depth studies of structure−propertyrelationships of advanced fibers28−30 led to their extraordinaryproperties and their presence as a dominant force among thestructural materials today. A similar understanding of structure/property relationships in electrospun nanofibers can produce thenext generation of advanced fibers. However, electrospun NFsare 2−4 orders of magnitude thinner than the conventionalmechanically spun fibers, making interrogation of individualnanofilaments extremely difficult. On the other hand, inherentinstabilities of the electrospinning process27 result in difficultieswith preparation of perfectly aligned monodisperse NF bundlesthat could, in theory, overcome the problems associated withexperimental analysis of individual NFs. Due to these difficulties,to date, only a limited number of studies of polymer chainorientation in electrospun NFs was performed31 (see Table S1in the Supporting Information, SI, for a list of prior orientationstudies in electrospun NFs).

Figure 2. Toward the development of a supernanofiber. Manufacturing/structure/properties relationship schematic. The bottom left panelillustrates the critically needed information on polymer chain orientation in ultrafine individual NFs with low crystallinity.

ACS Nano Review


C

http://pubs.acs.org/doi/suppl/10.1021/acsnano.8b08725/suppl_file/nn8b08725_si_001.pdf


The goal of this review is to critically evaluate characterizationtechniques that can provide orientation information for fine NFsproduced by electrospinning or other nanofabrication methods(see bottom left panel in Figure 2). The discussion is alsorelevant to biological materials with nanofibrillar architecture.Several traditional as well as emerging orientation character-ization techniques are reviewed. Their applicability forexamination of fine NFs is demonstrated and analyzed onexperimental examples that were conducted specifically for thepurpose of this paper. It is shown that each of the describedtechniques can provide valuable information but also faceunique challenges. In the discussion below, the methods aregrouped into those that require bundles/mats of NFs and thosethat can be used on individual NFs. Advantages and limitationsare discussed within the context of ultrafine-diameter NFs.Polarized Raman microscopy, as a technique with highimmediate potential, is examined in-depth. Possible artifactsrelated to applications of different techniques to sub-wave-length-diameter NFs (i.e., simultaneously strong and tough NFsof the highest interest for structural applications) are describedin-depth for the first time, and ways to overcome them areproposed. Although the artifacts are described on an example ofpolarized Raman method, the discussion is relevant to otherspectroscopic techniques, as well. Finally, future outlook andoutstanding issues are discussed.

METHODS OF ORIENTATION EVALUATION THATREQUIRE NANOFIBER MATS OR BUNDLES

X-ray Diffraction. X-ray diffraction (XRD) relies on theinteraction of X-rays with crystalline material. Diffractionpatterns obtained are then used to recreate the three-dimensional arrangement of atoms within the material crystals.XRD can be used to estimate the degree of crystallinity insemicrystalline polymers. The interaction between X-rayradiation and the material is generally weak. As a result, evenradiation-sensitive materials such as polymers can be examinedby XRD without significant damage. However, due to this weakinteraction, the observed signal is also relatively weak.Consequently, XRD requires a relatively large amount ofmaterial and cannot generally be used to examine individualNFs (see discussion on potential use of X-ray microscopy inindividual NF studies described below). Random NF mats canbe used for the quantification of polymer crystallinity,12 whereasoriented bundles can generally be used for orientationstudies.32−35

An example of the application of XRD to examine crystalorientation in PAN NF bundles is shown in Figure 3. Bundleswith different average diameters were examined (see Figure 3a,dfor scanning electron microscopy (SEM) images). Diameterdistribution (see Figure 3b,e) and relative degree of fiberalignment within the bundle (see Figure 3c,f) were constructed.

Figure 3. Evaluation of crystal orientation using XRD. (a,d) SEM images of the examinedNF bundles with two different average diameters. (b,e)NF diameter distribution within the bundles. (c,f) Orientation distribution of NFs within the bundles. (g,h) Two-dimensional XRD images ofthe different bundles in panels (a) and (d). The left part of the diffractograms is for as-spun bundles, and the right part is for the same bundlesannealed at 130 °C at constant bundle length. (i) Extracted azimuthal scans for the arcs corresponding to the crystal peaks. The intensity isintegrated between the two concentric circles, as shown in panel (g).

ACS Nano Review


D


A 2DXRDpattern for the bundles was recorded (see the left partof Figure 3g,h).Consistent with the previous results,12 crystallinity in as-spun

bundles was low (as evidenced by the diffuse arcs in the left partof Figure 3g,h). The 2D plots can generally be used toqualitatively compare the crystal orientation in the differentsamples. For quantitative examination of crystal orientation, anazimuthal scan of intensity as a function of angle from the bundleaxis can be obtained (see Figure 3i) by integrating the intensityunder the crystalline reflection at 2θ ∼ 17° (see the area markedby two concentric circles in Figure 3g). The degree of preferredcrystal orientation is then sometimes expressed by evaluating thefull width at half-maximum (fwhm) of the peaks correspondingto the arcs in the XRD images. This approach is commonly used,for example, for examining commercial carbon fibers.36 Asecond-order parameter of the orientation distribution function,also called the Herman function, ⟨P2⟩, can also be extracted andanalyzed.37

As seen in the left parts of Figure 3g,h, preferred crystalorientation was observed in as-spun NF bundles. However, lowcrystallinity led to very low signal-to-noise ratio (see Figure 3i).As a result, evaluation of fwhm corresponding to the crystal arcswas difficult.Crystallinity of polymer NFs can be increased by annealing.

The right part of Figure 3g,h shows the diffractogram for thebundles annealed at 130 °C. More pronounced preferred crystalorientation (as expressed by sharper peaks in Figure 3i) isevident. The fwhm can generally be computed for the crystallinearcs in the annealed samples. However, crystallization is acomplex phenomenon. It is impacted by additional parametersbeyond the degree of macromolecular alignment in as-spunNFs.As a result, caution should be exercised in relating theinformation on the crystal orientation in the annealed samplesto the original chain orientation in largely amorphous as-spunNFs.Two additional effects complicate the relationship between

chain orientation in the NFs and the one observed in XRD. First,as can be seen in Figure 3c,f, NFs are not perfectly aligned alongthe bundle axis. Consequently, the orientation seen in XRD is aconvolution of NF orientation within the bundles and of crystalswithin individual NFs. The second effect is that of NF diameterpolydispersity (see diameter distributions in Figure 3b,e). Thedegree of preferred orientation is expected to be different forNFs with different diameters. As a result, NF diameterpolydispersity will confound the interpretation of XRDorientation results.

Polarized Fourier Transform Infrared Spectroscopy.Polarized Fourier transform infrared (FTIR) spectroscopy isanother technique to investigate orientation in polymers. Eachinfrared band corresponds to a specific vibrational mode withinthe polymer structure, thereby providing orientation informa-tion with molecular selectivity. Jasse and Koenig38 provided adetailed description of the use of polarized FTIR for quantitativecharacterization of polymer chain orientation. The polarizedabsorption of a given IR band is proportional to the square of thedot product between the transition dipole moment vector andthe electric field polarization vector of the incoming radiation.For a simple localized vibration such as the nitrile stretchingband of PAN, the transition dipole moment is along the bondaxis. Consequently, orientation information can be obtainedfrom dichroic ratio (ratio of absorption intensities oforthogonally polarized light).39,40 When the conformation ofthe molecule is known, the dichroic ratio can be converted into aHerman orientation function, ⟨P2⟩ (see details in the SI).Contrary to XRD, polarized FTIR can provide information on

orientation in both crystalline and amorphous phases. However,it is still limited in its applicability to NFs. In particular, theminimal spot size achievable by focusing the beam using a far-field microscope is on the order of 5−10 μm because of the largewavelengths of IR radiation. This strongly reduces the signal-to-noise ratio for individual NFs because their diameter issignificantly smaller than the spot size and leads to quantificationerrors due to stray light.41 As a result, similarly to XRD, FTIRmeasurements are normally conducted on large bundles or matsof NFs using a large spot size of several hundreds ofmicrometers.32,33,42−44

An example of polarized FTIR examination was carried onPAN NF bundles. In the case of PAN, orientation informationcan be obtained by examining the bands associated with thenitrile group and the different CH vibrations.45 There are severaloverlapping peaks in the CH region, which require band fittingand assignment. On the other hand, the nitrile stretching band isprominent and is isolated from other bands. As a result,evaluation of the infrared dichroism of the nitrile stretchingmode is the most common approach. The molecularconformation of PAN is that of an irregular helix.46 The anglesbetween the nitrile groups and the axis of this helix are notconstant, but several studies have estimated the average α angleto be in the vicinity of 70°.39,47 As a result, if preferred chainorientation is present, the absorbance with the polarizationparallel to the fiber is expected to be smaller than the one withpolarization perpendicular to the fiber, resulting in dichroic

Figure 4. Polarized infrared spectra of bundles of PAN nanofibers. (a) Original transmission spectra for the bundle prepared from the 8%solution, showing large scattering. (b) Baselined transmission spectra in the nitrile stretching band region showing anomalous dispersion andincorrect relative intensities. (c) Transflection spectra of fibers prepared from 8 and 11% solutions and normalized using a band from residualdimethylformamide.

ACS Nano Review


E



ratios less than 1. However, because the α angle is significantlysmaller than 90°, the dichroic ratio of the PAN nitrile band is notexpected to be very small even in the presence of a significantdegree of preferred backbone chain orientation (R isapproximately 0.32 for ⟨P2⟩ of 0.9). Herman orientation factorwas extracted in the past for uniaxially drawn PAN films.39,47

Examination of oriented electrospun NF bundles was attemptedas well (see discussion below).15

Oriented bundles of NFs with different average diameterswere examined in this work by polarized FTIR in transmission(suspended fibers) and in transflection (fibers on an aluminumfoil) modes. The transmission FTIR spectra shown in Figure 4afor the bundle with an average NF diameter of 500 nm present avery pronounced baseline drift toward higher wavenumbers dueto light scattering (similar results are obtained by transflection).Such scattering is commonly observed for samples dispersed in amatrix, such as KBr pellets, due to a refractive index contrastbetween the sample and the matrix. It is maximal at wavelengthssimilar to the size of the particles. In the case of nanofibers,scattering is due to the refractive index mismatch between thefibers and the air between fibers. It therefore depends on theNFs’ diameter and average distance between NFs in the bundle.The scattering amplitude is systematically larger for lightpolarized parallel to the fibers. This could be explained bylarge differences in the spatial variation of refractive indexbetween the fibers and air (amplitude and periodicity) with lightpolarized along or perpendicular to the fiber bundle axis, as wellas differences in the curvature of the objects being probed witheach polarization (more details will be given below).Figure 4b shows a zoomed-in image of the nitrile stretching

band after baseline correction. Surprisingly, and contrary towhat was expected based on the known conformation of PAN,the intensity of the nitrile band was stronger with light polarizedparallel to the NF bundle than with the incoming light polarizedperpendicular to the bundle. The dichroic ratio for the nitrileband varied between 1.2 and 1.9 for all samples, either intransmission or in transflection, whereas values smaller than 1were expected. In addition, the carbonyl stretching band at 1666cm−1, which is due to the presence of residual dimethylforma-mide (DMF) in the fibers that is usually expected to be isotropic(see qualification below), is also more intense in the parallel-polarized spectra. This indicates that spectral distortionsstrongly affect the quantitative and even qualitative informationin the spectra of PAN nanofibers.It can also be seen that the nitrile band has a pronounced

derivative shape that distorts the expected band shape and leadsto an apparent shift of the maximum intensity by 2 cm−1 towardlow wavenumbers with parallel polarization compared toperpendicular polarization. This derivative-like shape is due tothe reflection of the IR beam at the fiber surface in addition toabsorption by the sample. The absorption depends only on theimaginary part of the complex refractive index of the sample (theabsorption index, k), whereas the reflection depends on both kand the real part (the refractive index, n) of the complexrefractive index. The “anomalous dispersion” of n (n is lower onthe high-wavenumber side of the band and is higher on the low-wavenumber side) affects the fraction of reflected light andprovokes the derivative-like band shape. As a consequence, theapparent absorbance does not simply depend on the absorptioncoefficient of the fibers as one would expect. In fact, suchdispersive band shape can be seen in most IR spectra of fiberbundles reported in the literature.15,48−51 The shift in bandposition between polarized spectra can also be observed in some

cases when the spectra are shown at a scale allowing theobservation.52,53 This emphasizes that the spectral distortionsreported here for PAN fibers are common when probing NFbundles by FTIR and can affect the quantified orientation valuesreported.In an attempt to quantify the orientation in spite of the

spectral distortions, the intensity of the nitrile band wasnormalized with respect to that of DMF. The transflectionspectra were used because of a smaller effect of anomalousdispersion (see Figure 4c). When doing this, orientationparameters ⟨P2⟩ of 0.14 ± 0.03 and 0.33 ± 0.04 were obtainedfor the samples with average diameters of approximately 750 and500 nm, respectively. These results demonstrate that properlyanalyzed FTIR data can provide useful information on polymerchain orientation in NF bundles. However, it must be noted thatthe normalization procedure relied on the assumption that theresidual solvent molecules are randomly oriented. In the past,Bashir et al.39,54 showed that some solvent molecules can formpolymer/solvent complexes with PAN. As a result, solventmolecules were found to have preferred orientation. If this is truefor DMF, the normalization procedure will yield incorrectresults, and the spectra cannot be corrected.

Differential Scanning Calorimetry Combined withDielectric Relaxation Spectroscopy. A combined differ-ential scanning calorimetry (DSC)/dielectric relaxation spec-troscopy (DRS) examination is a powerful tool that can provideinformation on polymer structure. DSC examines changes inmaterial heat capacity as a result of phase transformations,whereas DRS looks at changes in dielectric properties caused bythe same transformations.DSC is widely used to evaluate crystallinity of polymers and

other aspects of polymer structure. A traditional two-phasemodel of semicrystalline polymers is often insufficient toproperly describe the polymer structure. The amorphousphase of the polymer can be subdivided into rigid and mobilefractions (RAF andMAF, respectively). RAF, which results fromincomplete decoupling between the crystal and amorphousphases, is conceptually similar to an interphase in modernmicrostructural models of composites. In addition, a phase withan intermediate level of order between the crystal andamorphous phases, called mesophase, can also appear.DSC can also be used to gain information on polymer chain

orientation by looking at the polymer glass transition. Changesin glass transition temperature and the shape of the stepobserved in DSC reflect differences in mobility of themacromolecular segments. In the absence of changes inchemical structure, these differences are usually associatedwith changes in supramolecular structure such as degree ofcrystallinity and macromolecular alignment. Many studies55−59

reported broadening of the glass transition when the amorphousphase mobility was hampered by crystalline growth ororientation.DSC andmodulated temperature DSC (MT-DSC) have been

used to investigate the microstructure of electrospun fibers inseveral polymer systems.60−68 The presence of both RAF andmesophase was observed.62,69 The latter is likely due to the non-equilibrium nature of the electrospinning process, whichproduces microstructures with the amorphous phase sufficientlyoriented to allow the creation of a mesophase but prevents theformation of crystals.Quantification of molecular mobility in highly oriented low-

crystalline systems is very challenging. Relaxation spectroscopy,more specifically, DRS, has been shown to be capable of

ACS Nano Review


F


investigating the evolution of local and/or delocalized molecularmobility during chain orientation processes.57 These processeshave been associated with the appearance of oriented crystallinephases during strain-induced crystallization and orientation ofamorphous chains.70 DRS monitors the relaxation processesassociated with local or cooperative molecular mobility inamorphous fractions.71 DRS is also sensitive to the presence ofinterfaces between mobile (mainly amorphous) and crystallinephases. In polymers, these interphasic responses are classicallyattributed to RAF.Few studies reported the use of DRS for investigating

molecular mobility in electrospun nanofibers. The amorphousfraction in NFs is expected to be highly constrained by both themacromolecular orientation and the presence of a mesophase62

that can act as the RAF. Estimating the mobility of the mobileamorphous fraction (MAF) is challenging due to the presence ofRAF. This limitation is similar to studies of bulk polymers wherethe growth of constrained amorphous phase (RAF) isaccompanied by a gradual disappearance of theMAFmobility.72

In addition, due to the difficulty of investigating individual NFs,most of the studies are carried out on NF mats (see Figure 5a).NF mats have a large amount of interfaces between the fibers,leading to Maxwell−Wagner−Sillars interfacial polariza-tion.73−75 This effect increases the complexity of studyingmolecular mobility relaxation processes (orientation polar-ization effects), especially at low frequencies. In order tosuppress this effect, we removed the Teflon film in ourdemonstration experiments, which is otherwise usually posi-tioned at the sample/sensor interface.An example application of a combined DSC/DRS was

performed on random PAN NF samples with different average

NF diameters. Additional information on the experimentalprocedures and data reduction for the DSC and DRSmeasurements can be found in the SI.Different samples exhibited pronounced differences in the in-

phase component C′ of the complex heat capacity behavior inthe glass transition domain localized around 80 °C (see Figure5b). Analysis showed that the glass transition temperatureincreased from 78 to 87 °C with the decrease of the average NFdiameter. Furthermore, the glass transition was significantlybroadened until the heat capacity step became invisible and wastransformed into a continuous variation of the heat capacity withtemperature (sample with the smallest average diameter of 232nm). This behavior is the signature of a hindered relaxationdynamics due to increased level of constraints in the amorphousphase for the sample with the smallest average NF diameter.Although such a change in material behavior can also be causedby an increase in polymer crystallinity, we know from previousexaminations that the crystallinity of PAN nanofilamentsdecreases with the decrease in their diameter. Furthermore,the mat with the average fiber diameter of 232 nm exhibited thehighest heat capacity. It is the signature of a higher content ofamorphous phase relaxing at the glass transition. Thus, thisresult supports the assumption that reduction of the fiberdiameter is accompanied by a decrease in polymer crystallinity.Consequently, the change in this case is caused by increasedmacromolecular alignment. It is important to note thatconstraint due to increase in surface-to-volume ratio for smallerNFs (confinement) should lead to the opposite effect of reducedTg.This conclusion is supported by the examination of the real

and imaginary part of the permittivity obtained from DRS

Figure 5. MT-DSC/DRS examination of PANNFmats with different average nanofiber diameters. (a) Fiber mat of an average fiber diameter of232 nm before the sample was fixed between DRS electrodes. (b) MT-DSC in-phase (C′) component of the complex heat capacity versustemperature. (c) Imaginary part of the complex permittivity (ε*(ν)) as a function of frequency and temperature for a fiber mat with average NFdiameter of 785 nm. (d) Temperature fluctuation associated with the α-relaxation from MT-DSC and DRS measurements.

ACS Nano Review


G



measurements. DRS classically provides the signature of the α-relaxation (dielectric manifestation of the glass transition) attemperatures above the calorimetric glass transition temper-ature. In Figure 5c, it appears as a peak on the imaginary part, forwhich the position on the temperature axis is dependent on theapplied frequency. Coupling DRS with calorimetry techniques,such as MT-DSC or fast scanning calorimetry (FSC),76 canextend the range of investigation for the temperature depend-ence of the relaxation time. Temperature fluctuation, δT,associated with the α-relaxation is a signature of the distributionof the relaxation times in the amorphous phase. It increases withtemperature but can also vary due to structural causes, reflectingstructural differences between samples when examined atcomparable temperatures. As can be seen in Figure 5d, theinitial difference in temperature fluctuation observed at the glasstransition from MT-DSC disappears at higher temperaturesrecorded in DRS. As the melting temperature of PAN is about300 °C, this change could not be caused by the disappearance ofthe crystalline phase. However, it might be attributed to theconstraint slackening in the initially oriented structure (entropicrelaxation of polymer chains) when crossing the glass transitiontemperature.Therefore, this set of results indicates that electrospinning

induces orientation of macromolecules, which becomes morepronounced in samples with smaller NF diameters. Thisorientation significantly impacts structural dynamics aroundthe glass transition. At higher temperatures, the orientationgradually vanishes, as macromolecules becomemore flexible andstress relaxation occurs. Overall, the example shows that changesin relaxation behavior around Tg can provide information onpolymer chain alignment that complements results obtained byother methods. Relaxationmechanisms related to chain mobilityare inherently mechanical in nature and can therefore providevaluable insight into mechanisms of NF mechanical behaviorand toughness. Analysis of Tg is especially useful for polymers inwhich the melting temperature is higher than the onset ofthermal degradation, as is the case with PAN.Nuclear Magnetic Resonance Spectroscopy. Nuclear

magnetic resonance (NMR) spectroscopy is a spectroscopictechnique that can provide detailed information on molecularstructure, conformation, and dynamics (relaxation). In thistechnique, the sample is placed in a high magnetic field and isexcited by a radio frequency pulse from the ground state to anexcited state. The frequency required to flip the nucleus from theground to the excited state is characteristic of the nucleusobserved and is sensitive to the local electronic environment(shielding vs deshielding). The most common NMR techniquesinvolve nuclei with spin I = 1/2, such as 1H and 13C. Nuclei withspin I = (2n + 1)/2, such as 17O and 27Al, can also be examined.Nuclei with spin I = 0 (12C, 18O) are not NMR-active.Solid-state NMR (SSNMR) is less developed than solution

NMR and can suffer from sensitivity and technologicalproblems. However, it emerged as an important technique forstudying polymer structure. The technique has severaladvantages, such as the ability to study insoluble materials andconformations not attainable in solution over the solution-basedNMR.77

In general, SSNMR can be used to study important aspects ofsolid-state polymers such as internuclear distance, atomiccoordinates, backbone torsion angle, crystallinity, and the mixof different crystal structures, as well as orientation in both thecrystal and the amorphous phases.77,78 NMR interactions suchas dipolar (between nuclei with spin I ≠ 0), quadrupolar (nuclei

with spin I > 1/2), scalar coupling (indirect interaction betweentwo spins), and chemical shielding are inherently anisotropic.This anisotropy has been used to study orientation in drawnfilms and fibers.77,78 Multiple approaches exist. The mostpopular uses a chemical shift anisotropy tensor of rare nucleisuch as 13C79−81 and 15N,82−84 although naturally abundantnuclei such as 19F were also used.85

The chemical shift, observed as the different peaks on theNMR spectrum due to chemical shielding, is a second-ordertensor. The principal tensor components can generally beobtained from a static examination of an isotropic sample (suchas a powder sample).78,86 Some techniques also require an“isotropic” value of the chemical shift to analyze the differencebetween this parameter and the observed chemical shift values atdifferent fiber/magnetic field orientations. The isotropicchemical shift can be obtained either from averaging theprincipal components of the tensor or by carrying out magicangle spinning (MAS) (rapidly spinning the sample at the angleof 54.74° relative to the applied field) experiments on thepowder sample. Once the principal components of the tensor areknown, several approaches to obtaining orientation informationexist.The comprehensive approach analyzes and fits line shapes of

the chemical shift as a function of the angle between themagnetic field and the fiber/draw axis to obtain multiple-orderorientation parameters.85,87,88 Such an investigation is time-consuming and complex. Alternatively, in a simplified approach,measurements can be done either in one (parallel)82,84 or two(parallel and perpendicular)83 configurations. To obtainorientation information, Gaussian distribution of the orientationfunction is generally assumed. Fitting the “parallel” spectrumproduces a set of possible Euler angles for the rotation of thechemical shift tensor from the principal component axis systemto the fiber axis system. Additional fitting of the “perpendicular”spectrum and error minimization procedures are needed toobtain orientation parameters as expressed by the width at half-maximum of the orientation distribution. This orientationparameter does not relate directly to structural features but to anaverage angle between the main axis of the principal componenttensor and the fiber systems. However, if molecularconformation is known, this orientation parameter can beconverted into a distribution of bond angles in the fiber.83

Although these approaches are less complex than thecomprehensive one, they still require multiple spectraacquisition, examination of powder samples, and relativelycomplex fitting procedures.A further simplification is possible. It has been shown that the

chemical shift in the configuration where the magnetic field isparallel to themajor sample axis linearly correlates to orientationparameters obtained by other technique.79 Consequently, it canbe used as an indicator of orientation for comparative studieswithout obtaining the chemical shift tensor components.The different orientation studies were so far demonstrated for

drawn films,85,87 conventional,84,89 and biological90,91 fibers.SSNMR has also been used to determine conformationalchanges in electrospun silk fibers.92 In principle, the above-described approaches can be extended to the study oforientation in NFs. However, to our knowledge, such studieshave not yet been conducted or published.SSNMR is generally less sensitive compared to the solution

technique and can suffer from lower signal-to-noise ratio.77 Insolution, anisotropic NMR interactions are averaged over thetime domain due to rapid random tumbling for small molecules

ACS Nano Review


H


and rapid segmental motion for highmolecular weight polymers.The nuclear dipole−dipole interactions average to zero, whereaschemical shift anisotropy averages to a nonzero value. Thisaveraging produces sharp, well-resolved peaks. On the otherhand, in solids, molecular motion is restricted, and theanisotropic interactions as well as the presence of multiplepolymer conformations can cause severe line broadening.77 Thisproblem can become more acute for samples with lowcrystallinity and amorphous structures as well as in the caseswhere different chemical shifts are not sufficiently spectrallyseparated.77

Several approaches exist to partially remove the linebroadening in SSNMR. MAS produces a similar effect tomolecular motion in solution. MAS speed requirement in orderto suppress line broadening changes depending on theinteraction. However, it is generally on the order of severalkilohertz and increases linearly with the increase in themagnitude of the magnetic field used in the NMR system.77

This can lead to technological problems of spinning the sampleat the required speed. In addition, quadrupolar interaction isonly partially averaged byMAS and leaves line broadening in thecases where this interaction exists.Multiple pulse sequences are used for dipolar decoupling

(both for homo-93 and heteronuclear94 interactions). Cross-polarization (technique where polarization from abundant spins

such as 1H or 19F is transferred to dilute spins such as 13C or 15N)is used to enhance the signal-to-noise ratio.77 In addition, the useof naturally rare nuclei such as 13C or 15N removes linebroadening due to homonuclear dipolar coupling as thedistances between the adjacent studied nuclei are very large.Specific site or general labeling, using rare isotopes, can also beused to improve sensitivity and signal-to-noise ratio. Morerecently, cryogenic probes have gained use in NMR. Suchprobes reduce thermal noise and improve the signal-to-noiseratio, allowing for shorter (up to four times) acquisition times.95

An example of SSNMR application to PAN powder can befound in the SI. In solution, the chemical shift of the carbon inthe nitrile group appears as a sharp singlet.96 In contrast, theSSNMR spectrum in our experiment is very broad. Theobserved broadening is likely the result of convolution of thesignals from the crystal and the amorphous phases in thesemicrystalline polymer. A series of experiments on samples withdifferent polymer crystallinities with and without MAS couldhelp identify different signals. In particular, examination of a fullyamorphous and a highly crystalline polymer would be of mosthelp. Labeling the samples (adding to the natural abundance ofthe nuclei).92,97,98 can significantly improve sensitivity of thistechnique. However, possible effects of labeling on theelectrospinning process and resulting nanofibers need to becarefully examined.

Figure 6. Effect of NF diameter polydispersity on the relative volume associated with each diameter within the distribution. Synthetic diameterdistributions: (a) unimodal normal distribution with average diameter of 450 nm and a standard deviation of 120 nm; (b,c) bimodal normaldistribution with average diameters of 450 and 700 nm and a standard deviations of 120 and 250 nm at different mode mixes. Insets in (a−c)show changes in cumulative distributions when changing from distribution by number to a distribution by volume. Dashed lines show the shiftin the median diameter between the number and volume distributions. (d) Real diameter distribution measure for a NF mat. The inset showsthe relative cross-sectional area for NFs from a 100 nm and up to a 1500 nm diameter.

ACS Nano Review


I



Limitations of Techniques Requiring Bundles ofNanofibers. The techniques described in this section requirerelatively large material volume to produce good signal-to-noiseratios. As a result, these techniques rely on examination ofbundles to extract average orientation information for electro-spun NFs. Two problems complicate the analysis, in addition tothe experimental artifacts described in the previous sections.First, in the case of XRD, FTIR, and SSNMR, aligned NF

bundles are needed. Alignment of the NFs within the bundles isnot perfect (see examples of NF orientation distribution withinthe bundles in Figure 3c,f). The above techniques cannotdistinguish between NF orientation and the orientation ofpolymer chains and crystals within the NFs. As a result,misorientation of NFs within the bundles introduces significantmeasurement errors, leading to underestimation of theorientation of the structural features. The situation is furtherexacerbated when the degree of NF alignment in the bundleschanges between samples, preventing even comparative studies.Generally, misalignment of NFs in the bundles increases withthe decrease in NF diameter. Consequently, some of theexpected improvement in the degree of preferred orientation infine NFs can be masked by increased NF misalignment. Theproblem in the case of SSNMR is even more severe as largeramounts of material are needed.Second, within the bundle, the NF diameter is not

monodisperse. The level of polydispersity depends on samplefabrication parameters and will generally vary with the averageNF diameter in the bundle (see, for example, Figure 3b,e).Larger NFs within each bundle occupy larger volume.Consequently, most of the signal comes from thicker nanofila-ments (see Figure 6 for an example of the effect of NF diameterdistribution on proportion of the volume taken by larger NFs).Although a degree of polydispersity is inherent in samples of

electrospun fibers throughout the fiber diameter range, it isespecially difficult to obtain bundles with ultrafine averagediameters without significant artifacts. Producing NF bundleswith smaller average diameters usually requires reduction ofsolution viscosity. Electrospinning is a complex multiphysicsprocess involving a delicate balance between viscosity, surfacetension of the solution, and the applied electric field. At lowviscosities, the surface tension dominates and can lead tocapillary breakup of the electrospinning jet or to so-calledbeaded NFs.99 Polymer chain alignment (and likely crystal-linity) in beads is different from that in the uniform parts of theNFs.100 The presence of beads can constitute a large volume

fraction of the sample. The techniques described in this sectionprovide averaged information for the examined volume. As aresult, it is not possible to obtain correct information oncrystallinity and degree of macromolecular alignment in NFs inthe presence of nonuniformities and other artifacts related tosample preparation.Because of these issues, techniques which rely on nanofiber

bundles or mats for the examination of orientation facesignificant challenges when dealing with the NFs in the ultrafinediameter range.

METHODS OF ANALYSIS APPLICABLE TO INDIVIDUALNFS

In general, evaluation of individual NFs is preferred forstructural studies. Such an evaluation would eliminate theproblems associated with NF bundles described above. Inaddition, use of such techniques in conjunction with the newlydeveloped mechanical testing protocols12 would allow struc-ture−property relationships to be examined on the sameindividual NF. Direct correlation of structure and propertieson single nanofiber specimens would significantly reduce scatter,provide more precise understanding of the mechanisms ofmechanical behavior, and result in quicker and more robustproperty optimization through structural control.

SelectedArea ElectronDiffraction. Selected area electrondiffraction (SAED) is a crystallographic technique that istypically used inside a transmission electron microscope(TEM). The principles governing electron diffraction are similarto those for X-ray diffraction, with some notable differences.Electrons are charged particles, and thus their interaction withthe atoms in the matter is much stronger. In addition,wavelengths corresponding to electrons inside a TEM aremuch shorter than for X-ray diffraction, resulting in significantlysmaller diffraction angles.Because of the short wavelength of electrons, very small spot

sizes in SAED can be achieved by focusing the electron beam.Strong interaction between electrons and material produces asufficiently strong signal-to-noise ratio. As a result, SAED isnaturally suitable for examination of small individual NFs. Theprocedure for evaluating the degree of preferred orientation inSAED is similar that with XRD. SAED was successfully used inthe past to evaluate crystal orientation in carbon nanofibers.4,101

It was also used on several polymer systems with relatively highcrystallinity.16,18,43,102,103

Figure 7. SAED from an individual PAN NF. (a) Area from which the electron diffraction was taken. (b) SAED of as-spun NF, showing anextremely diffuse signal characteristic of a largely amorphous structure. (c) SAED from a NF annealed at 180 °C, showing signs of crystallinityand preferred crystal orientation. The change in the crystal structure is consistent with the results for annealing the NFs above the glasstransition temperature of PAN, shown in previous work.12

ACS Nano Review


J


Application of SAED for polymers with low and variablecrystallinity is more problematic. An example of SAEDexamination of individual PAN NF is shown in Figure 7. TheSAED pattern from a single as-spun NF (see Figure 7a for aTEM image of the fiber) showed an extremely weak and diffusering that did not allow for orientation examination (see Figure7b). This result is qualitatively similar to the results obtained fora NF bundle by XRD (see Figure 3). This is because, similarly toXRD, SAED can only provide information about crystalorientation in the fibers. For this reason, examination of low-crystalline or amorphous structures becomes difficult orimpossible.Unlike the case of XRD, annealing at 130 °C did not result in

significant improvement in the observed SAED. Only afterannealing the NFs at 180 °C, more pronounced crystalline arcswith preferred crystal orientation became visible (see Figure 7c).However, these arcs were not as pronounced as those in XRD.As a result, examination of orientation was still difficult. Onepossible reason for such a weak diffraction is partial damage tothe NF structure from interaction with the electron beam.The energy in the electron beam is focused on a much smaller

area than in XRD. The wavelength of electrons is approximately2 orders of magnitude shorter than that of X-rays. This meansthat the energy of individual electrons is much higher than theenergy of X-ray photons. Charged electrons produce a strongerinteraction with the material. As a result, polymers and otherorganic materials are significantly more susceptible to radiationdamage in SAED than in XRD. The level of damage to the NFstructure is difficult to evaluate. Therefore, SAED datainterpretation for polymers with low crystallinity is difficult.Changes in polymer crystallinity due to annealing and/orradiation damage in SAED can also change the originalorientation and therefore lead to errors.Polarized Optical Light Microscopy. Measurements of

Optical Anisotropy. Polarized optical microscopy measuresbirefringence, the difference in material refractive indexdepending on light polarization. Noncubic crystal structuresand oriented polymers are often birefringent. As a consequence,results from polarized optical microscopy experiments can oftenbe used as an indicator of the degree of macromolecularorientation.104 If the intrinsic birefringence (for perfectly alignedsystem) of the polymer is known, it is possible to convert the

measured birefringence value into the order parameter ⟨P2⟩ (seea typical schematic of the experimental setup in Figure 8a and adetailed explanation in the SI).Figure 8b,c shows the change in light intensity in drawn and

electrospun DNA fibers of 4 μm and 400 nm in diameter,respectively, as a function of the angle between the polarizer andthe fiber. The insets in the panels show the variation in thecorresponding light intensity. Theoretically, retardation (thedifference in the optical path) can be obtained either from the fitor from measurements at any specific angle (generally θ = 45° ispreferred in this case because it produces maximum intensity). Ifsample thickness can be accurately measured, birefringence canthen be calculated. Different orders of birefringence are possible(due to the periodicity of the intensity function). The differentorders can be distinguished through the use of waveplates or bymeasuring birefringence at several wavelengths.Another approach is polarization−modulation (PM) micros-

copy. This technique applies a PM spectroscopic approach in amicroscopy setup to measure differential absorption due toelectronic transitions of the sample (its dichroism) as a functionof the incoming light polarization. Existing commercialequipment allows PM spectroscopic measurements to be carriedout using ultraviolet, visible, and infrared light. The spatialresolution of PM microscopy can also be improved beyond thediffraction limit, using near-field microscopy (see discussionbelow). In the PM setup, the incoming light is modulated(switched) between two orthogonal linearly polarized states athigh frequency. The changing absorption at different polar-izations can be related to the molecular features of absorbingpolymer and the differential absorption can be used as anindicator of molecular orientation. The technique allows fordetection of small differences in the transmitted intensity.105 Assuch, it is theoretically advantageous for samples smaller thanthe spot size (such as NFs), where the absorption is expected tobe small.

Problems with Data Interpretation for SubwavelengthNFs. The techniques described above have several advantages.When visible light is used (similarly to Raman spectroscopydiscussed later), the spot size is significantly smaller than that inFTIR, allowing for a better signal-to-noise ratio when examiningindividual NFs. In principle, ultraviolet (UV) light can be usedto further reduce the spot size. However, care needs to be taken

Figure 8. Polarized optical microscopy experiment. (a) Schematic of a typical experimental setup. (b) Mechanically drawn from solution DNAfiber, 4 μm in diameter, observed at different angles between the fiber and the polarizer. (c) Electrospun DNA fiber, 400 nm in diameter,observed at different angles between the fiber and the polarizer. The insets in panels (b,c) show the variation in the measured intensity as afunction of the angle between the fiber and the polarizer. In the case of mechanically drawn fiber (b), the intensity was measured in the fiberregion away from the “bulge” visible in optical photographs. The electrospun nanofiber was uniform in diameter. The results indicate thepresence of significant preferred orientation in the fibers.

ACS Nano Review


K



because UV light can potentially damage the polymer structure(UV light is especially prone to breaking CC double bonds).On the other hand, when visible light is used, damage tomaterials can generally be avoided. In addition, simple,inexpensive polarized light microscopic instrumentation iswidely available.Despite these advantages, the techniques face several

challenges. As mentioned above, in the case of birefringencemeasurements, there is a need to account for variation of samplethickness over the projected sample width (i.e., over diameter inthe lateral direction of the fiber). The second problem incomparative studies is the changing crystallinity. The crystallineand the amorphous phases have different intrinsic birefringence,and the relative orientation of the phases can also differ. As aresult, the measured birefringence is going to produce acomposite signal. Disentangling the different components ofbirefringence for comparison of orientation between NFs withdifferent diameters would require preparation of multiplesamples with different degrees of crystallinity and orientation.However, independent variation of crystallinity and orientationis very difficult, especially in samples with low crystallinity/highmacromolecular alignment, such as ultrafine NFs.In the case of PM optical microscopy, the absorption of the

crystal and the amorphous phases can be different, as well,leading to a similar problem with data interpretation if they arenot well-separated spectroscopically.Finally, as both techniques measure the intensity of the

transmitted light, the amount of scattered/reflected light willplay a role. The amount of reflected/scattered light will dependon the degree of crystallinity and size of the crystal and on thecurvature of the NF surface (see discussion below). Smaller NFswill scatter more light, leading to problems in comparativestudies. In addition, the light polarized parallel and perpendic-ular to the nanofiber will be scattered differently (see additionaldiscussion below). This different scattering is potentially thereason for the shift in the maximum intensity in the inset inFigure 8c from the 45° angle. The severity of this problem willagain be diameter (curvature)-dependent.Furthermore, because light has approximately Gaussian

distribution of intensity across the spot, the location of thespot center relative to the fiber axis will also play a significant rolein the amount of reflected light and hence measured transmittedintensity. This problem will become more acute as the fiber sizebecomes comparable to the spot size.As a consequence, even though birefringence is shown to be

detectable in individual NFs (see Figure 8), the problemsdescribed above can generally lead to quantitative or evenqualitative errors. More studies are needed to establishapplicability of the above techniques for orientation analysis ofNFs with subwavelength diameters.

EVALUATION OF POLYMER CHAIN ORIENTATIONUSING POLARIZED RAMAN MICROSCOPYRaman spectroscopy is a technique based on inelastic scatteringof monochromatic light by matter. As the scattering is inelastic,the scattered light has a different wavelength from the incominglight. The difference between the two wavelengths is oftenreferred to as the Raman shift. The observed peaks in the Ramanspectrum correspond, similarly to FTIR, to different vibrationalmodes within the material. However, the two techniques aremore sensitive to different types of vibrations. FTIR is sensitiveto the changes in permanent dipoles associated with thevibration. On the other hand, Raman signal relies on variation of

the polarizability (change in the dipole moment) associated withthe vibration. Bonds with weaker permanent dipoles generallyhave higher polarizability and vice versa. As a result, bands cansignificantly change in intensity or appear/disappear, dependingon their symmetry, when comparing the two techniques.Importantly, Raman spectroscopy is often significantly moresensitive to the polymer backbone vibrations as these vibrationsgenerally tend to have weak dipole moments and therefore highpolarizability.106 Raman spectroscopy has been used extensivelyto obtain information on internal stresses,107 chemical structureand composition of polymers and blends, as well as on physicalchanges in polymer conformation and crystallinity,106 includingfor sub-micron-sized structures.108

Raman microscopy is a powerful technique that can generallybe used to interrogate individual nanofibers.31 Advances in lasertechnology, optics, and detectors in modern confocal Ramanmicroscopes allow for small spot sizes for higher signal-to-noiseratio (with spot sizes smaller than 1 μm currently possible),excellent spatial and spectral resolutions, as well as improvedsignal-to-noise ratios. Both spot size and focal depth areproportional to the laser wavelength. The smallest possiblelateral spot size is desirable for best spatial resolution and signal-to-noise ratio when probing a nanosized sample. On the otherhand, small focal depth can sometimes be beneficial (forexample, for depth profiling or in examining thin films) ordetrimental because of the reduction in the signal intensity dueto small confocal volume. A wide array of quality lasers from thenear-infrared to the ultraviolet is available. This allows for achoice of laser based on the required spatial resolution,sensitivity of the detector and the material to specificwavelengths, and other considerations, such as the need toseparate the Raman signal from fluorescence. Confocal Ramanmicroscopy also allows obtaining the signal from the samesample area that is visually observed. This capability is essentialwhen focusing on fibers with submicron and nanometerdiameters. As a result, investigations of individual NFs withdiameters as small as 100 nm and even below are theoreticallypossible.Polarized Raman microscopy is a method that allows for

examination of orientation in the material. Contrary to SAED,polarized Raman analysis can generally provide information onchain orientation in either amorphous or crystalline phases.Consequently, evaluation of the degree of macromolecularorientation in individual NFs with low crystallinity is possible.

Complete Raman Analysis of Polymer Chain Orienta-tion. Both Raman and FTIR are vibrational spectroscopicmethods. However, FTIR is a one-photon absorption technique.Raman spectroscopy, on the other hand, is a two-photonscattering method. This results in several key differences in theanalysis compared to FTIR. The intensity of Raman scatteredlight is proportional to the square of the magnitude of theinduced polarization vector. The polarization vector, in turn, isproportional to the polarizability. The Raman tensor is thespatial derivative of the polarizability with respect to localmolecular coordinates.Examples of the use of polarized Raman microscopy to

determine polymer chain orientation were provided byBower109,110 and can also be found in Jasse and Koenig.38 Thecomplete experimental and data reduction procedures arecomplex. They require 12 Raman spectra measured in threedifferent laser/sample/detector geometries to fully resolve theorientation functions.106,111

ACS Nano Review


L


Common confocal Raman microscopy systems are limited tobackscattering geometry (i.e., the detector collects the signalfrom the sample scattered back toward the objective rather thanat 90° to the direction of the incident light propagation. Aschematic of a typical optical path in polarized Ramanmicroscopy can be found in Figure S1 in the SI). In addition,samples like thin films and fibers cannot usually be placed withthe axis of their large dimension(s) along the light propagationdirection.19 As a result, only a set of intensities under fourpolarization combinations can normally be measured. With thislimitation, additional approximations are needed. It iscommonly assumed that the Raman tensor for a specificvibration has cylindrical symmetry. This assumption is notalways valid, as certain vibrations were shown in the past to havea noncylindrical Raman tensor.112 However, violation of thisassumption does not usually have a large effect on orientationmeasurements in samples with uniaxial orientation. In addition,the Raman tensor for a particular vibration is usually assumed tobe independent from the degree of polymer chain alignment andpolymer crystallinity. This assumption was also shown not to betrue.113 In our experience, the effect of changes in the Ramantensor due to changes in crystallinity is more significant than dueto noncylindrical symmetry.113 With these approximations, thesecond- and the fourth-order Herman orientation factors, ⟨P2⟩and ⟨P4⟩, can be calculated (see the SI).114

In order to obtain the orientation factors, four Raman spectraneed to be collected from the same location on the fiber. Thefour spectra consist of two sets with the incoming light and theanalyzer polarized in the same direction parallel andperpendicular to the fiber, IVV and IHH, respectively, and twosets of cross-polarizations, IVH and IHV. The ratios between theintensities (depolarization ratios) under cross- and parallelpolarizations can then be used to calculate the orientationfactors. Here and in the rest of the paper, intensity refers tointegrated intensity (the area under the spectral peak).Measurements of these ratios for isotropic samples are alsoneeded to fully solve the equations.114 The latter requirementcan create significant experimental problems, as adequateisotropic samples are not always available or possible. Thisproblem can be avoided by additional assumption of a normal

orientation distribution function.113 Amore detailed descriptionof both the experimental and the mathematical treatment can befound in previous studies.19,115 This method will be referred inthe rest of this paper as “comprehensive evaluation/analysismethod”.Obtaining multiple spectra for ultrafine NFs can suffer from a

variety of problems and instrumental artifacts. These are mostlythe consequence of slight drifts of the laser spot relative to theNF (see discussion below; such drift can occur either due toinstrumental issues or due to NF movement).19,100 Thecomprehensive method relies on ratios of absolute intensitiesunder different polarizations. Even small shifts can producesignificant changes in these intensities. These errors willcompound and can prevent reliable data reduction to obtainthe depolarization ratios. The possibility of errors will alsoincrease with the increase in spectrum acquisition time (forsmall samples, longer acquisition times are needed to producebetter signal-to-noise ratios).In general, a simplified approach, utilizing reduced number of

spectra would be of great value both to reduce experimentalerrors and for fast sample evaluation and comparison. This isespecially true for ultrafine NF specimens.

Orientation Evaluation Using Intensity Ratio for aSingle Band under Two Polarizations: An UnexpectedArtifact. A quantitative indicator of polymer chain orientationcan be obtained by examining only VV and HH polarizationcombinations for a single band. For properly acquired spectra,the IVV/IHH intensity ratio will indicate the degree of orientationof a Raman tensor for a certain vibration along the V axis. Thisindicator does not provide an absolute value of orientationfactors, but with the knowledge of molecular conformation, itcan be used for comparative studies. Frisk et al.111 compared theorientation factors extracted from the comprehensive evaluationto sample birefringence and the IVV/IHH ratio in conventionalpoly(propylene terephthalate) fibers. They found goodcorrelation between P2, birefringence, and 1−IVV/IHH. There-fore, the above simplified approach can theoretically be used toexamine size effects in individual NFs. Qualitative comparison ofthe IVV and IHH spectra was performed on electrospunpoly(ethylene oxide) NF bundles.32 However, only spectra for

Figure 9. Simplified examination of polymer chain orientation using the same band at two polarizations in Raman. Reprinted from ref 116.Copyright 2018 American Chemical Society. In polarized Raman measurements of polymer chain orientation, a large difference in IVV and IHHintensities indicates a large preferred orientation of nitrile groups. (a) Raman spectrumof a drop-cast PAN film in two polarization orientations,showing no preferred orientation. Inset shows a schematic of how polarization relates to the film (the measurements are done for both thepolarizer and the analyzer in the same orientation). (b) Raman spectrum of an individual PAN NF (∼300 nm) in two polarization orientationsin the range showing the nitrile band and the CH region, indicating significant apparent orientation. The inset shows an optical microscopeimage of the NF and the respective polarization orientations. The relationship between the respective intensities is contrary to what is expectedbased on molecular models of PAN. (c) Raman spectrum of a commercial PAN microfiber (∼30 μm) in two polarization orientations(orientation is the same as in (b)). The relationship between the respective intensities is as expected based on molecular models of PAN.

ACS Nano Review


M




a single aligned NF bundle were shown, and the IVV/IHH ratiowas not calculated.To evaluate the applicability of this simplified technique to

PANNFs, Raman spectra for a drop-cast film (with no preferredorientation) and an electrospun PAN NF under two differentpolarizations were collected and compared (see Figure 9a,b).The polarization of the laser and the analyzer was parallel, andthe orientation of this polarization was changed relative to thefilm and fiber (see insets in Figure 9a,b).All spectra exhibited very strong, sharp band around 2242

cm−1, corresponding to the nitrile stretching mode in PAN, andseveral strong overlapping bands in the wavenumber range of2800−3100 cm−1, corresponding to different CH vibrations. Asdiscussed in the context of FTIR, both regions can be used fororientation studies in PAN. However, the nitrile band isgenerally preferred because it is spectrally isolated from otherbands. In the case of the drop-cast film (Figure 9a), as expected,the spectra showed no change for different polarizations. In thecase of the fiber, the spectra were markedly different for differentpolarizations (see Figure 9b). As per the description above, thisdifference should indicate preferred orientation of the polymerchains within the NF.However, unexpectedly (and similarly to the bundle

examination by FTIR), the IVV intensity of the nitrile stretchingmode was stronger than the IHH (IVV/IHH > 1). Several NFs withdiameters below ∼500 nm were tested, producing qualitativelysimilar results. In fact, the IVV/IHH ratio increased with thedecrease in NF diameter. These results were independentlyreplicated in the laboratory of one of the co-authors of this study(C.P.), indicating a systematic nature to the artifact. This ratio ofintensities of the band under different polarizations is reversedrelative to what is expected based on the molecularconformation of PAN. As described above, the nitrile groupsare arranged almost perpendicular to the main chain. Given thislarge angle between the nitrile group and the main chain, IHHshould be stronger than IVV when preferred axial orientation ofthe main chain is present. A similar relationship between I⊥ andI∥ absorptions was observed in the past using FTIR dichroic ratioboth in drawn films39,45,47 and electrospun PAN NF bundles.15

Raman examination of commercial conventional PAN micro-fiber, Dralon X250 (approximately 30 μm in diameter), alsoexhibited this expected relation between the intensities (seeFigure 9c). Both past FTIR examinations of bundles and ourRaman examination of a commercial PAN fiber dealt withdiameters significantly larger than the wavelength of the lightsource. On the other hand, spectra in the Figure 9b are for anindividual NF smaller than ∼300 nm in diameter (less than halfthe wavelength of the laser). The change in the IHH to IVVrelationship in this case indicates a potential size-related artifactrelated to small NF diameters.Sources of Artifacts in Polarized Raman Analysis of

Sub-wavelength-Diameter Nanofibers. Both the compre-hensive analysis and the simplified approach described aboverequire acquisition of multiple spectra under different polar-ization combinations. The use of multiple polarizationcombinations can lead to several potential artifacts. Some ofthese trace back to instrumental issues. Others are inherent inNF geometry (size and curvature) and birefringence. Thedescription in this section is not meant to be an exhaustive list ofpotential artifacts, and additional studies are needed to identifythe potential sources of errors.Instrumental Sources. One of the well-known instrumental

sources of artifacts in polarized Raman studies is the polar-

ization-dependent response of the system. The response of thegrating, which is used in Raman microscopes to achieve spectralseparation of the signal, is usually dependent on the polarizationof the incoming light. The difference in the light intensity afterthe grating can be sometimes as large as a factor of 3 dependingon the incident light polarization. This poses a significantchallenge for polarized Raman studies.In order to avoid this issue, the linear polarization of the light

after the analyzer needs to be converted into a circularpolarization or depolarized (through the use of a quarter-wave-plate or a scrambler). In circular polarization, twoperpendicular polarizations of light with equal amplitude and a90° phase shift are combined. This results in constant lightintensity for all polarizations. As a consequence, conversion oflinear polarization into a circular one avoids the problem ofpolarization-dependent response of the grating.Linear light polarization is commonly converted into a

circular one by inserting a quarter-wave-plate in the optical path(see Figure S2 and section Explanation of Wave-PlateOperation in the SI). The wave-plate direction needs to be at45° relative to the direction of the incoming polarization.However, most Raman systems use a quarter-wave-platesintended to be used with a range of wavelengths. Such wave-plates are imperfect. As a result, instead of a circular polarization,they produce an elliptical one. An elliptical polarization iscreated when the amplitudes of the two perpendicularlypolarized light-waves are unequal or if the phase shift is not90° (this can also happen if the quarter-wave-plate is notperfectly aligned). As a result, light intensity in ellipticalpolarization is not equal for all polarizations. The ratio ofthese intensities will change depending on the direction of theincident linear polarization. Consequently, even with a quarter-wave-plate, residual problems with polarization dependence ofthe grating response remain.This problem is also accentuated by imperfections in the

analyzers that are generally used. A perfect analyzer fullysuppresses polarization perpendicular to the axis of the analyzer.Crystal analyzers can approach perfect ones with large ratios ofsuppression. On the other hand, common film analyzers allow acertain amount of perpendicular polarization through. As aresult, the light polarization after the analyzer is not linear butelliptical. Incoming elliptical polarization on the quarter-wave-plate produces further distortions in the light polarization afterthe plate (in addition to distortions caused by plateimperfections). This, in turn, leads to unequal grating response.The problems with imperfect quarter-wave-plate and analyzer

can largely be overcome by one of the following instrumentaladjustments. First, an achromatic quarter-wave-plate can beused. Such wave-plates produce almost perfectly circularpolarization (within 2−3%) in broad spectral range (forexample, 630−835 nm). Further, a crystal-based analyzer withhigh suppression ratio can be used. However, in most cases suchanalyzers are tailored for a specific wavelength. As a result, adifferent analyzer is needed for every laser in the system.Currently, such wave-plates are relatively expensive.Second, a crystal with well-defined, known orientation can be

examined. In such a case, relative intensities of the bands underdifferent polarizations can be predicted. Consequently, theresponse of the grating under different polarizations can becalibrated and taken into account in the examination ofunknown systems.The third option is to use a grating with the smallest

polarization dependence in its response. For example, a response

ACS Nano Review


N



of the 1800 gr/mm grating is strongly polarization-dependent,whereas that of a 600 gr/mm is not. It should be noted, however,that the dispersion (separation of different wavelengths) of thegrating is proportional to the grating density. As a result, the useof lower density grating reduces spectral resolution of the signal.This reduction can be partially alleviated by the use of longfocusing distance spectrographs. However, it can still be an acuteproblem if the examined bands are close (spectrally) to otherbands or if the examined band is very narrow (spectrally).Additionally, even with the use of 600 gr/mm, some residualpolarization dependence still remains (on the order of 5%change in intensity depending on the incident polarization).An additional significant potential artifact is due to slight

movement of the nanofiber during the experiment or thepolarization switch. This movement can be caused by slightvibrations in the system, not suppressed by the optical table, ordue to physically rotating a half-wave-plate that is inserted in theoptical path to rotate the laser polarization. Alternatively, thismovement can be the result of areas of the nanofiber suspendedin the air over the substrate. This generally happens when thesample is prepared by transferring the nanofiber to the substraterather than spinning directly on it. The nanofiber movement canbe either vertical or lateral (see Figure S3 in the SI forcomparison of this effect in fibers with 1 μm and 100 nmdiameters and system parameters used in our experiments).In cases where large fiber movements are detected, the

measurements should be discarded. However, subtle move-ments may be not detectable with current sample trackingtechniques. Even small shifts of the fiber can cause significantartificial changes in the signal. The focal depth and the spot sizeare proportional to the laser wavelength and inverselyproportional to the square of the numerical aperture of theobjective used (see Figure 10a). For example, focal depth for thelaser and the numerical aperture used in this experiment (633nm laser wavelength, and numerical aperture of 0.9) isapproximately 3 μm and the spot size is approximately 0.86 μm.NF shift changes the confocal volume (volume producing the

Raman signal) for different NF diameters. In addition, laser

intensity has an approximately Gaussian distribution withrespect to the focal plane (as a function of depth) and thecenter of the spot (laterally). A shift in the NF axis relative to thefocal plane either vertically or laterally will change the amount ofmaterial in the focal volume (at or close to the optimal laserintensity). For sufficiently small diameters, the shift can movethe focal point outside of the NF altogether. As a result, thinnerNFs can produce artificially different intensities at differentpolarizations due to NF movement during the switch. Lateralmovement will produce a significantly larger effect due to thefact that the lateral spot size is significantly smaller than thevertical dimension of the confocal volume (see Figure S3 in theSI).The artifacts due to nanofiber shift will be present in any

sample. However, their effect, when the sample dimensions (orradius) are larger than roughly twice the wavelength of the laser,diminish dramatically. This relationship between the laserwavelength and the “critical” sample dimensions is true formicroscope objectives with large numerical aperture (0.9 in thecase of this study). In cases of low numerical aperture (low-magnification) objectives, the laser spot size can increasedramatically (spot size would be ∼3 μm for the same laser asused in this study and numerical aperture of 0.25 compared to∼0.86 μm for the settings used). Although this increase in spotsize will reduce the artifacts caused by fiber movement (due to amore uniform light intensity around the center of the spot andthe focal plane), it will also drastically decrease the signal-to-noise ratio. Under these conditions, the “critical” sampledimensions will be significantly larger than half the laserwavelength.Because the direction of the nanofiber shift is not systematic,

the overall effect of these artifacts will be to increase scatter inthe results as nanofiber diameter decreases. This problem willbecome especially acute for NF diameters comparable to orbelow the vertical step of the typical microscope stage used inRaman systems (∼100 nm).

Artifacts Due to Nanofiber Birefringence and SurfaceCurvature. An additional effect on the signal under different

Figure 10. Sources of artifacts in polarized Raman measurements, using multiple polarizations. (a) Schematic of spot size and focal depth inconfocal Raman spectroscopy. (b) Changes in the incidence angle due to curvature of the NF surface (top) and relation between the plane ofincidence and the laser polarization in NFs resulting in s-polarized and p-polarized light (bottom). (c) Effect of diameter on total reflectance ofthe fiber. The panel compares the distribution of laser intensity in the fiber cross-sectional for diameters of 500 and 100 nm (drawn to scale).For larger fibers, laser intensity at large incidence angles is small, reducing the amount of reflected light. For smaller fibers, laser intensity isalmost constant for all incidence angles, and the total reflectance is large.

ACS Nano Review


O




polarizations is the difference in transmittivity/reflectivity of thematerial. This difference changes the amount of the incidentlight going into thematerial to produce the Raman signal and theamount of the Raman signal coming out.Materials with preferred polymer chain orientation are

expected to be birefringent (see discussion of birefringenceabove). In conventional fibers, differences in refractive index fordifferent light polarizations were shown to be as large as ∼0.1(absolute value) depending on the spinning conditions and thepolymer intrinsic birefringence.117

For a refractive index of 1.5 (typical for polymers), thereflectivity from a flat surface when light propagatesperpendicular to the surface is∼4%. An increase in the refractiveindex to 1.6 would change the reflectivity to ∼5%. Con-sequently, this effect is present in any birefringent media(including films and large fibers), but it is largely negligible.However, NF surface is not flat. As a result, the incidence angle(angle between the normal to the surface and the lightpropagation direction) on most of the illuminated fiber isdifferent from 0° (see Figure 10b). Reflectivity of the surfacegenerally increases for large incidence angles (see discussion ofwhen this is not the case below), approaching 100% at a 90°incidence angle. Consequently, surface curvature of NFs leads tosignificant increases in reflectivity. A small relative change inreflectivity will be much more significant when the totalreflectivity is large. Therefore, the impact of changes in refractiveindex become more pronounced.Light intensity has an approximately Gaussian distribution

across the laser spot. As a result, for larger fibers, laser intensity atlarge incidence angles is small, reducing the amount of reflectedlight. For smaller fibers, laser intensity is almost constant overthe fiber projection, that is, for all incidence angles, and the totalreflectance is large (see Figure 10c for comparison between NFswith diameters of 500 and 100 nm). In addition, thebirefringence of the NFs generally also increases with thedecrease in their diameter (due to improved polymer chainorientation), increasing the difference in reflectivity whenswitching the polarization of the incident light. Consequently,a combination of these two effects can produce a significantdiameter dependence in reflectivity between the differentincident polarizations of the laser.Thus far, only the effect of changes in refractive index due to

changes in light polarization was considered. However, inaddition, reflectivity will change depending on light polarizationin the case of NFs even without the changes in refractive index.For collimated beams or small numerical apertures, lightpolarized parallel to the fiber is perpendicular to the plane ofincidence (plane created by the normal to the surface and thelight propagation direction). It is thus s-polarized. On the otherhand, light polarized perpendicular to the fiber is parallel to theplane of incidence, and thus p-polarized (see Figure 10b). In thecase of large numerical apertures (see Figure S4 in the SI fordifferences between objectives with large and small numericalapertures) the situation is more complicated as a mix of p- and s-polarized light is present. However, even for large numericalapertures (due to laser spot symmetry) light polarized parallel tothe fiber will be predominantly s-polarized, whereas lightpolarized perpendicular to the fiber will be predominantly p-polarized.Reflectivity of any surface at most angles of incidence is

significantly larger for the s-polarized light (the difference issmallest for small angles of incidence). At a certain incidenceangle (Brewster angle), p-polarized light has perfect trans-

missivity (no light is reflected). This angle is∼56° for a materialwith refractive index of 1.5 in vacuum.As this effect is a result of sample surface curvature, it can be

neglected for large samples since the curvature of their surface issmall within the laser spot cross-section. However, similarly tothe effect of birefringence, due to the distribution of laserintensity along the NF surface (Figure 10c), its impact willincrease with decrease in NF diameter. In large fibers, laserintensity impinging the sample near the Brewster angle is smallcompared to small incidence angles. On the other hand, laserintensity in small NFs will be almost constant for all incidenceangles. As a result, the relative change in the total reflectancebetween the two polarizations will increase for smaller NFdiameters.This source of artifacts is, again, largely unique to NFs. Both

for thin films and spherical nanoparticles, an approximatelyequal mix of p-polarized and s-polarized light will always bepresent regardless of the direction of laser polarization. As aresult, no artificial change in intensity due to change in the laserpolarization will occur. On the other hand, in the case of NFs,laser polarization perpendicular and parallel to the fiber axis willproduce an unequal mix of p- and s-polarized light at allincidence angles, including the Brewster angle.

Combined Effects and Approaches to Overcome theArtifacts. All of the above effects result in an intensity fromone of the polarizations being artificially stronger than the other.Some of the effects are random in the sense of the difference (forexample, changes due to the NFmovement). On the other hand,effects due to differences in the refractive index and due to s- andp-polarization of the light are systematic. All of the describedeffects (both random and systematic), with the exception ofpolarization dependent response of the grating, will increase inmagnitude with the decrease in NF diameter. Generally, thedescribed effects are expected to act simultaneously. Thiscomplicates any attempt to extract orientation factors from a fullset of polarization combinations for a broad range of NFdiameters. These artificial changes also have a pronounced effectwhen indicators of orientation (such as IVV/IHH ratio) are used.It should be noted that the artifacts related to material propertiesand effects of NF curvature will similarly impact techniques suchas FTIR and polarized light microscopy that were described inprevious sections when applied to NFs with diameters similar tothe wavelength of the light source. This can significantly limit theapplicability of these techniques to ultrafine NFs. This limitationis going to be even more severe in the case of FTIR because ofthe long wavelengths used.The situation may be further exacerbated when the main

component of the Raman tensor for the chosen band is notperfectly aligned parallel or perpendicular to the backbone chainof the polymer. For example, in the case of PAN the averageangle between the nitrile group and the main chain is∼70°. As aresult, the depolarization ratios for the nitrile band are notexpected to be as large or change as much with changes inorientation (similarly to the dichroic ratio in FTIR). Therefore,the above artifacts can overwhelm changes in intensities due tochanges in the degree of polymer chain alignment. Thus, theobserved reversed relationship in the VV and HH intensitiesfound for PANNFs in Figure 9b is likely to be the result of one ofthe above effects or their combination.In summary, caution is needed when methods requiring

acquisition of multiple spectra under different polarizations areused. As we have shown here, the artifacts described in thissection can significantly skew the results of orientation

ACS Nano Review


P



measurements. Example of PAN shows that, in extreme cases,even the sense of the relationship between intensities underdifferent polarizations can change, leading to qualitativelyincorrect conclusions.The artifacts described above result from the use of multiple

polarization combinations. One way to overcome these artifactsis by using an “internal standard” band to normalize theintensities of the bands used for orientation studies.106

A polarization insensitive band has a constant intensityregardless of the polarization of the incoming light.118,119 Such aband can appear, for example, when the transition vector of thevibration mode is out of the plane created by the twopolarization directions. However, for cylindrically symmetricNFs, any vibration mode will always have an “in-plane”component. Alternatively, a band that produces constantdepolarization ratios under different polarizations regardless ofpolymer chain conformation can be used.106 Unfortunately,theoretically predicting whether the band will be affected bymacromolecular conformation is extremely difficult. Exper-imental verification is similarly complex. As a result, most of theapproaches simply use a band that “appears to be” constant inintensity.106

Care needs to be taken when using the internal standardmethod. For example, McGraw used this approach to studycorrelation between band intensity ratios and density ofpoly(ethylene terephthalate).120 One of the bands was used tonormalize the intensity of other bands. However, Melveger later

showed121 that the intensity of the internal standard band usedin ref 120 was not, in fact, a true “internal standard” but wasdependent on molecular orientation.An alternative method that can overcome the artifacts was

recently proposed and implemented for amorphous fibers.19,100

In this method, the intensity ratio between different bands undersingle polarization was correlated to the orientation factorsobtained from the comprehensive orientation evaluation. Such acorrelation can be found for large samples (such as drawn films,for example) with different degrees of preferred orientations.The degree of macromolecular alignment in these samples canbe probed by either the comprehensive Raman analysis, or byFTIR. The orientation parameters are then correlated to a ratiobetween two different bands under a single polarization, creatinga calibration curve. This calibration curve can be used forexamination of individual NFs under single polarization. At thisstage, no change in polarization is required. This eliminates theartifacts described above.This approach and analysis requires multiple samples, and

acquisition of multiple spectra. However, for a particularmaterial, once the calibration curve is constructed, investigationof NFs was relatively straightforward and was successfullyimplemented on amorphous systems.19,100

However, both the internal standard and the calibrationapproach can be further complicated by changes in crystallinity(in particular, additional corrections might be needed to thecalibration curve). Effects of crystallinity changes on the Raman

Figure 11. Examination of polymer chain orientation using two band intensity ratio at single polarization combination in Raman spectroscopy.(a) Raman bands used to calculate intensity ratio as a measure of polymer chain orientation. (b) Size effects in the Raman band intensity ratiomeasured for individual nanofibers of different diameters, compared to the ratios for the as-drop-cast (green line) and annealed (orange line)PAN films. Comparative examination of I1355/Initrile corrected for changing crystallinity for PAN NFs spun from (c) DMF and (d)dimethylacetamide (DMAc). Adapted from ref 116. Copyright 2018 American Chemical Society.

ACS Nano Review


Q


spectrum are complicated and depend on the examined bandandmaterial. They can include changes in the Raman shift of thebands, appearance of bands associated with a “crystalline” phase,changes in the width of the band, and changes in band intensityratios.122−127 The intensity and the wavenumber of the Ramanbands depend on the local environment (i.e., proximity of othergroups, angles between the different bonds, etc.). Changes incrystallinity result in changes in the local environment. Highercrystallinity leads to a more uniform environment and, as aresult, to narrower bands (band broadening is the result, in part,of frequency shifts due to different environments present). Inaddition, crystalline phases can sometimes be associated withspecificmacromolecular conformations. This leads to changes inband intensity ratios for different conformations, whencrystallinity changes.Impacts of crystallinity and orientation are not easily

separated.123,124,126,128 To fully disentangle these effects,crystallinity and orientation need to be varied independently.In unoriented samples, degree of crystallinity can be controlledby annealing.122 However, for oriented samples, independentcontrol of macromolecular alignment and crystallinity isdifficult. As was discussed above, in conventional manufacturingtechniques, improvement in preferred orientation often leads toincreases in crystallinity. Thus, achieving high macromolecularalignment with low crystallinity is especially problematic. At thesame time, such orientation/crystallinity architecture is themostrelevant to electrospun NFs with unique mechanical perform-ance (simultaneously ultrahigh strength and toughness).12 As aresult, the internal standard and the calibration methods facesignificant challenges dealing with NFs with variable crystal-linity, such as PANNFs exhibiting unique beneficial mechanicalsize effects described in ref 12.Artifact-Free Evaluation of Orientation in Nanofibers

with SubwavelengthDiameter by Two Band Ratio underSingle Polarization. Similarly to the simplified Raman analysisdescribed above, ratios of intensity between different bandsunder a single polarization combination can be used asqualitative indicators of macromolecular orientation whenappropriate bands are chosen. For example, different bandratios were used in the past to examine polymer chainorientation and crystallinity in conventional poly(ethyleneterephthalate) fibers.124,125

A simple, quantitative comparison between different samplescan be performed, using a single polarization (either VV or HH),if two bands are differently polarized (i.e., their polarization is ata different angle to the polymer backbone chain). An optimalchoice of bands (to get the maximum change in intensity ratiowith the change in polymer chain orientation) would utilizebands polarized parallel and perpendicular to the backbonechain. The ratio between such bands will increase under VVpolarization with the improvement in preferred orientation ofthe backbone chain along the NF axis. HH polarization and theinverse ratio can also be used.Such an analysis was recently performed on individual PAN

NFs.116 From Huang and Koenig’s work,129 the band at 1355cm−1 in PAN (exact position of the band depends on multipleparameters such as crystallinity and internal stress) is polarizedparallel to the main chain. On the other hand, the nitrile band ispolarized close to perpendicular to the backbone chain. Thesetwo bands and their relative intensities are shown in Figure 11a.Because the nitrile band is significantly stronger, VV polarizationwas chosen for the experiments to maximize the I1355/Initrile ratio.

Note that the intensity ratio is not an orientation factor like ⟨P2⟩and does not necessarily linearly correlates to it.Variation in measured I1355/Initrile ratio is shown in Figure 11b.

Scatter in the results was relatively high. That is typical forstudies of experimental fibers. However, general trend in theRaman intensity ratio as a function of NF diameter is still readilyapparent, showing significant increases for fine nanofilaments.This size effect in polymer chain orientation correlates well withthe increases in NF modulus (see Figure 1a).The observed significant increase in the I1355/Initrile ratio is

especially impressive given the fact that the nitrile group is notfully perpendicular to the main chain. This fact reduces thechanges in the I1355/Initrile ratio with the improved chainorientation.As mentioned, changes in crystallinity in PAN NFs can have

an effect on the Raman intensity of the different bands. Thiseffect may overlap with and/or distort the effect of polymerchain alignment. Effect of crystallinity on the I1355/Initrile Ramanintensity ratio was examined on films with no preferred polymerchain orientation, and a correction was introduced116 (see the SIfor typical XRD images used for crystallinity calculations). Themodified method was used for first comparative study of size-dependent orientation in NF families spun from differentsolvent systems (see Figure 11c,d).116

In most previous studies of orientation in polymer NFs, only amodest, gradual increase was demonstrated with the decrease inNF diameter.34 This is most likely due to the fact that NFbundles were examined. However, examination of the Ramanintensity ratio in Figure 11c,d shows a sharp dependence on NFdiameter. The ratio increased approximately 8-fold for NFdiameter decrease from ∼500 to ∼140 nm. Size effect inorientation factor that was qualitatively more similar to thisstudy was observed in recent studies of atactic polystyrene.100

Note, however, that the thinnest NFs in that work were ∼500nm in diameter. In this study, the smallest diameter (∼140 nm)was approximately 3.5 times smaller and 4.5 times smaller thanthe laser wavelength used (633 nm). To our knowledge, this isthe thinnest individual NF examined by polarized Ramantechnique to date. This diameter is well within the range ofdiameters of highest interest for mechanical propertiesoptimization (<250 nm).12 Demonstrated applicability of thenew method for quantitative comparative analysis of nanofibersspun from different solvent systems shows high potential of thistechnique for fast inexpensive structural size effect studies andfor quantitative parametric analyses of effects of processing onorientation. Coupled with size-dependent mechanical evalua-tion, this will enable uncovering the fundamental processingstructure−property relationships for various nanofibers.

EMERGING TECHNIQUES FOR EVALUATION OFORIENTATION

Detection of Fine-Scale Oriented Features by AtomicForce Microscopy. Atomic force microscopy (AFM) is a well-established technique for imaging and mapping local top-ography and properties of materials. AFM utilizes sharpmechanical probes that can be operated in contact, intermittentcontact (tapping), and noncontact modes, each having distinctadvantages for imaging. AFM can resolve single molecules withhigh-resolution130,131 and is sensitive enough to characterizebiological macromolecules, such as DNA, in both liquid andnonliquid environments.132,133

AFM can be useful for orientation characterization throughdetection of local topographical or property profiles with high

ACS Nano Review


R



spatial resolution. Early investigations of conventional micro-fibers using AFMwere restricted to characterizing fiber surfaces.These studies revealed highly oriented surface features andordered hierarchical (fractal) structures in high-performancefibers.134,135 A critical advantage of AFM is its ability to probethe local mechanical response of materials. Mechanical propertymapping, in conjunction with traditional topographic imaging,has been demonstrated on advanced polymer fiber surfaces.136

Oriented features were identified along the fiber axis, and themodulus maps revealed regions of decreased stiffness betweenthese features. Similar AFM capabilities were demonstrated onsurfaces of electrospun NFs.137−139 In one of the latter studies,phase contrast imaging revealed enhanced surface orientation ofthe phase-separated structure along the axis of an elongatedpolyurethane nanofiber. However, relating surface informationto orientation of substructures within the NF volume is notstraightforward and needs to be further studied.Recently, a novel sample preparation technique utilizing a

focused ion beam (FIB) was developed where reciprocalnotches are milled into individual fibers to expose the internalsurface through guided shear (see schematic in Figure 12a).140

The resulting surfaces are subsequently characterized usingAFM. In theory, the FIB notches need not to be symmetrical andexamination of fiber structure at different depths is possible. Thistechnique has been used to study structure and mechanicalproperties of several high-performance fibers.141−143 Topo-graphic and transverse stiffness maps acquired simultaneouslyrevealed highly oriented fibrillar structure with nanoscale fibrils

oriented along the fiber axis. Such fibrils are usually associatedwith the crystal phase in the highly crystalline advancedmicrofibers. Using samples prepared by the above FIB-notchtechnique, AFM has been recently utilized to characterizeinterfibrillar interactions in ultrahigh molecular weight poly-ethylene fibers through AFM-enhanced indentation to separatethe neighboring nanofibrils.144

Extending the FIB-notch sample preparation technique toelectrospun NFs should allow, in theory, examination of theirinternal structure, such as dimensions and orientation ofcrystalline and amorphous regions and interfaces, by mappingtopography, stiffness, and other mechanical characteristics.However, achieving shear fracture needed to expose internalstructure may be more difficult in NFs (see comparison of FIB-notch geometry between conventional and electrospun fibers inFigure 12b,c). Notch geometry and spacing will need to beoptimized. Gallium liquid metal ion sources are used in mostFIB systems, but the combination of small NF dimensions andthe low thermal conductivity of polymers in general may requirealternative techniques for milling high aspect ratio features, suchas beams used in helium ion microscopy.145,146

One of the consequences of high macromolecular orientationin NFs is their expected high mechanical anisotropy. Thedifference between the longitudinal and transverse moduli inhighly oriented polymer systems and fibers can exceed 2 ordersof magnitude. This anisotropy can have significant effects on theeffective properties of nanofiber reinforced composites. AFMcan theoretically be used to evaluate such mechanical

Figure 12. AFM orientation measurements using FIB-notch technique. (a) Schematic of the application of FIB-notch technique to expose theinternal structure of the fiber. Adapted with permission from ref 140. Copyright 2016 Elsevier. Comparison of notch geometry between (b)conventional and (c) electrospun fibers, showing significantly better notch aspect ratio achieved in conventional fiber. (d) Schematic ofmechanical anisotropy measurements on longitudinal and transverse cross sections of a nanofiber embedded in the matrix using AFM. The tipsize starts to approach NF and interphase dimensions, requiring explicit 3D models for data reduction and property estimation. (e)Topography, (f) DMTmodulus, and (g) PeakForce QNMdeformationmaps of the same internal area of a high performance poly(p-phenylene-2,6-benzobisoxazole) (PBO) fiber. The property maps provide resolution of features not distinguishable in the topography map.

ACS Nano Review


S


anisotropy, which can serve as a measure of polymer chainorientation. To obtain anisotropic mechanical data, nanofibersamples can be embedded into a matrix and longitudinal andtransverse sections of the fiber can then be examined (seeschematic in Figure 12d). An example of an experiment of thistype was performed on a conventional high-performancefiber.147 Similar studies have been attempted on NFs.148,149

However, quantitative extraction of NF mechanical propertiesusing this technique may be significantly more complicated forfine NFs. Currently, elastic moduli in AFM and nanoindentationexperiments are computed based on one of relatively simpleclassical indentation models. These models consider probes ofvarious geometries interacting with a homogeneous isotropicelastic half space. Such models may be appropriate for theanalysis of microfibers with fiber diameters significantly largerthan the diameter of the probe tip. For ultrafine nanofibers,however, the fiber diameter is comparable to the diameter of theAFM tip. In addition, the interphase region (a volume known toform at interfaces in composites due to physicochemicalinteractions between the fiber and matrix) may now besufficiently large compared to the fiber (see illustration inFigure 12d). Such an interphase would need to be taken intoaccount. As a result, a full multiphase 3D elastic model needs tobe used for accurate data reduction. Nanofiber anisotropy andthe generally unknown stiffness distribution through thethickness of the interphase layer further complicates the solutionand requires appropriate mechanical assumptions. To ourknowledge, no heterogeneous mechanical models have beendeveloped to extract anisotropic fiber properties from nano-mechanics experiments to date. Note that, even for conventionalmicrometer-sized fibers, the above-described heterogeneity andanisotropy need to be taken into account if the regions of thefiber close to the interface are probed.Recently, enhanced AFM techniques such as bimodal

amplitude modulation−frequency modulation (AMFM) modeand peak force quantitative nanomechanical mapping (PF-QNM) have been shown capable of producing high-resolutionmaps of mechanical properties. The former uses two distinctlasers to modulate the cantilever at the first and secondresonance frequency simultaneously.150 Amplitude modulationcorresponds to changes in sample topography, whereasmodulation of the second resonance frequency is related tochanges in material stiffness. In PF-QNM, a high-frequency (2kHz) force curve mapping technique generates high-resolutionmaps for modulus, adhesion, dissipation, and deformation, inaddition to the topography map.151 Figure 12e−g shows thatsimultaneous mapping of different material properties applyingPF-QNM in the section of an advanced PBO fiber prepared byFIB-notch technique resolved ultrafine substructures that wereundetected in the topography maps (compare gradual variationof topographic height in Figure 12e to much finer, higherfrequency features on the mechanical maps in Figure 12f,g).Similar differences in topographic and property maps of fiberswere observed using AMFM technique.141−143 Combined withthe innovative sectioning techniques, demonstrated highmechanical sensitivity and ultrahigh spatial resolution providean intriguing possibility of performing internal featuremetrologyin NFs with ultrafine crystalline and pseudocrystalline (ormesophase) structures that remain elusive to other character-ization techniques. The latter structures are of special interest astheir mechanical properties may be only subtly different fromthe properties of the surrounding amorphous phase. Informa-tion about presence, size, orientation, and mechanical character-

istics of such structures would be invaluable for building moreaccurate nanofiber structural models. Such models, in turn, canproduce critical advances in understanding nanofiber mechan-ical behavior.

Micro- and Nanofocused X-ray Microscopy withAdvanced Light Sources. Initial experiments in quasi X-raymicroscopy trace back to the discovery of X-rays in the late 19thcentury. In the 1950s, the first X-ray microscopy experimentsusing grazing incidence optics were developed. First lightsource-based X-ray microscopes were constructed in 1970s.Finally, large expansion of interest in X-ray microscopy, whichcontinues to this date, started in the 1990s with the introductionof new equipment and new light sources, especially synchrotronbased X-ray sources. Since then, X-ray microscopy has gained awide variety of applications from soil science to biology andpolymer science.152

In principle, the limitations of small interaction volumes of X-ray scattering from nanoscale fibers, which led to the demand forthe examination of NF bundles, can be overcome by the use ofmicro- and nanofocused synchrotron X-ray beams. Thisfocusing can be accomplished either through reflection,refraction, or diffraction of an incident synchrotronbeam153,154 with geometries that can be designed to matchsource parameters to a desired spot size and divergence. Grazingincidence Kirkpatrick−Baez focusing mirrors are typically usedto provide high flux micron and submicron beam spots formicroscopy.155 Fresnel zone plate diffractive optics have beendemonstrated to provide focused beam spots in the tens ofnanometers and are commonly used for nanoscale microscopyin both the soft (100−1000 eV)156 and more recently hard(7000−20000 eV)157,158 X-ray regimes. The fundamentaladvantage of this approach is the creation of real-spacemicroscopy. The samples can be raster scanned to create aspatial map of sample properties, using any of the fundamentalcontrast modes of X-ray microscopy such as absorption,diffraction or fluorescence.Scanning transmission X-ray microscopy methods in the soft

X-ray regime have been demonstrated to map local orientationand heterogeneity of biofibers at sub-100 nm spatialresolution.159 This was accomplished by near-edge absorptionfine structure spectromicroscopy,160 where the dichroicabsorption contrast relative to a linearly polarized X-ray beamis used. A similar technique revealed local radial alignment and adegree of orientation of functional groups in Kevlar fibers.161

The potential of this method for soft material studies should becarefully viewed in the context of dose limitations due toradiation damage. Even though the X-ray probe is fundamentallymore weakly interacting than an electron probe on a per-interaction-event basis, the dose required for either electron orX-ray microscopy to image volumes at a comparable voxelresolution and imaging contrast is comparable or in some caseshigher for X-ray microscopy. This has been explicitly calculatedin the case of biological materials and is represented by thehighest achievable imaging resolution that can be attained beforethe sample is exposed to a maximum allowed X-ray dose (wherethe maximum dose threshold is typically defined as a 50%reduction in scattered intensity). This results in a calculatedlimit of ∼10 nm spatial resolution for X-ray microscopyfollowing the Rose criterion162,163 versus a ∼5 nm resolutionlimit for electron microscopy,164 showing the two methods arefundamentally comparable in this regime for biomaterials. TheX-ray beam damage can be caused by a variety of interactionswith an ionizing beam such as free radical generation, oxidation,

ACS Nano Review


T


chemical activation, differential charging, or local heating,leading to a loss of mass or loss of crystallinity in polymermaterial.165 Some of these pathways have been shown to triggerthe crystallization of polymer fibers166 and even to beunexpectedly present in more dose-tolerant semiconductormaterials.167 Consequently, it is difficult to estimate the damagepotential of beam interaction in complex systems.It has been attempted to overcome these limits by using

coherent diffraction imaging techniques which both remove theneed for a nanofocused beam and make an explicitly phase-sensitive measurement by using the scattered intensitydistribution over a wide range of momentum transfer toiteratively reconstruct a real-space image.168 However, reso-lutions even with coherent imaging techniques have not yet beendemonstrated past the theoretical limit of direct imagingtechniques in the case of biological materials.169,170 This ismost likely due to the fact that the damage limits calculated arebased on fundamental contrast efficiency, which remains thesame regardless of imaging methodology.In the hard X-ray regime, the expandedmomentum transfer of

signal from microfocused beams provide a mechanism for bothX-ray diffraction and small-angle X-ray scattering (SAXS)contrast microscopies that are sensitive to local orientationand spacing of fiber bundles.171 This technique has recentlybeen demonstrated to create a 6D SAXS tomographic map, thatis, internal 3D orientation of a 3D voxel, at 100 μm3 real-spacevoxel size (see Figure 13).172

Nanofocused X-ray Bragg diffraction microscopy methods todate are typically applied to hard materials in highly crystallinestates with known orientations such as epitaxial films andmicrocrystal materials in both 2D173,174 and 3D.175 Theexperimental realization of nanofocused hard X-ray diffractionmicroscopy methods gives the tantalizing possibility of single-fiber strain analysis. However, significant challenges are presentfor attempting orientation studies, using a nanofocused beam ona single fiber as one loses the passive benefit of an orientationalaverage of multiple fibers being present at random positions and

attitudes within a larger X-ray beam volume. To attain a singlefiber diffraction signal, for example, the passive orientationaldiversity present in conventional XRD studies must be replacedby a tilt series, typically a rotational scan of a fiber necessitating2D maps at each angular position to realign the nanobeam ontothe nanofiber, resulting in significant dosing of the sample toattain any diffraction signal. Planned synchrotron sourceupgrades to diffraction-limited storage rings such as theAdvanced Photon Source Upgrade project (https://www.aps.anl.gov/APS-Upgrade) are expected to create 2 orders ofmagnitude improvement in nanofocused hard X-ray flux whichmay make these measurements feasible in the future. Even morebrilliant X-ray sources such as Free Electron Lasers (FELs) givethe intriguing possibility of avoiding beam damage limitationsaltogether in “diffract-before-destroy” single pulse imagingmodes which use a femtosecond X-ray pulse so short in timethat the entire scattering experiment is concluded before thebeam damage can alter the structure of the sample.176,177

However, these techniques have generally yet to be applied topolymer science.

Fast Scanning Calorimetry. Another potential emergingtechnique is FSC, also called flash DSC. In this technique, a verysmall sample with mass ranging from 50 to 200 ng is placeddirectly on the sensor rather than into a crucible, like inconventional DSC. The required sample weight corresponds tothe weight of 50 mm long section of single nanofiber withdiameter in the range from 150 to 300 nm, that makes thistechnique potentially useable for interrogation of individualnanofibers. The measured heat flux is proportional to samplemass and heating rate. As a result, ultrafast heating and coolingrates can be used, while retaining sensitivity. This technique canprovide information on nucleation178,179 and crystallizationkinetics,180 and on crystalline reorganization.181 High scanningrate available with this technique allows determination of theheat capacity in polymers that cannot be easily amorphisized. Itis also well suited for studying structural recovery182,183 andearly stages of polymer degradation.184 It offers the possibility to

Figure 13. Real-space image of collagen fiber within a tooth sample, showing 3D orientation distribution. Reprinted by permission from ref 172.Copyright 2015 Springer Nature.

ACS Nano Review


U

https://www.aps.anl.gov/APS-Upgrade

https://www.aps.anl.gov/APS-Upgrade


magnify changes in recovery kinetics due to structuralconstraints,185 which could become a criteria for investigationof macromolecular alignment. In addition, this technique hasshown great potential for investigating thermal events inmaterial for which degradation under standard DSC conditionscan occur. Examples of such materials include PAN, various silkproteins186 and starch.187

Combination of FSC with other techniques such as X-raymicroscopy offers an intriguing possibility of in situ evaluation ofstructure formation. FSC was used in conjunction withsynchrotron WAXD to study crystallization processes in highdensity polyethylene and polyamide.188 It was also used togetherwith nanofocus X-ray scattering to investigate the phenomenonof multiple melting behavior of poly(trimethylene terephtha-late).189

Some non-equilibrium structures, which are hard to observeotherwise, can be produced in FSC. Microfocus wide-angle X-ray scattering has been used to analyze polymorphism inisotactic polypropylene and polyamide 6 crystallized in a fastscanning chip calorimeter.190 This capability may be ofsignificant interest to studying structures of electrospun NFs,as fast solvent evaporation can theoretically create non-equilibrium structures similar to the ones produced by fastquenching.One of the challenges for accurate FSC measurements is

achieving good thermal contact between the sample and thesensor. An imperfect thermal contact is likely to result inbroadening of the signatures related to each thermal event. Poorthermal contact can be severe in the case of electrospun NFs,when porous mats of overlapping fibers are used. Low intrinsicthermal conductivity of polymers coupled with the fact that thebulk of the nanofiber sample will not be in direct contact with thesensor can lead to slower and more inhomogeneous heating. Anexample of consequences of poor thermal contact can beobserved in the investigation of melting behavior of electrospunpolyamide fibers.191 The authors observed thermal lag thatintensified and became significant for heating rates above 500 Kmin−1. Recently, possibility of fast scanning calorimetry withheating rates up to 2000 K s−1 was demonstrated on electrospunfiber bundles.192 Fibers of poly(ethylene terephthalate) weredirectly spun on transmission electron microscopy copper grids.The grid was cut and placed on the chip sensor. Direct spinningis claimed to have improved the thermal conductivity betweenthe sample and the sensor. However, provided images ofnanofiber samples allow to suggest that only a few nanofibers inthe bottom layer of the bundle were in direct thermal contactwith the copper grid and there were fibers spanning holes in thegrid, suggesting that thermal conductivity needs to be furtherimproved.The problem of poor thermal contact can possibly be

alleviated by better sample preparation. For example, rarifiedaligned NF arrays (with minimum or no NF intersectionsseeexamples of highly aligned spaced nanofibers in ref 6) can bespun on metallic substrate. This will ensure good contact with athermally conductive substrate. Alternatively, longer sections ofindividual NFs can be used. In the latter case, the NF sampleneeds to be coiled on a thermally conductive surface. Coiling canbe achieved by proper nanofiber deposition control.To our knowledge, no investigations of individual NFs using

FSC were carried out to date. Our recent attempt to examineindividual NFs using this technique showed promise, leading usto believe that, with improved sample preparation, suchexamination is possible.

Near-Field Scanning Optical Microscopy (NSOM/SNOM). Near-field scanning optical microscopy (NSOM orSNOM) is a family of experimental techniques that can provideoptical information with subdiffraction-limit resolution. Intraditional, far-field optics, the resolution is limited toapproximately half the wavelength of the light. In order toovercome this limitation, near-field microscopy makes use ofevanescent or nonpropagating fields that are present only nearthe surface of the object. Imaging beyond the diffraction limitwas first proposed in 1928.193 However, technical limitationslimited widespread application of the approach until much later.Because the intensity of the evanescent field drops offexponentially with the increase in distance from the surface,the probe needs to be placed very close (typically within a fewnanometers) to the surface. This distance is generally keptconstant through either shear or normal force feedbackmechanisms.Near-field microscopy can examine several phenomena that

have an equivalent in classical optics, such as reflection,absorption, scattering, and emission. In addition, it can provideinformation on uniquely near-field phenomena such as Forstertransfer of excitation energy and localized plasmon reso-nance.194

Two different approaches are used to confine the light to thesmall area of the sample. The first one (aperture or a-NSOM)uses subwavelength aperture (on the order of 50−100 nm). Inthis approach the probe is generally either a tapered optical fibercoated with metal except for the opening at the end, or atraditional AFM tip with a hole in the center. The secondapproach focuses the light on a solid tip and is called scatteringor s-NSOM.

Aperture Near-Field Scanning Optical Microscopy (a-NSOM). a-NSOM can be used in several configurations. In theillumination mode, the light is provided through the probe andthen collected either in transmission, reflection, or scattering. Inthe collection mode, the light is generally provided from thebottom of the sample by a far-field source, whereas the collectionis done by the probe through the aperture.In general, a-NSOM can be used in a similar manner to

traditional optical microscopy, including with polarizationcontrol. However, it faces several significant challenges. Metalliccoating around the aperture produces significant losses thatincrease exponentially with increase of the ratio of lightwavelength over the size of the aperture.194 Polarization of thelight also deteriorates as the signal propagates through probe.195

As a consequence, the practical resolution limit of a-NSOM is onthe order of 50 nm, when used with visible light,194 long scantimes are needed, and vibrational spectroscopies such as FTIRare not practical for submicron feature sizes.194 Although Ramanspectroscopy using a-NSOM was demonstrated for somematerial systems,196 it is generally impractical due to weaksignal. In addition to the above limitations, a-NSOM techniquecan suffer from a variety of artifacts. The most severe of these arethe so-called “z-motion” artifacts.197,198 These artifacts cause thesignal to correspond to sample topography rather than opticalproperties.197 As a result, these artifacts can severely limit theapplication of a-NSOM.198

Despite the above limitations, a-NSOM version of techniquessuch as fluorescence spectroscopy,199 PM optical microsco-py,200,201 as well as measurements of birefringence202 aretheoretically possible. Application of these techniques formacromolecular orientation quantification is similar in theoryto their far-field counterparts (see above) but with better spatial

ACS Nano Review


V


resolution. In the case of NF examination, these techniques havean advantage of the interaction area being smaller or comparableto NF diameter. This can potentially allow for orientationmapping along and across the NF.However, a-NSOM measurement of birefringence faces

limitations similar to its far-field counterpart (see above), suchas difficulty with disentangling the signal from the crystalline andthe amorphous phases, which have different intrinsic birefrin-gence, and different reflectivity from the curved surface of theNFs for light polarized parallel and perpendicular to the NF. Theproblem of the effects of the crystal versus amorphous phasesmight be even more acute in the case of a-NSOM version of thetechnique. The aperture size (and lateral resolution) iscomparable to crystal sizes observed in NFs. As a result, at thevery least, this will cause additional scatter of the results due tothe waveguide tip movement over the different phases.In addition, mapping of curved surfaces (such as in the case of

NFs) faces a further complication. Intensity of the near-fieldsignal deteriorates with the increase in distance between theprobe and the subject. If the curvature is significant, there will bea distribution of distances between the probe and the sampleacross the spot size even if constant central distance ismaintained (see example of a 50 nm sized aperture probe overNFs with diameters of 500, 200, and 50 nm in Figure 14a). Thisdistribution will change depending on the location in the map.As a result, the observed intensity will be artificially changedregardless of the orientation features. The problem will intensifyas the curvature becomes larger (smaller NF diameter) relativeto the aperture size. In comparative studies, even if the apertureis perfectly centered over the NFs, the distribution of distanceswill change for different NF diameters (see the example of NFwith 50 nm diameter in Figure 14a). In general, the interactionbetween evanescent field and the curved surface needs to befurther studied.Quantification of macromolecular orientation in electrospun

NFs was attempted using PM near-field optical microscopy(PM-NSOM).203 The authors examined ribbon-shaped NFswith approximately 1.6 μm× 140 nm cross section and observed

significant preferred orientation at the center (“core”) of theNFs as expressed by the dichroic ratio. However, surprisingly,the dichroic ratio changed sign near the edges of the NF (thevolume called “sheath” by the authors). This result wasinterpreted as indication of radial arrangement of the polymerchains within the NF sheath.203

Radial arrangement of polymer chains in the “sheath” isunlikely from the physical standpoint. In fact, as the solventevaporation rate is highest at the surface of the electrospun jet,the degree of macromolecular orientation is expected to behighest in that area. Consequently, the observed switch in thedichroic ratio at the edges of the NF was likely the result ofchanging reflectivity between the different light polarization inthe areas with pronounced curvature, as described above. Thisresult indicates that the artifacts described in this work,stemming from different reflectivity of the curved surface dueto polarization switch, could limit the application of a-NSOMversion of the optical approaches described above, as well.Similar limitations would apply to any other technique thatrequires a switch in the polarization of the incoming light.

Scattering Near-Field Scanning Optical Microscopy (s-NSOM). In s-NSOM, light is focused on a sharp metalized tip(generally an AFM tip) that is used as the probe. The apex of theprobe tip enhances the light field in the immediate vicinity,acting as an antenna. The high evanescent field is confined to avolume with lateral dimensions of approximately one radius ofcurvature of the tip apex. In the case of s-NSOM, theconfinement effect depends only on the apex radius, and isindependent of the wavelength of the incoming light, whichdistinguishes it from both the far-field optics and from a-NSOM.As a result, this technique can produce nanometer resolutionwith a broad range of light wavelengths, including the mid-infrared region.194 Several configurations of s-NSOM exist.Illumination can be brought in from the bottom in the case oftransparent samples or from the side. Signal collection can alsobe placed either below or above the sample.The current model of s-NSOM is the “lightning rod” model,

which builds on the early “dipole” models.206 It is generally

Figure 14. NSOMapplication. (a) Effect of NF diameter and curvature on the distribution of distances between the a-NSOMprobe aperture andthe surface. The distributed distance is expected to affect the evanescent light region and, therefore, the collected light characteristics. (b)Traditional application of s-NSOMwhere the tip is illuminated by light at an angle to the surface, producing vertical and horizontal polarization.Adapted with permission from ref 204. Copyright 2014 AAAS. (c) Use of inclined probe to vary the proportion of in-plane and out-of-planepolarization. Adapted from ref 205. Copyright 2018 American Chemical Society.

ACS Nano Review


W


believed that the light in the evanescent region is polarized alongthe probe axis (i.e., perpendicular to the sample surface) with noin-plane polarization components. This makes approaches suchas PM optical microscopy or traditional birefringence measure-ments currently impossible with s-NSOM. However, recentexperiments showed that an in-plane polarization may exist, aswell.204 This may be due to the fact that the p-polarizedincoming light delivered through the beam with nonzeroincident angle (in the most popular s-NSOM configuration, asshown in Figure 14b) always has an in-plane polarizationcomponent (in addition to the vertical component exciting the“dipole”). In addition, it was recently shown that the lightpolarization direction can be altered through the use of aninclined probe (see Figure 14c).205 Implications of thispossibility for macromolecular orientation quantification andthe application of the above techniques can be profound andneed to be further explored. It is, however, likely that the s-NSOM versions of PM microscopy and birefringence measure-ments would face the same challenges due to high samplecurvature as the approaches based on a-NSOM.In context of macromolecular orientation in fine NFs, s-

NSOM versions of infrared and Raman spectroscopies would beof particular interest. Application of s-NSOM for infraredmicroscopy will be discussed below.Tip-enhanced Raman spectroscopy (TERS) is a version of

surface-enhanced Raman spectroscopy (SERS) where theRaman signal is enhanced by the presence of metallicnanoparticles. It can be considered as a s-NSOM version ofRaman microscopy. TERS has been extensively studied,207−210

and large enhancements of the signal were reported in the past.The enhancement calculations take into account the differencein the signal producing volume. Significant experimentation withtip coating and probe-sample distance needs to be undertaken inorder to obtain best signal enhancement. However, in practice,

the most important parameter is not the signal enhancement,but contrast (the ratio between the near-field to far-fieldsignals).209 This contrast determines whether the near-fieldsignal can be separated from the background. Far-field signal canbe partially suppressed by crossed polarization schemes in orderto improve the contrast.209,211 However, conditions formaximum contrast and maximum signal intensity are not thesame, and often contradictory.209 In addition, care needs to betaken as some spurious far-field signals, such as reflections fromthe pyramidal tip surfaces rather than signals from the apexregion, can still remain even after suppression.209 Importantly,for the purposes of this paper, the near-field Raman signalappears to be less polarization sensitive and might not be suitedfor orientation measurements.209 Better understanding ofpolarization of the evanescent light in s-NSOM/TERS isneeded. A version of the simplified Raman approach, using twobands ratio, proposed above, may be applicable. It should benoted, though, that the correction for crystallinity discussedearlier in this paper will be more difficult in the case of TERSthan in the far-field Raman microscopy. Measuring localcrystallinity under the tip is difficult if not impossible. Inaddition, the crystal size in NFs is similar to the tip size. Both ofthese facts would lead to increased scatter in the results.

Combined Infrared Spectroscopy and Atomic ForceMicroscopy. Several promising techniques combining infraredspectroscopy and AFM have been introduced in recent years.These techniques preserve the rich structural informationprovided by IR spectroscopy but provide much improvedspatial resolution approaching the AFM tip radius, well belowthe far-field diffraction limit at IR wavelengths. The two maintechniques are photothermal induced resonance (PTIR)spectroscopy, also called NanoIR or AFM-IR, and infraredscattering-type near-field optical microcopy (IR s-SNOM), alsocalled nano-FTIR when using a broadband IR source. Although

Figure 15. Applications of PTIR technique in the literature. (a) Top: original AFM-IR configuration with bottom side illumination and thesample mounted on an infrared-transparent prism. Bottom: topside illumination, enabling sample measurements on arbitrary substrates. (b)AFM images and IR spectra of individual electrospunNFs, showing different phases as a result of different collectionmethods. Adapted from ref50. Copyright 2015 American Chemical Society. (c) IR absorbance images of two PVDF fibers collected with the pulsed laser source tuned to1404 cm−1 by using light polarized in the two orthogonal directions indicated. Lighter color indicates stronger absorbance at that wavenumber.Adapted with permission from ref 212. Copyright 2012 SAGE Publications, Ltd.

ACS Nano Review


X


both techniques combine AFM and infrared spectroscopy, theirprinciple of operation is different.PTIR212,213 uses a tunable infrared laser to sweep through the

mid-IR region. Absorption at a certain wavelength results inheating and thermal expansion of the sample. This localizedexpansion is recorded by AFM by probing the deflection of theAFM tip (see Figure 15a). Spectra obtained by this techniquewere shown to correlate well to those obtained by FTIR as thephotothermal effect is proportional to the absorption index.Recently, this technique was used to distinguish betweendifferent crystal phases in individual NFs (see Figure 15b)50 andeven to map the distribution of crystal structures in thenanofiber, revealing the formation of a 10 nm thin shellcontaining an unusual crystal polymorph.214 By rotating theintrinsically polarized laser light, the method has been used toprobe anisotropy in organic thin films and plasmonic systems215

and is also expected to be suitable for examining polymer chainorientation in individual NFs. Qualitative examination of singlePVDF fibers with diameters of ∼1 μm was demonstrated (seeFigure 15c).212 However, suitability of this technique forquantitative analysis of NFs in the ultrafine diameter region isyet to be demonstrated. A limitation of earlier generations ofPTIR instruments was the need to use the AFM tip in contactmode which is not well suited to the cylindrical shape of NFs.Recent instrumental developments, in particular, the resonanceenhancement of the signal when using quantum cascade lasersand the implementation of tapping mode AFM sampling, makePTIR a promising technique to study the molecular structure inindividual nanofibers.In nano-FTIR, a technique derived from scanning near-field

optical microscopy discussed in the previous section, a sharpmetal-coated AFM tip is illuminated by a broadband IR sourceand the tip-scattered light (typically backscattered) is collectedas a function of the tip position and analyzed interferometri-cally.216−218 In contrast to PTIR, analysis of the opticalamplitude and phase of the scattered light enables extractingthe real (n) and imaginary (k) parts of the complex refractiveindex of the sample, the latter being generally comparable tostandard FTIR absorption spectra.218

The technique has been used for chemical mapping withnanoscale resolution218 and for structural analysis and mappingof individual protein complexes.219 Figure 16 illustrates the useof nano-FTIR to map different phases in individual electrospun

NFs (images and spectra were acquired using a neaSNOMmicroscope from neaspec GmbH, Germany). However, directapplicability of this technique for quantification of macro-molecular orientation is yet to be demonstrated. As in other s-NSOM-based techniques, light in nano-FTIR is primarilypolarized perpendicular to the surface due to antenna effect ofthe metal-coated AFM tip. This enables probing out-of-planeorientations, but as was discussed in the context of traditionalFTIR, polarization parallel to fiber axis is also needed in order toobtain orientation factors of NFs.As was mentioned in the previous section, recent experiments

showed possible presence of in-plane polarization componentsin the evanescent s-NSOM volume. If these components can becontrolled to produce polarization in the direction of the fiber, itmight be possible to obtain infrared dichroism values, and,consequently, orientation parameters. Such a technique couldstill be affected by some of the artifacts described above.However, as the lateral dimensions of the examined area in thecase of s-NSOM are comparable to the size of the tip apex, theycan be made very small (5−10 nm). Consequently, the effects oflocal curvature are likely to be less severe than in the case ofRaman microscopy.Even if full examination of infrared dichroism by nano-FTIR is

not possible, qualitative indicators of orientation can beobtained by examining the ratios of bands that are differentlypolarized (similar to the polarized Raman approach describedabove). These orientation indicators can then be used forcomparative studies. Use of nano-FTIR for orientationevaluation was demonstrated in organic semiconductors221

and collagen nanofibrils.222 However, to our knowledge,polymer chain orientation in electrospun NFs was not yetevaluated using this approach. It is also important to note thatsimilarly to the Raman approach, effects of changing crystallinitywill need to be addressed.

ADDRESSING SAMPLE PREPARATION LIMITATIONS

NF Bundles. In addition to new or modified interrogationtechniques, one can attempt to improve sample preparationmethods. Currently, oriented NF bundles are prepared by directelectrospinning using one of several available NF alignmentmethods.6 The resulting partially aligned NF assemblies orsheets are sometimes further postprocessed using techniquessuch as bunching, rolling, twisting, or stretching (note that the

Figure 16. Nanoscale spectroscopic mapping in individual electrospun silk NFs using nano-FTIR. (a) Nano-FTIR technique based on s-SNOMusing broadband IR radiation focused in the near-field using an AFM tip. Modified with permission from ref 220 under a Creative CommonsLicense Attribution 4.0 International CC BY 4.0 https://creativecommons.org/licenses/by/4.0/). (b) AFM imaging of individual electrospunNFs. (c) Nano-FTIR absorption spectra measured at three different locations on the NF marked in (b). The inset shows AFM topography lineprofile corresponding to the dashed line in the left image in (b).

ACS Nano Review


Y

https://creativecommons.org/licenses/by/4.0/


latter may affect native chain orientation in individual nano-fibers). The above protocols result in samples with distributedNF diameters and orientations within the bundle. Thesedistributions often become broader as NF diameter decreasesand/or the bundle diameter (i.e., sample size) increases.It might be possible to apply some of the techniques described

above (e.g., XRD, FTIR, and DSC) to smaller NF bundles. If thebundles are small enough, it should be possible to constructthem by picking and assembling selected individual NFs. Suchan approach could result in narrower NF diameter andorientation distributions. Automating this proposed protocolof nanofiber-by-nanofiber sample assembly can lead to largeruniform samples and more accurate orientation informationobtained using the techniques discussed in the previous sections.Another possibility is modeling-assisted process control.

Electrospinning is a complex multiparameter process. Empiricaloptimization of such processes has limits. Models, describing jetinstabilities and deposition on static or moving substrates223 canguide experimental NF assembly and orientation control. Nearlyperfect alignment is possible in special cases with proper processoptimization.6 Multiphysics continuum jet models that includesolvent evaporation would be invaluable for precision control ofNF diameter.Individual NFs. As discussed above, NF properties and

structure are strongly diameter-dependent in the ultrafinediameter range. Methods of chain orientation evaluationapplicable to individual NFs are best suited for studies ofstructural size effects. Precise diameter measurement andensuring diameter uniformity within the interrogated sectionof the NF are critical for such studies. Pronounced beading dueto onset of capillary breakup instability in dilute solution jets cansignificantly affect the outcomes as beads can occupy substantialfraction of the sample volume. However, this instability usuallyproduces beads in the micrometer diameter range that arerelatively easy to detect and avoid by discarding the affectedsamples. Subtler variation of NF diameter that is more difficultto detect can still adversely affect the results by increasing theexperimental scatter (in the case the diameter measurementerror is randomized) or even leading to systematic errors.Accurate measurement of NF diameter is not trivial. Optical

evaluation is generally diffraction limited and cannot achieveaccurate measurement for NFs in the diameter region of highestinterest. Currently, the most reliable method to measure thediameter of ultrafine NFs is scanning electron microscopy(SEM). However, diameter measurement on the section to bestructurally interrogated can affect the sample, even if environ-mental SEM modes are used with no fiber coating. NFcontinuity and generally weak diameter variation along thenanofiber axis (in absence of beading) allows to overcome thisproblem. Currently the best practice employed in ourlaboratories is to collect and analyze two adjacent sections ofthe same fiber. One section is evaluated in an SEM to accuratelymeasure NF diameter, whereas the other is used for structuralstudies. This approach was used in our previous mechanicalstudies.12,224 NF diameter is measured in several locations alongthe fiber, and if significant variation is observed, the sample isdiscarded. For samples with no systematic diameter variation,structural studies are carried out on the second virgin NFsection.The above approach can be further expanded and used with

several structural characterization techniques or simultaneousstructural and mechanical characterization on the samenanofiber. Such multimodal characterization would require

collection of three or more adjacent sections of NF. Evaluationof the structure of the same NF by several methods cansignificantly improve the accuracy of the method comparisonstudies. Such studies are important for evaluation of newtechniques through their comparison to established methods.Use of the same NF for structural and mechanical testing (thelatter known to produce especially high experimental scatter)can substantially reduce scatter and enable direct correlation ofstructural features to observedmechanical behavior of individualNFs. Such correlation would enable better identification ofmechanisms of NF deformation and failure and could accelerateanalysis of fundamentals of the unique mechanical size effects.

SPECIAL CASES AND CURRENT LIMITATIONSChanging Phase Composition. So far in this review, we

have considered the case of ultrafine NFs with low and variablecrystallinity. We have shown that the degree of macromolecularalignment in such NFs can be quantified, for example, by amodified polarized Raman spectroscopic technique, which canalso account for changing crystallinity of the NFs with differentdiameters. Other techniques such as nano-FTIR can potentiallybe used in a similar way, although accounting for crystallinestructure can be more difficult in these cases. However, thediscussion so far focused on a relatively simple case when thestructure of the polymer is described by a simple two-phase(amorphous−crystalline) material model. Already in ourdiscussion of DSC/DRS experiments, we touched on the factthat such a model is most likely insufficient, as the amorphousphase of the polymer can include both mobile and rigidfractions. In addition, presence of a mesophase is a possibility inelectrospun NFs, as well.19

Our recent examination of electrospun Nylon-6 NFs showedthat a further complication is possible. XRD, DSC, and Ramanexaminations indicate that not only the crystallinity of NFschanges with the decrease in their diameter, but the crystal phasemix changes, as well. Our preliminary investigation showed thatthinner nanofilaments contained increasing amounts ofmetastable γ phase and decreasing amounts of α phase of thepolymer (to be submitted). Under these conditions, disen-tangling the changes in the band ratio (for the simplifiedpolarized Raman analysis) stemming from changes incrystallinity, crystal mix, and orientation is extremely challeng-ing. One way to solve this problem is to find bands that are onlyaffected by two out of the three factors (orientation, crystallinity,and phase composition). This may or may not be possible for aparticular polymer. In the latter case, a complicated series ofexperiments where the degree of crystallinity and the crystal mixare changed independently in unoriented samples is needed todeconvolute the different effects and to introduce the neededcorrections when NFs are examined. Such experiments are notstraightforward. Additional studies are needed in these specialcases to obtain reliable indicators of polymer chain orientation.

Chain Orientation versus Chain Extension. The degreeof polymer chain alignment has been identified above as thecritical parameter determining the mechanical properties ofultrafine NFs with low crystallinity. Another potentiallyimportant parameter is polymer chain extension. It is believedthat fully extended polymer chains are responsible for highmechanical properties of conventional advanced polymerfibers.225,226 Chain extension is normally achieved by preventingchain entanglement. Traditionally, in flexible chain polymerssuch as ultrahigh molecular weight polyethylene (UHMWPE),this is achieved through techniques such as gel spinning.

ACS Nano Review


Z


Recently, an alternative route that involves specialized polymer-ization procedure to produce a solid preform in an unentangledstate was proposed and implement to produce high modulusUHMWPE tapes.227

In electrospinning, NFs are generally produced fromsemidilute polymer solutions. As a result, polymer chainentanglement is expected to be low, and it is likely that theresulting NFs possess fully or almost fully extended chains. Thishypothesis is supported by major increases in modulus andstrength of thinner nanofilaments observed in multiple polymersystems. These high property values are generally well in excessof properties of conventional fibers made from the same polymerand are achieved before significant process optimization. Still,experimental confirmation of fully extended polymer chain stateand what are the critical parameters needed to obtain it can beessential to optimize NF properties.One of the ways to examine the degree of chain extension is by

examining the radius of gyration of the polymer chain paralleland perpendicular to the fiber axis using small angle neutronscattering (SANS).228 This method, which uses the elasticscattering of neutrons to probe length scales from approximately1 to 100 nm, requires the addition of a deuterated reporterpolymer to the electrospun solution. Ideally, the deuteratedpolymer should have the same molecular weight as the mainpolymer being investigated to avoid modification of theviscoelastic properties of the solution and any effect of molecularweight on chain extension and relaxation. This is only possiblefor some polymers such as polystyrene. As a consequence, SANSis essentially only applicable tomodel systems. It also needs largebundles of fibers to be examined and produces averagedinformation. Finally, SANS experiments can only be conductedin a limited number of dedicated facilities worldwide.Another approach is to use fluorescence spectroscopy.

Fluorescence microscopy has been used in the past to examinethe length of DNA229 and actin230 molecules. It was also used toascertain extension of DNA molecules embedded in anelectrospun fiber matrix.231 Two approaches can be undertaken.A marker fluorescent molecule with known length can be addedto the electrospinning solution in very small quantities to ensurespatial separation. The length and extension of the molecule canthen be examined postspinning using fluorescence microscopy.Alternatively, a small quantity (again to ensure spatialseparation) of the main electrospinning polymer molecule canbe fluorescently labeled. The average distance between thedifferent fluorescent sites needs to be measured prior toelectrospinning. This can be accomplished, for example, byexamining the labeled molecule stretched in a microfluidicchannel. After spinning, the degree of chain extension can bethen estimated by measuring the distance between thefluorescent sites in the NF.Similar approaches with better, subdiffraction limit resolution,

can be undertaken using emerging super resolution microscopytechniques such as structured illumination microscopy,simulated emission depletion, stochastic optical reconstructionmicroscopy, or photoactivated localization microscopy. Thetechniques have already been used to examine spatial arrange-ment of protein nanofibrils,232−234 measure the length ofstretched DNA segments,235 and resolve the distances betweenadjacent fluorescent sites with spatial resolution unavailable inregular fluorescent microscopy.236 It is likely that suchapproaches can be adopted to examine chain extension inelectrospun NFs.

In the above techniques, only a small amount of moleculesneeds to be fluorescent in order to provide sufficient spatialseparation between the fluorescent chains. This would allowsimultaneous evaluation of polymer chain alignment andextension. However, it needs to be noted that in bothapproaches described above (addition of fluorescent markermolecule or labeling a small number of native polymermolecules), changes are made to the fiber-forming solution. Inelectrospinning nanofabrication process, even small changes insolution parameters can produce significant changes in theresulting NF morphology and structure. Consequently, effectson processing need to be studied and taken into account whileinterpreting the results of the proposed experiments.

CONCLUSIONS: CURRENT STATUS AND PROSPECTSIn-depth analysis of structure−property relationships inadvanced fibers in the second half of the 20th Century has ledto development of fibers with extraordinary strength andmodulus. These fibers are now ubiquitous in a myriad ofapplications from sporting goods to aerospace structures.Ultrafine nanofibers have recently demonstrated unique ultra-high toughness in addition to high strength and modulus. Hightoughness is believed to be due to low crystallinity as a result ofultrafast solvent evaporation during the electrospinning process.Crystallization in the thin electrospun jets is retarded by fastsolidification, despite high macromolecular alignment inultrafine-diameter nanofilaments. The degree of polymer chainalignment is therefore the main structural parameter determin-ing nanofibers mechanical and physical properties. Polymerchain orientation in ultrafine NFs is responsible for their highstrength and modulus. Chain sliding at large deformations,enabled by low NF crystallinity that is responsible for ultrahightoughness, is also affected by chain orientation. Quantification ofthe degree of macromolecular alignment is, therefore, criticallyneeded for better understanding and control of unusualmechanical behavior and size effects in NFs.If there is any one central message that the authors would like

to convey in this review, it is that evaluation of macromolecularchain orientation in fine polymer NFs is a formidable challenge.Many of the traditional techniques can provide valuableinformation but all face significant difficulties when applied toultrafine, sub-250 nm diameter NFs (see Table S2 summarizingadvantages and limitations of different techniques in the SI).Techniques such as XRD and SAED that examine the

crystalline phase of the polymer are limited in their utility due tolow NF crystallinity. Results from techniques that need largeamounts of material such as XRD, FTIR, DSC/DRS, and NMRare further confounded by the inherent and often severemisalignment and polydispersity of NF samples produced by theinstability-driven nanofabrication process. Sample inhomoge-neity usually increases with reduction of average NF diameter.Additionally, low viscosities, required to achieve ultrafinediameters, can lead to beading detrimental to structuralevaluation in NF bundles. Consequently, the techniques relyingon NF interrogation of bundles are currently providing onlyqualitative information when applied to ultrafine NFs. Suchinformation is useful for qualitative evaluation of trends andvisualization. It can provide complementary information forcomparative studies of different nanofiber systems. Averagedinformation on bundles is also valuable for applications utilizingexact same bundles or other nanofiber assemblies with similarNF polydispersity. However, in-depth fundamental studies andnanofiber properties optimization would generally require

ACS Nano Review


AA



higher accuracy. Better sample preparation techniques, such asthe ones described in above, can significantly improve theoutcomes for the bundle-based methods and elevate them to thequantitative level.Currently, techniques interrogating individual NFs are

preferred for structural size effects studies in the ultrafine NFdiameter range. Note, however, that all relevant spectroscopictechniques that rely on measuring light intensities underdifferent polarizations face a variety of instrumental and inherentlight-sample interaction-related artifacts due to high curvature ofNFs with subwavelength diameters. Such artifacts, discussed indetail on a polarized Ramanmicroscopy example, were shown tolead to quantitatively and even qualitatively erroneous results.Some of these artifacts can also appear in examination of NFbundles when a switch in polarization is needed (e.g., in usingpolarized FTIR). Further carefully designed studies are neededto better understand and avoid these artifacts.A modified polarized Raman technique described above

emerges as a relatively simple option suitable for quickcomparative orientation studies. It is inexpensive, relativelyfast, uses widely available instrumentation, and is able toexamine individual NFs. The technique was implemented forboth amorphous and low-crystalline polymer systems. It wasdemonstrated to produce quantitative information that can beused to distinguish structural size effects in NFs produced usingdifferent nanofabrication parameters. The latter opens uppossibilities for more detailed processing structure−propertyanalyses that can lead to accelerated research and developmentof supertough strong continuous nanofibers. We expect that theartifact-free approach recently developed and demonstrated bythe coauthors of this study for Raman spectroscopy (i.e., analysisof multiple bands with different orientation under single lightpolarization) should be applicable to other spectroscopictechniques discussed in the review. It can be used to avoiderrors due to light interaction with subwavelength-diameter(ultrafine) NF samples with high curvature.Several emerging techniques discussed above show high

potential for orientational studies on individual nanofibers orsmall bundles. However, their ability to reliably quantifypolymer chain orientation in ultrafine electrospun NFs is yetto be demonstrated. Several anticipated problems, described inthis review, need to be further analyzed and addressed.Fast solvent evaporation in electrospinning can create

metastable phases and unusual crystal phase mixes in somepolymer systems. The content as well as orientation of thesephases can vary with nanofiber diameters. Further studies areneeded for quantification of macromolecular alignment andstructural size effects in these more complex systems.Currently, no single technique can provide comprehensive

evaluation of orientation in ultrafine low-crystalline NFs.Techniques described in this review should be used with theappropriate caution, and several techniques should be employedto both provide complementary information and to cross-reference the results for possible artifacts.Reliable quantification of polymer chain alignment in ultrafine

electrospun NFs will lead to better understanding of structuralmechanisms behind their unique property combinations andwill significantly accelerate property optimization. Developmentof simultaneously ultratough and strong continuous nanofiberscan revolutionize structural nanomaterials research and lead tonext generation bulk nanostructured materials and nano-composites for safety-critical applications.

ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsnano.8b08725.

Orientation studies in nanofibers; orientation evaluationusing FTIR; experimental details for DSC and DRS tests;SSNMR examination of PAN powder; orientationevaluation using polarized light microscopy; schematicof the optical path in polarized Raman microscopy;comprehensive orientation evaluation using polarizedRaman spectroscopy; explanation of wave-plate oper-ation; schematic of the effect of fiber movement on laserintensity that produces the Raman signal for different NFdiameters; schematic of the difference between objectiveswith small and large numerical aperture; typical XRDpatterns for PAN film and electrospun mat; summary ofthe techniques evaluated in this review (PDF)

AUTHOR INFORMATIONCorresponding Author*E-mail: [email protected] Delpouve: 0000-0002-6064-7151Laurent Delbreilh: 0000-0002-9322-7153Vinayak P. Dravid: 0000-0002-6007-3063Christian Pellerin: 0000-0001-6144-1318Yuris A. Dzenis: 0000-0002-6645-3294Present Addresses▲Department of Surgery, University of Nebraska MedicalCenter, Omaha, NE 68198-7690.▽Department of Mechanics, Huazhong University of Scienceand Technology, Wuhan, Hubei 430074, China.NotesThe authors declare no competing financial interest.

ACKNOWLEDGMENTSThis work was supported in part by the grants from ONR(N000141410663), NSF (DMR-1310534, CMMI-1463636),and NIH (1R01HL125736-01). C.P. acknowledges the supportfrom the Natural Sciences and Engineering Council of Canada.M.H. acknowledges the support of the Center for NanoscaleMaterials, a U.S. Department of Energy Office of Science UserFacility under Contract No. DE-AC02-06CH11357. Theauthors thank Bruce Chase of University of Delaware, DmitriBasov of Columbia University, Alan D. English of Macro-molecules journal, and Martha Morton of University ofNebraskaLincoln for reading sections of this work andvaluable comments; Tobias Gokus and Nicolai Hartmann fromneaspec GmbH for their help with nano-FTIR measurements;and Joel Brehm of UNL for his help with creating graphicalschematics for this paper.

VOCABULARYadvanced fibers, high-performance synthetic fibers with higherstrength and modulus than typical textile fibers; most currentadvanced fibers were developed in the second half of the 20thCentury; diameters of modern advanced fibers range fromseveral to tens of micrometers; continuous nanofibers, fiberswith submicrometer diameters produced by electrospinning orother techniques; strength and stiffness, properties describingthe ability of a material to withstand an applied load without

ACS Nano Review


AB

http://pubs.acs.org

http://pubs.acs.org/doi/abs/10.1021/acsnano.8b08725


mailto:[email protected]

http://orcid.org/0000-0002-6064-7151

http://orcid.org/0000-0002-9322-7153

http://orcid.org/0000-0002-6007-3063

http://orcid.org/0000-0001-6144-1318

http://orcid.org/0000-0002-6645-3294


failure and its resistance to elastic deformation under load,respectively; high strength and stiffness of advanced polymerfibers require highly oriented polymer molecules achieved byspecially designed fiber spinning and drawing processes;toughness, property describing the ability of a material toabsorb energy without fracturing; one definition of toughness isthe amount of energy per unit volume that a material can absorbbefore failure; high toughness requires a degree of ductility orplastic deformation; electrospinning, most popular techniqueproducing continuous polymer nanofibers by jetting semidilutepolymer solutions or melts in high electric fields; macro-molecular chain orientation, preferred orientation achievedthrough various flow and drawing processes and/or self-assembly; preferred orientation is described by the orientationdistribution function or by parameters such as second and fourthLegendre order parameters ⟨P2(cos θ)⟩ and ⟨P4(cos θ)⟩;polymer crystallization, process involving folding and/orassociation of polymer chains to form ordered regions; polymerscan crystallize upon cooling from the melt or solventevaporation; most bulk polymers exhibit partial crystallinitywhen crystalline regions are mixed with amorphous regions;oriented polymers normally exhibit higher degree of crystallinitydefined as a ratio of crystalline regions to overall volume;methods of orientation analysis, characterization techniquesmeasuring the orientation distribution function (such as X-rayand electron diffraction) or specific order parameters (such aselectronic, infrared and Raman spectroscopies); Raman spec-troscopy, vibrational spectroscopy technique based on theinelastic scattering of light that enables measuring orderparameters in individual nanofibers

REFERENCES(1) Dzenis, Y. A. Structural Nanocomposites. Science 2008, 319, 419−420.(2) Ritchie, R. O.; Dzenis, Y. A. The Quest for Stronger, TougherMaterials : A Letter and a Response. Science 2008, 320, 448a.(3) Papkov, D.; Beese, A. M.; Goponenko, A.; Zou, Y.; Naraghi, M.;Espinosa, H. D.; Saha, B.; Schatz, G. C.; Moravsky, A.; Loutfy, R.;Nguyen, S. T.; Dzenis, Y. Extraordinary Improvement of the GraphiticStructure of Continuous Carbon Nanofibers Templated with DoubleWall Carbon Nanotubes. ACS Nano 2013, 7, 126−142.(4) Papkov, D.; Goponenko, A.; Compton, O. C.; An, Z.; Moravsky,A.; Li, X.-Z.; Nguyen, S. T.; Dzenis, Y. A. Improved Graphitic Structureof Continuous Carbon Nanofibers via Graphene Oxide Templating.Adv. Funct. Mater. 2013, 23, 5763−5770.(5) Reneker, D. H.; Chun, I. Nanometre Diameter Fibres of Polymer,Produced by Electrospinning. Nanotechnology 1996, 7, 216−223.(6) Dzenis, Y. A. Spinning Continuous Fibers for Nanotechnology.Science 2004, 304, 1917−1919.(7) Lannutti, J.; Reneker, D.; Ma, T.; Tomasko, D.; Farson, D.Electrospinning for Tissue Engineering Scaffolds. Mater. Sci. Eng., C2007, 27, 504−509.(8) Sill, T. J.; von Recum, H. A. Electrospinning: Applications in DrugDelivery and Tissue Engineering. Biomaterials 2008, 29, 1989−2006.(9) Cho, H.; Min, S.-Y.; Lee, T.-W. Electrospun Organic NanofiberElectronics and Photonics.Macromol. Mater. Eng. 2013, 298, 475−486.(10) Camposeo, A.; Di Benedetto, F.; Stabile, R.; Neves, A. a R.;Cingolani, R.; Pisignano, D. Laser Emission from Electrospun PolymerNanofibers. Small 2009, 5, 562−566.(11) Sun, B.; Long, Y.-Z.; Chen, Z.-J.; Liu, S.-L.; Zhang, H.-D.; Zhang,J.-C.; Han, W.-P. Recent Advances in Flexible and StretchableElectronic Devices via Electrospinning. J. Mater. Chem. C 2014, 2,1209−1219.(12) Papkov, D.; Zou, Y.; Andalib, M. N.; Goponenko, A.; Cheng, S.Z. D.; Dzenis, Y. A. Simultaneously Strong and Tough UltrafineContinuous Nanofibers. ACS Nano 2013, 7, 3324−3331.

(13) Ritchie, R. O. The Conflicts between Strength and Toughness.Nat. Mater. 2011, 10, 817−822.(14) Fennessey, S. F.; Farris, R. J. Fabrication of Aligned andMolecularly Oriented Electrospun Polyacrylonitrile Nanofibers and theMechanical Behavior of Their Twisted Yarns. Polymer 2004, 45, 4217−4225.(15) Naraghi, M.; Arshad, S. N.; Chasiotis, I. Molecular Orientationand Mechanical Property Size Effects in Electrospun PolyacrylonitrileNanofibers. Polymer 2011, 52, 1612−1618.(16) Dersch, R.; Liu, T.; Schaper, A. K.; Greiner, A.; Wendorff, J. H.Electrospun Nanofibers: Internal Structure and Intrinsic Orientation. J.Polym. Sci., Part A: Polym. Chem. 2003, 41, 545−553.(17) Maleckis, K.; Dzenis, Y. Continuous DNA Nanofibers withExtraordinary Mechanical Properties and High Molecular Orientation.Macromol. Mater. Eng. 2018, 1800302.(18) Tosaka, M.; Yamaguchi, K.; Tsuji, M. Latent Orientation in theSkin Layer of Electrospun Isotactic Polystyrene Ultrafine Fibers.Polymer 2010, 51, 547−553.(19) Richard-Lacroix, M.; Pellerin, C. Orientation and Structure ofSingle Electrospun Nanofibers of Poly(Ethylene Terephthalate) byConfocal Raman Spectroscopy.Macromolecules 2012, 45, 1946−1953.(20) High-Performance Structural Fibers for Advanced Polymer MatrixComposites; National Academies Press: Washington, D.C., 2005.(21) Hou, H.; Ge, J. J.; Zeng, J.; Li, Q.; Reneker, D. H.; Greiner, A.;Cheng, S. Z. D. Electrospun Polyacrylonitrile Nanofibers Containing aHigh Concentration of Well-Aligned Multiwall Carbon Nanotubes.Chem. Mater. 2005, 17, 967−973.(22) Bunsell, R. A. In Fibre Reinforcements for Composite Materials;Pipes, B. R., Ed.; Composite Materials; Elsevier: New York, 1988; Vol.2.(23) Naraghi, M.; Filleter, T.; Moravsky, A.; Locascio, M.; Loutfy, R.O.; Espinosa, H. D. A Multiscale Study of High Performance Double-Walled Nanotube−Polymer Fibers. ACS Nano 2010, 4, 6463−6476.(24) Zhang, M.; Atkinson, K. R.; Baughman, R. H. MultifunctionalCarbon Nanotube Yarns by Downsizing an Ancient Technology.Science 2004, 306, 1358−1361.(25) Motta, M.; Moisala, A.; Kinloch, I. A.; Windle, A. H. HighPerformance Fibres from “Dog Bone” Carbon Nanotubes. Adv. Mater.2007, 19, 3721−3726.(26) Miaudet, P.; Badaire, S.; Maugey, M.; Derre, A.; Pichot, V.;Launois, P.; Poulin, P.; Zakri, C. Hot-Drawing of Single and MultiwallCarbon Nanotube Fibers for High Toughness and Alignment. NanoLett. 2005, 5, 2212−2215.(27) Wu, X.-F.; Salkovskiy, Y.; Dzenis, Y. A. Modeling of SolventEvaporation from Polymer Jets in Electrospinning. Appl. Phys. Lett.2011, 98, 223108.(28) Charch,W.H.;Moseley,W.W. Structure-Property Relationshipsin Synthetic Fibers. Text. Res. J. 1959, 29, 525−535.(29) Watt, W.; Phillips, L. N.; Johnson, W. High-Modulus, HighStrength Carbon Fibers. Eng. 1969, 221, 815.(30) Bastiaansen, C. W. M. Tensile Strength of Solution-Spun, Ultra-Drawn Ultra-High Molecular Weight Polyethylene Fibres: 1. Influenceof Fibre Diameter. Polymer 1992, 33, 1649−1652.(31) Richard-Lacroix, M.; Pellerin, C. Molecular Orientation inElectrospun Fibers: From Mats to Single Fibers.Macromolecules 2013,46, 9473−9493.(32) Kakade, M. V.; Givens, S.; Gardner, K.; Lee, K. H.; Chase, D. B.;Rabolt, J. F. Electric Field Induced Orientation of Polymer Chains inMacroscopically Aligned Electrospun Polymer Nanofibers. J. Am.Chem. Soc. 2007, 129, 2777−2782.(33) Kongkhlang, T.; Tashiro, K.; Kotaki, M.; Chirachanchai, S.Electrospinning as a New Technique To Control the CrystalMorphology and Molecular Orientation of Polyoxymethylene Nano-fibers. J. Am. Chem. Soc. 2008, 130, 15460−15466.(34) Arinstein, A.; Burman, M.; Gendelman, O.; Zussman, E. Effect ofSupramolecular Structure on Polymer Nanofibre Elasticity. Nat.Nanotechnol. 2007, 2, 59−62.

ACS Nano Review


AC


(35) Chen, W.; Tao, X.; Liu, Y. Carbon Nanotube-ReinforcedPolyurethane Composite Fibers. Compos. Sci. Technol. 2006, 66, 3029−3034.(36)Morgan, P.Carbon Fibers and Their Composites; CRC Press: BocaRaton, FL, 2005.(37) Ward, I. M.; Coates, P. D.; Dumoulin, M. M. Solid PhaseProcessing of Polymers; Hanser Publishers, 2000.(38) Jasse, B.; Koenig, J. L. Orientational Measurements in PolymersUsing Vibrational Spectroscopy. J. Macromol. Sci., Polym. Rev. 1979, 17,61.(39) Bashir, Z.; Tipping, A. R.; Church, S. P. Orientation Studies inPolyacrylonitrile Films Uniaxially Drawn in the Solid State. Polym. Int.1994, 33, 9−17.(40) Chan, K. H. K.; Wong, S. Y.; Li, X.; Zhang, Y. Z.; Lim, P. C.; Lim,C. T.; Kotaki, M.; He, C. B. Effect of Molecular Orientation onMechanical Property of Single Electrospun Fiber of Poly[(R)-3-Hydroxybutyrate-Co-(R)-3-Hydroxyvalerate]. J. Phys. Chem. B 2009,113, 13179−13185.(41) Griffiths, P. R.; De Haseth, J. A. Fourier Transform InfraredSpectrometry; Wiley-Interscience, 2007.(42) Ma, X.; Zachariah, M. R.; Zangmeister, C. D. CrumpledNanopaper from Graphene Oxide. Nano Lett. 2012, 12, 486−489.(43) Cheng, Y.-W.; Lu, H.-A.; Wang, Y.-C.; Thierry, A.; Lotz, B.;Wang, C. Syndiotactic Polystyrene Nanofibers Obtained from High-Temperature Solution Electrospinning Process. Macromolecules 2010,43, 2371−2376.(44) Tu, D.; Pagliara, S.; Camposeo, A.; Persano, L.; Cingolani, R.;Pisignano, D. Single Light-Emitting Polymer Nanofiber Field-EffectTransistors. Nanoscale 2010, 2, 2217−2222.(45) Koenig, J. L. Spectroscopy of Polymers, 2nd ed.; Elsevier, 1999.(46) Henrici-Olive, G.; Olive, S. Molecular Interactions andMacroscopic Properties of Polyacrylonitrile and Model Substances.In Chemistry. Advances in Polymer Science; Advances in PolymerScience; Springer: Berlin, 1979; Vol. 32, pp 123−152.(47) Litovchenko, G. D. Infrared Dichroism of the Band of ValenceVibrations of the C≡N Bonds in Polyacrylonitrile Fibers. J. Appl.Spectrosc. 1973, 18, 781−783.(48) Gu, S.; Wu, Q.; Ren, J.; Vancso, G. J. Mechanical Properties of aSingle Electrospun Fiber and Its Structures.Macromol. Rapid Commun.2005, 26, 716−720.(49) Pagliara, S.; Vitiello, M. S.; Camposeo, A.; Polini, A.; Cingolani,R.; Scamarcio, G.; Pisignano, D. Optical Anisotropy in Single Light-Emitting Polymer Nanofibers. J. Phys. Chem. C 2011, 115, 20399−20405.(50) Gong, L.; Chase, D. B.; Noda, I.; Liu, J.; Martin, D. C.; Ni, C.;Rabolt, J. F. Discovery of β-Form Crystal Structure in ElectrospunPoly[(R)-3-Hydroxybutyrate-Co-(R)-3-Hydroxyhexanoate](PHBHx) Nanofibers: From Fiber Mats to Single Fibers. Macro-molecules 2015, 48, 6197−6205.(51) Behabtu, N.; Young, C. C.; Tsentalovich, D. E.; Kleinerman, O.;Wang, X.; Ma, A.W. K.; Bengio, E. A.; ter Waarbeek, R. F.; de Jong, J. J.;Hoogerwerf, R. E.; Fairchild, S. B.; Ferguson, J. B.; Maruyama, B.;Kono, J.; Talmon, Y.; Cohen, Y.; Otto, M. J.; Pasquali, M. Strong, Light,Multifunctional Fibers of Carbon Nanotubes with Ultrahigh Con-ductivity. Science 2013, 339, 182−186.(52) Wang, C.; Chien, H.-S.; Yan, K.-W.; Hung, C.-L.; Hung, K.-L.;Tsai, S.-J.; Jhang, H.-J. Correlation between Processing Parameters andMicrostructure of Electrospun Poly(D,l-Lactic Acid) Nanofibers.Polymer 2009, 50, 6100−6110.(53) Isakov, D. V.; de Matos Gomes, E.; Vieira, L. G.; Dekola, T.;Belsley, M. S.; Almeida, B. G. Oriented Single-Crystal-like MolecularArrangement of Optically Nonlinear 2-Methyl-4-Nitroaniline inElectrospun Nanofibers. ACS Nano 2011, 5, 73−78.(54) Bashir, Z.; Church, S. P.; Price, D.M. The Formation of Polymer-Solvent Complexes of Polyacrylonitrile from Organic SolventsContaining Carbonyl Groups. Acta Polym. 1993, 44, 211−218.(55) Miller, R. W.; Murayama, T. Dynamic Mechanical Properties ofPartially Oriented Polyester (POY) and Draw-Textured Polyester(PTY) Yarns. J. Appl. Polym. Sci. 1984, 29, 933−939.

(56) Furushima, Y.; Ishikiriyama, K.; Higashioji, T. The CharacteristicLength of Cooperative Rearranging Region for Uniaxial DrawnPoly(Ethylene Terephthalate) Films. Polymer 2013, 54, 4078−4084.(57) Delpouve, N.; Delbreilh, L.; Stoclet, G.; Saiter, A.; Dargent, E.Structural Dependence of the Molecular Mobility in the AmorphousFractions of Polylactide. Macromolecules 2014, 47, 5186−5197.(58) Delpouve, N.; Lixon, C.; Saiter, A.; Dargent, E.; Grenet, J.Amorphous Phase Dynamics at the Glass Transition in Drawn Semi-Crystalline Polyester. J. Therm. Anal. Calorim. 2009, 97, 541−546.(59) Hamonic, F.; Miri, V.; Saiter, A.; Dargent, E. Rigid AmorphousFraction versus Oriented Amorphous Fraction in Uniaxially DrawnPolyesters. Eur. Polym. J. 2014, 58, 233−244.(60) Wang, C.; Lee, M.-F.; Jao, C.-H. Phase Transition ofPoly(Ethylene Terephthalate) in Nanofibers Electrospun fromPhenol-Based Solution. Eur. Polym. J. 2014, 52, 127−136.(61) Chen, H.; Liu, Z.; Cebe, P. Chain Confinement in ElectrospunNanofibers of PET with Carbon Nanotubes. Polymer 2009, 50, 872−880.(62) Ma, Q.; Pyda, M.; Mao, B.; Cebe, P. Relationship between theRigid Amorphous Phase andMesophase in Electrospun Fibers. Polymer2013, 54, 2544−2554.(63) Monnier, X.; Delpouve, N.; Basson, N.; Guinault, A.; Domenek,S.; Saiter, A.; Mallon, P. E.; Dargent, E. Molecular Dynamics inElectrospun Amorphous Plasticized Polylactide Fibers. Polymer 2015,73, 68−78.(64) Zhang, P.; Tian, R.; Na, B.; Lv, R.; Liu, Q. IntermolecularOrdering as the Precursor for Stereocomplex Formation in theElectrospun Polylactide Fibers. Polymer 2015, 60, 221−227.(65) Li, K.; Mao, B.; Cebe, P. Electrospun Fibers of Poly(EthyleneTerephthalate) Blended with Poly(Lactic Acid). J. Therm. Anal.Calorim. 2014, 116, 1351−1359.(66) Lv, R.; Na, B.; Tian, N.; Zou, S.; Li, Z.; Jiang, S. MesophaseFormation and Its Thermal Transition in the Stretched GlassyPolylactide Revealed by Infrared Spectroscopy. Polymer 2011, 52,4979−4984.(67) Ma, Q.; Mao, B.; Cebe, P. Chain Confinement in ElectrospunNanocomposites: Using Thermal Analysis to Investigate Polymer−filler Interactions. Polymer 2011, 52, 3190−3200.(68) Mao, B.; Geers, K.; Hu, S.; Mancera, M.; Sandoval, M.; Port, J.;Zhu, Y.; Cebe, P. Properties of Aligned Poly(L-Lactic Acid)Electrospun Fibers. J. Appl. Polym. Sci. 2015, 132, 41779.(69) Zong, X. H.; Kim, K.; Fang, D.; Ran, S.; Hsiao, B. S.; Chu, B.Structure and Process Relationship of Electrospun BioabsorbableNanofiber Membranes. Polymer 2002, 43, 4403−4412.(70)Hamonic, F.; Prevosto, D.; Dargent, E.; Saiter, A. Contribution ofChain Alignment and Crystallization in the Evolution of Cooperativityin Drawn Polymers. Polymer 2014, 55, 2882−2889.(71) Kremer, F.; Schonhals, A. Broadband Dielectric Spectroscopy;Springer: Berlin, 2003.(72) Esposito, A.; Delpouve, N.; Causin, V.; Dhotel, A.; Delbreilh, L.;Dargent, E. From a Three-Phase Model to a Continuous Description ofMolecular Mobility in Semicrystalline Poly(Hydroxybutyrate- Co-Hydroxyvalerate). Macromolecules 2016, 49, 4850−4861.(73) Tahalyani, J.; Datar, S.; Balasubramanian, K. Investigation ofDielectric Properties of Free Standing Electrospun Nonwoven Mat. J.Appl. Polym. Sci. 2018, 135, 46121.(74) Santos, J. P. F.; da Silva, A. B.; Arjmand, M.; Sundararaj, U.;Bretas, R. E. S. Nanofibers of Poly(Vinylidene Fluoride)/CopperNanowire: Microstructural Analysis and Dielectric Behavior. Eur.Polym. J. 2018, 101, 46−55.(75) Tahalyani, J.; Rahangdale, K. K.; Aepuru, R.; Kandasubramanian,B.; Datar, S. Dielectric Investigation of a Conducting FibrousNonwoven Porous Mat Fabricated by a One-Step Facile Electro-spinning Process. RSC Adv. 2016, 6, 36588−36598.(76) Dhotel, A.; Rijal, B.; Delbreilh, L.; Dargent, E.; Saiter, A.Combining Flash DSC, DSC and Broadband Dielectric Spectroscopyto Determine Fragility. J. Therm. Anal. Calorim. 2015, 121, 453−461.

ACS Nano Review


AD


(77) Komorski, R. A., Ed. High Resolution NMR Spectroscopy ofSynthetic Polymers in Bulk; VCH Publishers Inc.: Deerfield Beach, FL,1986.(78) Ando, I.; Asakura, T. Solid State NMR of Polymers; Elsevier, 1998.(79) Guo, W.; Fung, B. M. Determination of the Order Parameters ofLiquid Crystals from Carbon-13 Chemical Shifts. J. Chem. Phys. 1991,95, 3917−3923.(80) Wei, Y.; Lee, D.-K.; Ramamoorthy, A. Solid-State 13 C NMRChemical Shift Anisotropy Tensors of Polypeptides. J. Am. Chem. Soc.2001, 123, 6118−6126.(81) Henrichs, P. M. Molecular Orientation and Structure in SolidPolymers with 13C NMR: A Study of Biaxial Films of Poly(EthyleneTerephthalate). Macromolecules 1987, 20, 2099−2112.(82) Asakura, T.; Yeo, J. H.; Demura, M.; Itoh, T.; Fujito, T.; Imanari,M.; Nicholson, L. K.; Cross, T. A. Structural Analysis of UniaxiallyAligned Polymers Using Solid-State Nitrogen-15NMR.Macromolecules1993, 26, 6660−6663.(83) Yeo, J.-H.; Demura, M.; Asakura, T.; Fujito, T.; Imanari, M.;Nicholson, L. K.; Cross, T. A. Structural Analysis of Highly OrientedPoly(p-Phenylene-Terephthalamide) by 15N Solid-State NuclearMagnetic Resonance. Solid State Nucl. Magn. Reson. 1994, 3, 209−218.(84) Asakura, T.; Yeo, J.-H.; Ando, I. Structure of Polyamide Fibers inthe Non-Crystalline Domain Studied by 15N Solid State NMR. Polym.J. 1994, 26, 229−233.(85) Brandolini, A. J.; Alvey, M. D.; Dybowski, C. The CrystalliteOrientation Distribution Functions of Deformed Polytetrafluoro-ethylene: Multiple-Pulse 19F NMR Spectroscopy. J. Polym. Sci.,Polym. Phys. Ed. 1983, 21, 2511−2524.(86) Saito, H.; Ando, I. High-Resolution Solid-State NMR Studies ofSynthetic and Biological Macromolecules. Annu. Rep. NMR Spectrosc.1989, 21, 209−290.(87) Falk, U.; Spiess, H. W. Phase Behaviour and MacroscopicAlignment of Rigid Chain Polyesters with Short Flexible Side Chains.Makromol. Chem., Rapid Commun. 1989, 10, 149−155.(88) Hentschel, R.; Sillescu, H.; Spiess, H. W. OrientationalDistribution of Polymer Chains Studied by 2H n.m.r. Line Shapes.Polymer 1981, 22, 1516−1521.(89) Yeo, J.-H.; Demura, M.; Asakura, T.; Fujito, T.; Imanari, M.;Nicholson, L. K.; Cross, T. A. Structural Analysis of Highly OrientedPoly(p-Phenylene-Terephthalamide) by 15N Solid-State NuclearMagnetic Resonance. Solid State Nucl. Magn. Reson. 1994, 3, 209−218.(90) van Beek, J. D.; Hess, S.; Vollrath, F.; Meier, B. H. TheMolecularStructure of Spider Dragline Silk: Folding and Orientation of theProtein Backbone. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 10266−10271.(91) Asakura, T.; Ashida, J.; Yamane, T.; Kameda, T.; Nakazawa, Y.;Ohgo, K.; Komatsu, K. A Repeated β-Turn Structure in Poly(Ala-Gly)as a Model for Silk I of Bombyx Mori Silk Fibroin Studied with Two-Dimensional Spin-Diffusion NMR under off Magic Angle Spinning andRotational Echo Double Resonance. J. Mol. Biol. 2001, 306, 291−305.(92) Hyun Kim, S.; Sik Nam, Y.; Seung Lee, T.; Ho Park, W. SilkFibroin Nanofiber. Electrospinning, Properties, and Structure. Polym. J.2003, 35, 185−190.(93) Mote, K. R.; Agarwal, V.; Madhu, P. K. Five Decades ofHomonuclear Dipolar Decoupling in Solid-State NMR: Status andOutlook. Prog. Nucl. Magn. Reson. Spectrosc. 2016, 97, 1−39.(94) Mithu, V. S.; Pratihar, S.; Paul, S.; Madhu, P. K. Efficiency ofHeteronuclear Dipolar Decoupling Schemes in Solid-State NMR:Investigation of Effective Transverse Relaxation Times. J. Magn. Reson.2012, 220, 8−17.(95) Kovacs, H.; Moskau, D.; Spraul, M. Cryogenically CooledProbes-a Leap in NMR Technology. Prog. Nucl. Magn. Reson. Spectrosc.2005, 46, 131−155.(96)Matsuzaki, K.; Uryu, T.; Ishigure, K.; Takeuchi, M. NMR Spectraof Polyacrylonitriles. J. Polym. Sci., Part B: Polym. Lett. 1966, 4, 93−96.(97) Hong, M. Oligomeric Structure, Dynamics, and Orientation ofMembrane Proteins from Solid-State NMR. Structure 2006, 14, 1731−1740.

(98) Wang, Y.; Xu, L.; Wang, M.; Pang, W.; Ge, X. StructuralIdentification of Polyacrylonitrile during Thermal Treatment bySelective 13 C Labeling and Solid-State 13 C NMR Spectroscopy.Macromolecules 2014, 47, 3901−3908.(99) Fong, H.; Chun, I.; Reneker, D. H. Beaded Nanofibers Formedduring Electrospinning. Polymer 1999, 40, 4585−4592.(100) Richard-Lacroix, M.; Pellerin, C. Orientation and PartialDisentanglement in Individual Electrospun Fibers: Diameter Depend-ence and Correlation with Mechanical Properties. Macromolecules2015, 48, 4511−4519.(101) Papkov, D.; Beese, A. M.; Goponenko, A.; Zou, Y.; Naraghi, M.;Espinosa, H. D.; Saha, B.; Schatz, G. C.; Moravsky, A.; Loutfy, R.;Nguyen, S. T.; Dzenis, Y. A. Extraordinary Improvement of theGraphitic Structure of Continuous Carbon Nanofibers Templated withDouble Wall Carbon Nanotubes. ACS Nano 2013, 7, 126−142.(102) Yoshioka, T.; Dersch, R.; Tsuji, M.; Schaper, A. K. OrientationAnalysis of Individual Electrospun PE Nanofibers by TransmissionElectron Microscopy. Polymer 2010, 51, 2383−2389.(103) Ma, X.; Liu, J.; Ni, C.; Martin, D. C.; Chase, D. B.; Rabolt, J. F.Molecular Orientation in Electrospun Poly(Vinylidene Fluoride)Fibers. ACS Macro Lett. 2012, 1, 428−431.(104) Davidson, J. A.; Jung, H.-T.; Hudson, S. D.; Percec, S.Investigation ofMolecular Orientation inMelt-SpunHigh AcrylonitrileFibers. Polymer 2000, 41, 3357−3364.(105) Drake, A. F. Polarisation Modulation-the Measurement ofLinear and Circular Dichroism. J. Phys. E: Sci. Instrum. 1986, 19, 170.(106) Lewis, I. R.; Edwards, H. Handbook of Raman Spectroscopy:From the Research Laboratory to the Process Line; CRC Press, 2001.(107) Singamaneni, S.; Bertoldi, K.; Chang, S.; Jang, J.-H.; Young, S.L.; Thomas, E. L.; Boyce, M. C.; Tsukruk, V. V. Bifurcated MechanicalBehavior of Deformed Periodic Porous Solids. Adv. Funct. Mater. 2009,19, 1426−1436.(108) Kattumenu, R.; Lee, C. H.; Bliznyuk, V. N.; Singamaneni, S.Micro-Raman Spectroscopy of Nanostructures. Raman Spectroscopy forNanomaterials Characterization; Springer: Berlin, 2012; pp 417−444.(109) Bower, D. I. Investigation of Molecular OrientationDistributions by Polarized Raman Scattering and Polarized Fluo-rescence. J. Polym. Sci. Polym. Phys. Ed. 1972, 10, 2135−2153.(110) Bower, D. I. Raman Scattering from an Assembly of PartiallyOriented Scatterers. J. Phys. B: At. Mol. Phys. 1976, 9, 3275.(111) Frisk, S.; Ikeda, R. M.; Chase, D. B.; Rabolt, J. F. Determinationof the Molecular Orientation of Poly(Propylene Terephthalate) FibersUsing Polarized Raman Spectroscopy: A Comparison of Methods.Appl. Spectrosc. 2004, 58, 279−286.(112) Purvis, J.; Bower, D. I. Molecular Orientation in Poly(EthyleneTerephthalate) by Means of Laser−raman Spectroscopy. J. Polym. Sci.,Polym. Phys. Ed. 1976, 14, 1461−1484.(113) Richard-Lacroix, M.; Pellerin, C. Accurate New Method forMolecular Orientation Quantification Using Polarized Raman Spec-troscopy. Macromolecules 2013, 46, 5561−5569.(114) Rousseau, M.-E.; Lefevre, T.; Beaulieu, L.; Asakura, T.; Pezolet,M. Study of Protein Conformation and Orientation in Silkworm andSpider Silk Fibers Using Raman Microspectroscopy. Biomacromolecules2004, 5, 2247−2257.(115) Richard-Lacroix, M.; Pellerin, C. Raman Spectroscopy ofIndividual Poly(Ethylene Oxide) Electrospun Fibers: Effect of theCollector on Molecular Orientation. Vib. Spectrosc. 2017, 91, 92−98.(116) Papkov, D.; Pellerin, C.; Dzenis, Y. A. Polarized Raman Analysisof Polymer Chain Orientation in Ultrafine Individual Nanofibers withVariable Low Crystallinity. Macromolecules 2018, 51, 8746−8751.(117) Takajima, T. Advanced Fiber Spinning Technology; WoodheadPublishing, 1994.(118) Galvis, L.; Dunlop, J. W. C.; Duda, G.; Fratzl, P.; Masic, A.Polarized Raman Anisotropic Response of Collagen in Tendon:Towards 3D Orientation Mapping of Collagen in Tissues. PLoS One2013, 8, No. e63518.(119) Thomas, G. J.; Benevides, J. M.; Overman, S. a.; Ueda, T.;Ushizawa, K.; Saitoh, M.; Tsuboi, M. Polarized Raman Spectra ofOriented Fibers of ADNA and BDNA: Anisotropic and Isotropic Local

ACS Nano Review


AE


Raman Tensors of Base and Backbone Vibrations. Biophys. J. 1995, 68,1073−1088.(120) McGraw, G. E. Investigation of Polyester Structure by LaserRaman Spectroscopy. In Polymer Characterization: InterdisciplinaryApproaches; Craver, C. D., Ed.; Springer: Boston, MA, 1971; pp 37−46.(121) Melveger, A. J. Laser-Raman Study of Crystallinity Changes inPoly(Ethylene Terephthalate). J. Polym. Sci. Part A-2 Polym. Phys. 1972,10, 317−322.(122) Louden, J. D. Crystallinity in Poly(Aryl Ether Ketone) FilmsStudied by Raman Spectroscopy. Polym. Commun. 1986, 27, 82−84.(123) Everall, N. J.; Lumsdon, J.; Chalmers, J. M.; Mason, N. The Useof Polarised Fourier Transform Raman Spectroscopy in MorphologicalStudies of Uniaxially Oriented PEEK Fibressome PreliminaryResults. Spectrochim. Acta Part A Mol. Spectrosc. 1991, 47, 1305−1311.(124) Adar, F.; Noether, H. Raman Microprobe Spectra of Spin-Oriented and Drawn Filaments of Poly(Ethylene Terephthalate).Polymer 1985, 26, 1935−1943.(125) Bulkin, B. J.; Lewin, M.; DeBlase, F. J. Conformational Change,Chain Orientation, and Crystallinity in Poly(Ethylene Terephthalate)Yarns: Raman Spectroscopic Study. Macromolecules 1985, 18, 2587−2594.(126) Natarajan, S.; Michielsen, S. Determination of Density andBirefringence of Poly(Ethylene Terephthalate) Fibers Using RamanMicroscopy. J. Appl. Polym. Sci. 1999, 73, 943−952.(127) Schmidt, P.; Hendra, P. J. The Application of Fourier-Transform Raman Spectroscopy to the Determination of Conforma-tion in Poly(ε-Caprolactam) Chains. Spectrochim. Acta Part A Mol.Spectrosc. 1994, 50, 1999−2004.(128) Everall, N.; Tayler, P.; Chalmers, J. M.; MacKerron, D.;Ferwerda, R.; van der Maas, J. Study of Density and Orientation inPoly(Ethylene Terephthalate) Using Fourier Transform RamanSpectroscopy and Multivariate Data Analysis. Polymer 1994, 35,3184−3192.(129) Huang, Y. S.; Koenig, J. L. Raman Spectra of Polyacrylonitrile.Appl. Spectrosc. 1971, 25, 620−622.(130) Gross, L.; Mohn, F.; Moll, N.; Liljeroth, P.; Meyer, G. TheChemical Structure of a Molecule Resolved by Atomic ForceMicroscopy. Science 2009, 325, 1110−1114.(131) Fukuma, T.; Kilpatrick, J. I.; Jarvis, S. P. Phase ModulationAtomic Force Microscope with True Atomic Resolution. Rev. Sci.Instrum. 2006, 77, 123703.(132) Mou, J.; Czajkowsky, D. M.; Zhang, Y.; Shao, Z. High-Resolution Atomic-Force Microscopy of DNA: The Pitch of theDouble Helix. FEBS Lett. 1995, 371, 279−282.(133) Cerreta, A.; Vobornik, D.; Dietler, G. Fine DNA StructureRevealed by Constant Height Frequency Modulation AFM Imaging.Eur. Polym. J. 2013, 49, 1916−1922.(134) Rebouillat, S.; Donnet, J. B.; Wang, T. K. Surface Micro-structure of a Kevlar* Aramid Fibre Studied by Direct Atomic ForceMicroscopy. Polymer 1997, 38, 2245−2249.(135) Dzenis, Y. A.; Reneker, D. H.; Tsukruk, V. V.; Patil, R. FractalAnalysis of Surfaces of Advanced Reinforcing Fibers by Atomic ForceMicroscopy. Compos. Interfaces 1994, 2, 307−319.(136) Strawhecker, K. E.; Cole, D. P. Morphological and LocalMechanical Surface Characterization of Ballistic Fibers via AFM. J.Appl. Polym. Sci. 2014, 131, 40880.(137) Sakamoto, H.; Asakawa, H.; Fukuma, T.; Fujita, S.; Suye, S. I.Atomic Force Microscopy Visualization of Hard Segment Alignment inStretched Polyurethane Nanofibers Prepared by Electrospinning. Sci.Technol. Adv. Mater. 2014, 15, 015008.(138) Xu, Y.; Gao, Y.; Wang, X.; Jiang, J.; Hou, J.; Li, Q. InternalStructure of Amorphous Electrospun Nanofiber: Oriented MolecularChains. Macromol. Mater. Eng. 2017, 302, 70019.(139) Chlanda, A.; Kijen ska, E.; Rinoldi, C.; Tarnowski, M.;Wierzchon, T.; Swieszkowski, W. Structure and Physico-MechanicalProperties of Low Temperature Plasma Treated Electrospun Nano-fibrous Scaffolds Examined with Atomic Force Microscopy. Micron2018, 107, 79−84.

(140) Stockdale, T. A.; Strawhecker, K. E.; Sandoz-Rosado, E. J.;Wetzel, E. D. A Rapid FIB-Notch Technique for Characterizing theInternal Morphology of High-Performance Fibers. Mater. Lett. 2016,176, 173.(141) Strawhecker, K. E.; Sandoz-Rosado, E. J.; Stockdale, T. A.;Laird, E. D. Interior Morphology of High-Performance PolyethyleneFibers Revealed by Modulus Mapping. Polymer 2016, 103, 224.(142) Roenbeck, M. R.; Sandoz-Rosado, E. J.; Cline, J.; Wu, V.; Moy,P.; Afshari, M.; Reichert, D.; Lustig, S. R.; Strawhecker, K. E. Probingthe Internal Structures of Kevlar®fibers and Their Impacts onMechanical Performance. Polymer 2017, 128, 200.(143) Sandoz-Rosado, E.; Roenbeck, M. R.; Strawhecker, K. E.Quantifying High-Performance Material Microstructure Using Nano-mechanical Tools with Visual and Frequency Analysis. Scanning 2018,2018, 1−12.(144) McDaniel, P. B.; Strawhecker, K. E.; Deitzel, J. M.; Gillespie, J.W.Nanoscale Interfibrillar Adhesion inUHMWPEFibers. J. Polym. Sci.,Part B: Polym. Phys. 2018, 56, 391−401.(145) Hill, R.; Faridur Rahman, F. H. M. Advances in Helium IonMicroscopy. Nucl. Instrum. Methods Phys. Res., Sect. A 2011, 645, 96−101.(146) Livengood, R. H.; Tan, S.; Hallstein, R.; Notte, J.; McVey, S.;Faridur Rahman, F. H. M. The Neon Gas Field Ion Source - A FirstCharacterization of Neon Nanomachining Properties. Nucl. Instrum.Methods Phys. Res., Sect. A 2011, 645, 136−140.(147) McAllister, Q. P.; Gillespie, J. W.; VanLandingham, M. R.Evaluation of the Three-Dimensional Properties of Kevlar acrossLength Scales. J. Mater. Res. 2012, 27, 1824−1837.(148) Camposeo, A.; Greenfeld, I.; Tantussi, F.; Pagliara, S.; Moffa,M.; Fuso, F.; Allegrini, M.; Zussman, E.; Pisignano, D. LocalMechanical Properties of Electrospun Fibers Correlate to TheirInternal Nanostructure. Nano Lett. 2013, 13, 5056−5062.(149) Wang, X.; Xu, Y.; Jiang, Y.; Jiang, J.; Turng, L. S.; Li, Q. Core/Shell Structure of Electrospun Polycarbonate Nanofibers. Polym. Test.2018, 70, 498−502.(150) Kocun, M.; Labuda, A.; Meinhold, W.; Revenko, I.; Proksch, R.Fast, High Resolution, and Wide Modulus Range NanomechanicalMapping with Bimodal Tapping Mode. ACS Nano 2017, 11, 10097−10105.(151) Pittenger, B.; Erina, N.; Su, C. Application Note # 128Quantitative Mechanical Property Mapping at the Nanoscale withPeakForce QNM. Bruker Corp. 2012; 1−12.(152) Kirz, J.; Jacobsen, C. The History and Future of X-RayMicroscopy. J. Phys. Conf. Ser. 2009, 186, 012001.(153) Macrander, A. T.; Huang, X. Synchrotron X-Ray Optics. Annu.Rev. Mater. Res. 2017, 47, 135−152.(154) Roth, T.; Detlefs, C.; Snigireva, I.; Snigirev, A. X-RayDiffractionMicroscopy Based on Refractive Optics.Opt. Commun. 2015, 340, 33−38.(155) Liu, W.; Ice, G. E.; Tischler, J. Z.; Khounsary, A.; Liu, C.;Assoufid, L.; Macrander, A. T. Short Focal Length Kirkpatrick-BaezMirrors for a Hard x-Ray Nanoprobe. Rev. Sci. Instrum. 2005, 76,113701.(156) Sakdinawat, A.; Attwood, D. Nanoscale X-Ray Imaging. Nat.Photonics 2010, 4, 840−848.(157) Holt, M.; Harder, R.; Winarski, R.; Rose, V. Nanoscale Hard X-Ray Microscopy Methods for Materials Studies. Annu. Rev. Mater. Res.2013, 43, 183−211.(158) Mohacsi, I.; Vartiainen, I.; Rosner, B.; Guizar-Sicairos, M.;Guzenko, V. A.; McNulty, I.; Winarski, R.; Holt, M. V.; David, C.Interlaced Zone Plate Optics for Hard X-Ray Imaging in the 10 NmRange. Sci. Rep. 2017, 7, 43624.(159) Hernandez Cruz, D.; Rousseau, M.-E.; West, M. M.; Pezolet,M.; Hitchcock, A. P. Quantitative Mapping of the Orientation ofFibroin β-Sheets in B. Mori Cocoon Fibers by Scanning TransmissionX-Ray Microscopy. Biomacromolecules 2006, 7, 836−843.(160) Ade, H.; Stoll, H. Near-Edge X-Ray Absorption Fine-StructureMicroscopy of Organic and Magnetic Materials. Nat. Mater. 2009, 8,281−290.

ACS Nano Review


AF


(161) Ade, H.; Hsiao, B. X-Ray Linear DichroismMicroscopy. Science1993, 262, 1427−1429.(162) Howells, M. R.; Beetz, T.; Chapman, H. N.; Cui, C.; Holton, J.M.; Kirz, J.; Lima, E.;Marchesini, S.;Miao, H.; Sayre, D.; Shapiro, D. A.;Spence, J. C. H.; Starodub, D. An Assessment of the ResolutionLimitation Due to Radiation-Damage in X-Ray DiffractionMicroscopy.J. Electron Spectrosc. Relat. Phenom. 2009, 170, 4−12.(163) Holton, J. M. A Beginner’s Guide to Radiation Damage. J.Synchrotron Radiat. 2009, 16, 133−142.(164) Egerton, R. F. Control of Radiation Damage in the TEM.Ultramicroscopy 2013, 127, 100−108.(165) Coffey, T.; Urquhart, S. G.; Ade, H. Characterization of theEffects of Soft X-Ray Irradiation on Polymers. J. Electron Spectrosc. Relat.Phenom. 2002, 122, 65−78.(166) Cui, H.; Pashuck, E. T.; Velichko, Y. S.; Weigand, S. J.;Cheetham, A. G.; Newcomb, C. J.; Stupp, S. I. Spontaneous and X-Ray-Triggered Crystallization at Long Range in Self-Assembling FilamentNetworks. Science 2010, 327, 555−559.(167) Polvino, S. M.; Murray, C. E.; Kalenci, O.; Noyan, I. C.; Lai, B.;Cai, Z. Synchrotron Microbeam X-Ray Radiation Damage inSemiconductor Layers. Appl. Phys. Lett. 2008, 92, 224105.(168) Dierolf, M.; Menzel, A.; Thibault, P.; Schneider, P.; Kewish, C.M.; Wepf, R.; Bunk, O.; Pfeiffer, F. Ptychographic X-Ray ComputedTomography at the Nanoscale. Nature 2010, 467, 436−439.(169) Shapiro, D.; Thibault, P.; Beetz, T.; Elser, V.; Howells, M.;Jacobsen, C.; Kirz, J.; Lima, E.; Miao, H.; Neiman, A. M.; Sayre, D.Biological Imaging by Soft X-Ray Diffraction Microscopy. Proc. Natl.Acad. Sci. U. S. A. 2005, 102, 15343−15346.(170) Lima, E.; Wiegart, L.; Pernot, P.; Howells, M.; Timmins, J.;Zontone, F.; Madsen, A. Cryogenic X-Ray Diffraction Microscopy forBiological Samples. Phys. Rev. Lett. 2009, 103, 198102.(171) Chu, B.; Hsiao, B. S. Small-Angle X-Ray Scattering of Polymers.Chem. Rev. 2001, 101, 1727−1761.(172) Schaff, F.; Bech, M.; Zaslansky, P.; Jud, C.; Liebi, M.; Guizar-Sicairos, M.; Pfeiffer, F. Six-Dimensional Real and Reciprocal SpaceSmall-Angle X-Ray Scattering Tomography. Nature 2015, 527, 353−356.(173) Hruszkewycz, S. O.; Highland, M. J.; Holt, M. V.; Kim, D.;Folkman, C. M.; Thompson, C.; Tripathi, A.; Stephenson, G. B.; Hong,S.; Fuoss, P. H. Imaging Local Polarization in Ferroelectric Thin Filmsby Coherent X-Ray Bragg Projection Ptychography. Phys. Rev. Lett.2013, 110, 177601.(174) Holt, M. V.; Hruszkewycz, S. O.; Murray, C. E.; Holt, J. R.;Paskiewicz, D. M.; Fuoss, P. H. Strain Imaging of NanoscaleSemiconductor Heterostructures with X-Ray Bragg ProjectionPtychography. Phys. Rev. Lett. 2014, 112, 165502.(175)Hruszkewycz, S. O.; Allain, M.; Holt, M. V.;Murray, C. E.; Holt,J. R.; Fuoss, P. H.; Chamard, V. High-Resolution Three-DimensionalStructural Microscopy by Single-Angle Bragg Ptychography. Nat.Mater. 2017, 16, 244−251.(176) Chapman, H. N.; Fromme, P.; Barty, A.; White, T. A.; Kirian, R.A.; Aquila, A.; Hunter, M. S.; Schulz, J.; DePonte, D. P.; Weierstall, U.;Doak, R. B.; Maia, F. R. N. C.; Martin, A. V.; Schlichting, I.; Lomb, L.;Coppola, N.; Shoeman, R. L.; Epp, S. W.; Hartmann, R.; et al.Femtosecond X-Ray Protein Nanocrystallography. Nature 2011, 470,73−77.(177) Seibert, M. M.; Ekeberg, T.; Maia, F. R. N. C.; Svenda, M.;Andreasson, J.; Jonsson, O.; Odic, D.; Iwan, B.; Rocker, A.; Westphal,D.; Hantke, M.; DePonte, D. P.; Barty, A.; Schulz, J.; Gumprecht, L.;Coppola, N.; Aquila, A.; Liang, M.; White, T. A.; et al. Single MimivirusParticles Intercepted and Imaged with an X-Ray Laser. Nature 2011,470, 78−81.(178) Androsch, R.; Schick, C.; Schmelzer, J. W. P. Sequence ofEnthalpy Relaxation, Homogeneous Crystal Nucleation and CrystalGrowth in Glassy Polyamide 6. Eur. Polym. J. 2014, 53, 100−108.(179) Schawe, J. E. K.; Potschke, P.; Alig, I. Nucleation Efficiency ofFillers in Polymer Crystallization Studied by Fast ScanningCalorimetry: Carbon Nanotubes in Polypropylene. Polymer 2017,116, 160−172.

(180) Paolucci, F.; Baeten, D.; Roozemond, P. C.; Goderis, B.; Peters,G. W. M. Quantification of Isothermal Crystallization of Polyamide 12:Modelling of Crystallization Kinetics and Phase Composition. Polymer2018, 155, 187.(181) Schawe, J. E. K. Analysis of Non-Isothermal Crystallizationduring Cooling and Reorganization during Heating of IsotacticPolypropylene by Fast Scanning DSC. Thermochim. Acta 2015, 603,85−93.(182) Koh, Y. P.; Grassia, L.; Simon, S. L. Structural Recovery of aSingle Polystyrene Thin Film Using Nanocalorimetry to Extend theAging Time and Temperature Range. Thermochim. Acta 2015, 603,135−141.(183) Simon, S. L.; Koh, Y. P. The Glass Transition and StructuralRecovery Using Flash DSC. Fast Scanning Calorimetry; SpringerInternational Publishing: Cham, 2016; pp 433−459.(184) Schawe, J. E. K.; Ziegelmeier, S. Determination of the ThermalShort Time Stability of Polymers by Fast Scanning Calorimetry.Thermochim. Acta 2016, 623, 80−85.(185) Monnier, X.; Fernandes Nassar, S.; Domenek, S.; Guinault, A.;Sollogoub, C.; Dargent, E.; Delpouve, N. Reduced Physical Aging Ratesof Polylactide in Polystyrene/Polylactide Multilayer Films from FastScanning Calorimetry. Polymer 2018, 150, 1−9.(186) Cebe, P.; Partlow, B. P.; Kaplan, D. L.; Wurm, A.; Zhuravlev, E.;Schick, C. Silk I and Silk II Studied by Fast Scanning Calorimetry. ActaBiomater. 2017, 55, 323−332.(187) Monnier, X.; Maigret, J.-E.; Lourdin, D.; Saiter, A. GlassTransition of Anhydrous Starch by Fast Scanning Calorimetry.Carbohydr. Polym. 2017, 173, 77−83.(188) Baeten, D.; Mathot, V. B. F.; Pijpers, T. F. J.; Verkinderen, O.;Portale, G.; Van Puyvelde, P.; Goderis, B. Simultaneous SynchrotronWAXD and Fast Scanning (Chip) Calorimetry: On the (Isothermal)Crystallization of HDPE and PA11 at High Supercoolings and CoolingRates up to 200 °C s −1. Macromol. Rapid Commun. 2015, 36, 1184−1191.(189) Melnikov, A. P.; Rosenthal, M.; Rodygin, A. I.; Doblas, D.;Anokhin, D. V.; Burghammer, M.; Ivanov, D. A. Re-Exploring theDouble-Melting Behavior of Semirigid-Chain Polymers with an in-SituCombination of Synchrotron Nano-Focus X-Ray Scattering andNanocalorimetry. Eur. Polym. J. 2016, 81, 598−606.(190) van Drongelen, M.; Meijer-Vissers, T.; Cavallo, D.; Portale, G.;Poel, G. V.; Androsch, R. Microfocus Wide-Angle X-Ray Scattering ofPolymers Crystallized in a Fast Scanning Chip Calorimeter.Thermochim. Acta 2013, 563, 33−37.(191) Steyaert, I.; Delplancke, M.-P.; Van Assche, G.; Rahier, H.; DeClerck, K. Fast-Scanning Calorimetry of Electrospun PolyamideNanofibres: Melting Behaviour and Crystal Structure. Polymer 2013,54, 6809−6817.(192) Thomas, D.; Schick, C.; Cebe, P. Novel Method for FastScanning Calorimetry of Electrospun Fibers. Thermochim. Acta 2018,667, 65−72.(193) Novotny, L. Chapter 5 The History of Near-Field Optics. Prog.Opt. 2007, 50, 137−184.(194) Keilmann, F. Scattering-Type near-Field Optical Microscopy. J.Electron Microsc. 2004, 53, 187−192.(195) Ishibashi, T.; Cai, Y. Polarization Properties in Apertureless-Type Scanning Near-Field Optical Microscopy. Nanoscale Res. Lett.2015, 10, 10625.(196) Zhang, W.; Fang, Z.; Zhu, X. Near-Field Raman Spectroscopywith Aperture Tips. Chem. Rev. 2017, 117, 5095−5109.(197) Hecht, B.; Bielefeldt, H.; Inouye, Y.; Pohl, D. W.; Novotny, L.Facts and Artifacts in Near-Field Optical Microscopy. J. Appl. Phys.1997, 81, 2492−2498.(198) Gucciardi, P. G.; Bachelier, G.; Allegrini, M.; Ahn, J.; Hong, M.;Chang, S.; Jhe, W.; Hong, S. C.; Baek, S. H. Artifacts Identification inApertureless Near-Field Optical Microscopy. J. Appl. Phys. 2007, 101,064303.(199) Huckabay, H. A.; Armendariz, K. P.; Newhart, W. H.; Wildgen,S. M.; Dunn, R. C. Near-Field Scanning Optical Microscopy for High-

ACS Nano Review


AG


Resolution Membrane Studies. Nanoimaging; Humana Press: Totowa,NJ, 2013; Vol. 950, pp 373−394.(200) Gucciardi, P. G.; Micheletto, R.; Kawakami, Y.; Allegrini, M.Polarization-Modulation Techniques in Near-Field Optical Micros-copy for Imaging of Polarization Anisotropy in Photonic Nanostruc-tures. In Applied Scanning Probe Methods II; Bhushan, B., Fuchs, H.,Eds.; Springer: Berlin, 2006; pp 321−360.(201) Higgins, D. A.; Vanden Bout, D. A.; Kerimo, J.; Barbara, P. F.Polarization-Modulation near-Field Scanning Optical Microscopy ofMesostructured Materials. J. Phys. Chem. 1996, 100, 13794−13803.(202) Qin, J.; Umeda, N. Near-Field Birefringence Response of IPSLiquid Crystal Thin Film Detected by Bi-SNOM. In Proceedings of theSPIE 7133, Fifth International Symposium on Instrumentation Science andTechnology; Tan, J., Wen, X., Eds.; SPIE, 2008; p 71333D.(203) Camposeo, A.; Greenfeld, I.; Tantussi, F.; Moffa, M.; Fuso, F.;Allegrini, M.; Zussman, E.; Pisignano, D. Conformational Evolution ofElongated Polymer Solutions Tailors the Polarization of Light-Emission from Organic Nanofibers. Macromolecules 2014, 47, 4704−4710.(204)Dai, S.; Fei, Z.;Ma,Q.; Rodin, A. S.;Wagner,M.;McLeod, A. S.;Liu, M. K.; Gannett, W.; Regan, W.; Watanabe, K.; Taniguchi, T.;Thiemens, M.; Dominguez, G.; Castro Neto, A. H.; Zettl, A.; Keilmann,F.; Jarillo-Herrero, P.; Fogler, M. M.; Basov, D. N. Tunable PhononPolaritons in Atomically Thin van DerWaals Crystals of Boron Nitride.Science 2014, 343, 1125−1129.(205) Park, K. D.; Raschke,M. B. Polarization Control with PlasmonicAntenna Tips: A Universal Approach to Optical Nanocrystallographyand Vector-Field Imaging. Nano Lett. 2018, 18, 2912−2917.(206) McLeod, A. S.; Kelly, P.; Goldflam, M. D.; Gainsforth, Z.;Westphal, A. J.; Dominguez, G.; Thiemens, M. H.; Fogler, M. M.;Basov, D. N. Model for Quantitative Tip-Enhanced Spectroscopy andthe Extraction of Nanoscale-Resolved Optical Constants. Phys. Rev. B:Condens. Matter Mater. Phys. 2014, 90, 085136.(207) Hayazawa, N.; Tarun, A.; Inouye, Y.; Kawata, S. Near-FieldEnhanced Raman Spectroscopy Using Side IlluminationOptics. J. Appl.Phys. 2002, 92, 6983−6986.(208) Xue, L.; Li,W.; Hoffmann, G. G.; Goossens, J. G. P.; Loos, J.; DeWith, G. High Resolution Tip Enhanced Raman Mapping on PolymerThin Films. Macromol. Symp. 2011, 305, 73−80.(209) Lee, N.; Hartschuh, R. D.; Mehtani, D.; Kisliuk, A.; Maguire, J.F.; Green, M.; Foster, M. D.; Sokolov, A. P. High Contrast ScanningNano-Raman Spectroscopy of Silicon. J. Raman Spectrosc. 2007, 38,789−796.(210) Cialla, D.; Deckert-Gaudig, T.; Budich, C.; Laue,M.;Moller, R.;Naumann, D.; Deckert, V.; Popp, J. Raman to the Limit: Tip-EnhancedRaman Spectroscopic Investigations of a Single Tobacco Mosaic Virus.J. Raman Spectrosc. 2009, 40, 240−243.(211) Poborchii, V.; Tada, T.; Kanayama, T. Subwavelength-Resolution Raman Microscopy of Si Structures Using Metal-Particle-Topped AFM Probe. Jpn. J. Appl. Phys. 2005, 44, L202.(212) Dazzi, A.; Prater, C. B.; Hu, Q.; Chase, D. B.; Rabolt, J. F.;Marcott, C. AFM-IR : Combining Atomic Force Microscopy andInfrared Spectroscopy for Nanoscale Chemical Characterization. Appl.Spectrosc. 2012, 66, 1365−1384.(213) Dazzi, A.; Prater, C. B. AFM-IR: Technology and Applicationsin Nanoscale Infrared Spectroscopy and Chemical Imaging. Chem. Rev.2017, 117, 5146−5173.(214) Gong, L.; Chase, D. B.; Noda, I.; Marcott, C. A.; Liu, J.; Martin,D. C.; Ni, C.; Rabolt, J. F. Polymorphic Distribution in IndividualElectrospun Poly[(R)-3-Hydroxybutyrate- Co -(R)-3-Hydroxyhexa-noate] (PHBHx) Nanofibers. Macromolecules 2017, 50, 5510.(215) Hinrichs, K.; Shaykhutdinov, T. Polarization-DependentAtomic Force Microscopy−Infrared Spectroscopy (AFM-IR): InfraredNanopolarimetric Analysis of Structure and Anisotropy of Thin Filmsand Surfaces. Appl. Spectrosc. 2018, 72, 817−832.(216) Amarie, S.; Ganz, T.; Keilmann, F. Mid-Infrared near-FieldSpectroscopy. Opt. Express 2009, 17, 21794.

(217) Huth, F.; Schnell, M.; Wittborn, J.; Ocelic, N.; Hillenbrand, R.Infrared-Spectroscopic Nanoimaging with a Thermal Source. Nat.Mater. 2011, 10, 352−356.(218) Huth, F.; Govyadinov, A.; Amarie, S.; Nuansing, W.; Keilmann,F.; Hillenbrand, R. Nano-FTIR Absorption Spectroscopy of MolecularFingerprints at 20 Nm Spatial Resolution. Nano Lett. 2012, 12, 3973−3978.(219) Amenabar, I.; Poly, S.; Nuansing, W.; Hubrich, E. H.;Govyadinov, A. A.; Huth, F.; Krutokhvostov, R.; Zhang, L.; Knez,M.; Heberle, J.; Bittner, A. M.; Hillenbrand, R. Structural Analysis andMapping of Individual Protein Complexes by Infrared Nanospectro-scopy. Nat. Commun. 2013, 4, 2890.(220) Amenabar, I.; Poly, S.; Goikoetxea, M.; Nuansing, W.; Lasch, P.;Hillenbrand, R. Hyperspectral Infrared Nanoimaging of OrganicSamples Based on Fourier Transform Infrared Nanospectroscopy.Nat. Commun. 2017, 8, 14402.(221) Muller, E. A.; Pollard, B.; Bechtel, H. A.; van Blerkom, P.;Raschke, M. B. Infrared Vibrational Nano-Crystallography and-Imaging. Sci. Adv. 2016, 2, e1601006.(222)Wiens, R.; Findlay, C. R.; Baldwin, S. G.; Kreplak, L.; Lee, J. M.;Veres, S. P.; Gough, K. M. High Spatial Resolution (1.1 Μm and 20Nm) FTIR Polarization Contrast Imaging Reveals Pre-RuptureDisorder in Damaged Tendon. Faraday Discuss. 2016, 187, 555−573.(223) Liu, L.; Dzenis, Y. Simulation of Electrospun NanofibreDeposition on Stationary and Moving Substrates. Micro Nano Lett.2011, 6, 408.(224) Papkov, D.; Maleckis, K.; Zou, Y.; Andalib, M.; Goponenko, A.;Dzenis, Y. Nano to Macro: Mechanical Evaluation of MacroscopicallyLong Individual Nanofibers. InMEMS and Nanotechnology; Prorok, B.C., Starman, L., Eds.; Conference Proceedings of the Society forExperimental Mechanics Series; Springer International Publishing,2016; Vol. 5 SE-6, pp 35−43.(225) Lemstra, P. J.; Kirschbaum, R.; Ohta, T.; Yasuda, H. High-Strength/High-Modulus Structures Based on Flexible Macromole-cules: Gel-Spinning and Related Processes. Developments in OrientedPolymers2; Springer: Dordrecht, The Netherlands, 1987; pp 39−77.(226) Lemstra, P. J.; Bastiaansen, C. W. M.; Meijer, H. E. H. Chain-extended Flexible Polymers. Angew. Makromol. Chem. 1986, 145, 343−358.(227) Rastogi, S.; Yao, Y.; Ronca, S.; Bos, J.; Van Der Eem, J.Unprecedented High-Modulus High-Strength Tapes and Films ofUltrahigh Molecular Weight Polyethylene via Solvent-Free Route.Macromolecules 2011, 44, 5558−5568.(228) Mohan, S. D.; Mitchell, G. R.; Davis, F. J. Chain Extension inElectrospun Polystyrene Fibres: A SANS Study. Soft Matter 2011, 7,4397.(229) Smith, S. B.; Bendich, A. J. Electrophoretic Charge Density andPersistence Length of DNA as Measured by Fluorescence Microscopy.Biopolymers 1990, 29, 1167−1173.(230) Ott, A.; Magnasco, M.; Simon, A.; Libchaber, A. Measurementof the Persistence Length of Polymerized Actin Using FluorescenceMicroscopy. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip.Top. 1993, 48, R1642−R1645.(231) Bellan, L. M.; Cross, J. D.; Strychalski, E. A.; Moran-Mirabal, J.;Craighead, H. G. Individually Resolved DNA Molecules Stretched andEmbedded in Electrospun Polymer Nanofibers. Nano Lett. 2006, 6,2526−2530.(232) Buckers, J.; Wildanger, D.; Vicidomini, G.; Kastrup, L.; Hell, S.W. Simultaneous Multi-Lifetime Multi-Color STED Imaging forColocalization Analyses. Opt. Express 2011, 19, 3130.(233) Ries, J.; Udayar, V.; Soragni, A.; Hornemann, S.; Nilsson, K. P.R.; Riek, R.; Hock, C.; Ewers, H.; Aguzzi, A. A.; Rajendran, L.Superresolution Imaging of Amyloid Fibrils with Binding-ActivatedProbes. ACS Chem. Neurosci. 2013, 4, 1057−1061.(234) Pinotsi, D.; Kaminskie, G. S.; Kaminski, C. F. Optical Super-Resolution Imaging of β-Amyloid Aggregation In Vitro and In Vivo:Method and Techniques. Methods in Molecular Biology; Springer:Clifton, NJ, 2016; Vol. 1303, pp 125−141.

ACS Nano Review


AH


(235) Kim, N.; Kim, H. J.; Kim, Y.; Min, K. S.; Kim, S. K. Direct andPrecise Length Measurement of Single, Stretched DNA Fragments byDynamic Molecular Combing and STED Nanoscopy. Anal. Bioanal.Chem. 2016, 408, 6453−6459.(236) Rust, M. J.; Bates, M.; Zhuang, X. Sub-Diffraction-LimitImaging by Stochastic Optical Reconstruction Microscopy (STORM).Nat. Methods 2006, 3, 793−796.

ACS Nano Review


AI


Quantifying Polymer Chain Orientation in Strong and Tough ... · 5/13/2019 · solvent evaporation during electrospinning27 that can act similarly to fast quenching in traditional

Documents