Linear and Nonlinear Optical Properties of PalladiumNanoparticle Reinforced Fluoropolymer Composites

James Spicer1,2, Kevin See1, Yasushi Katsumi1, Dajie Zhang2, John Brupbacher3, andTerrence Vargo4

1 The Johns Hopkins University, Department of Materials Science and Engineering,Room 103 Maryland Hall, Baltimore, MD 21218, spicer@jhu.edu

2 The Johns Hopkins University, Advanced Technology Laboratory, Baltimore, MD 212183 The Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723

4 Integument Technologies, Inc., Tonawanda, NY 14150, tvargo@integument.com

ABSTRACT

The linear optical properties of a functionally graded,palladium nanoparticle reinforced fluorinated ethylenepolymer matrix nanocomposite (PMNC) were investigatedin this work. Synthesized through repeated infusions of apalladium organometallic precursor gas into an insulating,fully fluorinated polymer matrix film, the compositeconsists of discrete, palladium nanoparticles distributedthroughout the polymer matrix. Under controlledprocessing conditions, preferential near-surface nucleationof nanoparticles can be achieved and, as a result of increasesin nanoparticle density near the surface, percolation of thenucleated particles was observed. The presence of this nearsurface percolated layer significantly alters the opticalproperties of this materials system. The reflectivity of thenanocomposites was investigated in depth and compared tobehavior predicted using the Tourquato-Kreibig-Fresnel(TKF) theory. The nonlinear optical properties of thesematerials have also been investigated using optical limitingexperiments. Under relatively low fluence conditions, thepalladium nanoparticles are excited to effectively changethe nanocomposite into an efficient reflector as a result ofplasma formation at the material surface. With plasmaformation, optical transmission through the nanocompositedecreases significantly. Using a 8 ns laser pulse at 1064nm, the measured transmission characteristics show thatpulse fluences of nearly 1.3 x 104 mJ/cm2 reduced thetransmission coefficient to less than 10% of the linear value.

1. INTRODUCTION

Currently it is possible to synthesize a wide array ofnanocomposite materials systems with controllable physicalproperties [1-3]. Polymer Matrix Nanocomposites (PMNCs)synthesized by uniformly incorporating metal nanoparticleswithin an insulating polymer matrix are of interest owing tothe optical properties they exhibit that are distinct from thatof the bulk [4,5]. PMNCs can be synthesized throughestablished thin film processing technologies such as

physical vapor deposition and chemical infusion techniques[6,7].

A chemical infusion synthesis method was used for thisstudy since it presents a wide range of organic precursor andpolymer combinations that allow fabrication of a variety ofmetal-polymer nanocomposite materials systems withdesirable properties. The infusion synthesis method is atype of chemical vapor deposition process carried out in thefree volume of the polymer matrix material [8]. Thisprocess can be repeated to control the volume percentagecontent of metal nanoparticles in the polymer leading tosignificant changes in the optical properties of thesematerials [9].

2. THEORY: OPTICAL PROPERTIES

Torquato-Kreibig-Fresnel theory addresses thescattering characteristics of nanoparticles embedded in amatrix material that includes multipolar interactionsbetween the distributed particles in the matrix [10]. Byknowing the bulk properties of the nanoparticles and thematrix, as well as the volume fraction of the nanoparticles,the TKF model represents the effective dielectric constant,εe, as follows:

εe(ω) = εm1 + 2φB− 2(1− φ)ξ2B2

1 − 2φB− 2(1− φ)ξ2B2

(1)

whereφ is the particle volume fraction, 2ξ is the structureparameter, mε is the dielectric constant of the polymermatrix, and B is calculated using the equation

B =εp − εmε p + 2εm

(2)

where pε is the frequency and size dependent dielectricfunction of the nanoparticles. The structure parameter

789NSTI-Nanotech 2006, www.nsti.org, ISBN 0-9767985-6-5 Vol. 1, 2006

accounts for the multipolar interactions between theparticles that occur at high volume fractions [11]. As thevolume fraction becomes less than 0.1, the structureparameter approaches zero and the above equation becomesthe well-known Maxwell-Garnett dielectric function used tomodel composite systems with small embedded particles[12-14]:

εe(ω) = εm 1 +1 +3φ (εp −ε m) (ε p + 2εm )1− φ (ε p − εm) (ε p + 2εm )

. (3)

To apply this model to our system appropriately, baselinestructural and physical measurements were performed toestablish the size, shape, and distribution of Pdnanoparticles in the polymer matrix.

3. STRUCTURAL AND OPTICALCHARACTERIZATION

For th i s s tudy , e i ther semicrys ta l l inehexafluoropropylene-co-tetrafluoroethylene (FEP) orethylene-co-tetrafluoroethylene (ETFE) was used as thematrix polymer for nanoparticle synthesis. This synthesiswas performed through vacuum processing techniques toheterogeneously nucleate and grow nanoparticles within thepolymer [8]. Palladuim nanoparticles were producedthrough the repetitive infusion and decomposition of theorganometallic precursor Pd Acetylacetonate [Pd(C5H7O2)2]in the polymer matrix.

3.1 Structural Characterization Results

High resolution transmission electron microscope(TEM) micrographs showed that the Pd nanoparticles werespherical, ranged between 7-15 nm in diameter with thediameter increasing with successive infusion cycles. Figure1 shows the particle shape and distribution for a sample thathas been infused 15 times resulting in approximately 2 v/oPd in the composite. Under various processing conditions,particle size and particle density can be higher near thesurface due to the higher flux of precursor that passesthrough this region during infusion and decomposition.This is shown in Fig. 2 where the surface and near surfaceregions of one of the composites contains a densely packedpopulation of Pd nanoparticles.

Nanocomposite systems processed through hightemperature decomposition favor near surface nanoparticlenucleation and growth. Through repeated infusions underthese processing conditions, it is possible to produce aninterconnecting network of particles that give the particles apercolative structure. In the literature, percolation forrandomly packed, impenetrable spheres has been observedfor volume fractions as low as 16 % [15,16].

Figure 1: TEM micrograph of Pd-FEP nanocompositewith approximately 2 v/o Pd content. The Pd particle

diameter is approximately 10 nm.

Figure 2: TEM Micrograph of Pd Particles near the FEPsurface showing high population density. The

decomposition temperature used for the infused precursorgas was 207 0C for this sample.

For our samples, percolation was only observed in twoheavily infused samples within a sample set produced athigher decomposition temperatures. A second set ofsamples was processed at a slightly lower temperature thatdid not favor near surface nucleation and growth. This setconsisted of four samples: two lightly infused (3,5 times)and two heavily infused (15,17 times). Volume percentagesrange between 0.27% and 3.3% and vary linearly with thenumber of infusion cycles carried out during materialproduction for fewer than 18 cycles. Although the twosample sets were synthesized through slightly differentprocessing conditions and were visually distinguishable, theweight fraction gains with successive infusion cycles weresimilar for both.

3.2 Linear Optical Property Measurements

Having performed baseline structure analysis of thesenanocomposites, additional studies were carried out todetermine the optical properties of these materials. Sincemacroscopic sample curvature deformed the shape of thereflected beam, an integrating sphere, reflectivity apparatuswas used to measure the reflectivity of the PMNCs. Aschematic of the apparatus is shown in Fig. 3.

790 NSTI-Nanotech 2006, www.nsti.org, ISBN 0-9767985-6-5 Vol. 1, 2006

Figure 3: Integrating sphere reflectivity setup.

The reflected power measured using the integratingsphere contains contributions from both the diffuse and thespecular components [17], however, since the diffusecomponent was small compared to the specular reflection,the total power measured from the integrating sphere is agood estimate of the specular reflectivity of thenanocomposite system [18].

Although the PMNCs had low average volumepercentages determined using simple mass gainmeasurements, baseline characterization showed thatsamples with percolated layers exhibited much higher localvolume percentages. Figure 4 shows the measured andcalculated optical reflectivities of PMNCs as a function ofthe Pd content and indicates that the introduction ofnanoparticles into the polymer matrix caused the reflectivityto decrease compared with that of the uninfused FEP – aresult not predicted by the TKF model. Beyond a certainvolume fraction, the reflectivities for both sets of samplesincreased with increasing volume fraction of nanoparticles.

Figure 4: Measured and calculated optical reflectivitiesof Pd reinforced PMNCs as a function of the Pd content.

Samples processed to yield percolated surfaces exhibitedmuch higher reflectivities than calculated using TKF owingto the high local concentration of particles in the near-surface region of the composite. Materials processed atrelatively low temperatures with similar volume percentagesdisplayed optical reflectivities that followed trends similarto that shown in Fig. 4.

3.3 Nonlinear Optical Property Measurements

To complement the linear optical propertymeasurements, nonlinear optical properties were alsoinvestigated using simple optical limiting measurements.These were performed using the apparatus shownschematically in Fig. 5 where the transmittance of aNd:YAG laser pulse (1064 nm, 8ns pulse duration) wasmeasured as a function of the laser pulse fluence.

Figure 5: Experimental apparatus used for performingoptical limiting experiments on Pd-FEP nanocomposites.

These measurements were completed only on those samplescontaining the lowest content of Pd (1 or 3 infusions) sincethese transmitted sufficient laser pulse energy to bemeasured using standard laser pulse energy meters.

The results of these measurements are shown in Fig. 6where the absolute transmissivity is shown as a function ofthe incident laser pulse fluence. At low fluences, nonlineareffects are small and the linear transmittance is measured.

Figure 6: Results for the optical limiting performance ofPd-ETFE nanocomposite materials as a function of laser

pulse fluence.

These results show that the material infused once (1x)transmits approximately ten times as much optical energy asthe material produced using three infusion cycles (3x) – thisis in rough agreement with the Beer-Lambert Law. As thelaser pulse fluence increases, the transmitted pulse energygenerally decreases. Above a certain incident pulse fluence,plasma formation at the PMNC surface is initiated and isresponsible for significant reductions in the transmittedpulse energy owing to the optical properties of plasmas.This decrease appears to be exponential as is shown in Fig.

791NSTI-Nanotech 2006, www.nsti.org, ISBN 0-9767985-6-5 Vol. 1, 2006

6 by the solid lines fit to the data. Below this plasmaformation threshold, the material behavior is quitecomplicated and is currently being explored using moresophisticated measurement techniques.

4. DISCUSSION

For samples with a homogeneous distribution ofnanoparticles, our system satisfies the conditions assumedfor TKF, yet the introduction of nanoparticles in thepolymer matrix causes significant changes not predicted bythis model. Qualitatively, as metal nanoparticles with highdielectric constant are added to a polymer matrix, theoverall effective dielectric constant should rise. Instead, forall of the lightly infused samples, we observed ~ 45%reduction in the reflectivity from the uninfused FEP. Thissuggests that as small amounts of nanoparticles are added tothe polymer matrix, the overall dielectric constant of thenanocomposite decreases. The opposite result is expectedbased on most composite models including TKF andMaxwell-Garnett. Since the volume fraction of thenanoparticles was low, this effect might be attributed to thechanges in the dielectric constant of the matrix.

This decrease in the dielectric constant in the matrixcould result from chemical interactions that exist betweenthe palladium particles and the polymer matrix that lead tolong range variations throughout the materials system.Clearly, percolation phenomena as well as strong chemicalinteractions between the matrix and the polymer make itdifficult to simply model these polymer nanocompositematerial systems using effective medium theories.

5. CONCLUSION

The linear and nonlinear optical properties of palladium-polymer matrix nanocomposite systems were investigatedusing optical reflectivity measurements and optical limitingmeasurements. These polymer-metal nanocomposites werefabricated using a chemical infusion synthesis techniquecapable of producing stable, crystalline, spherical palladiumnanoparticles in an insulating polymer matrix. By choosingspecific processing conditions, it was possible to producesamples with near surface percolation of nanoparticles.Samples with percolated layers exhibited much higherreflectivites than non-percolated samples of similar volumefraction. The most heavily loaded percolated sample had areflectivity of 14.3%. It was also shown that chemicalinteraction between the polymer matrix and the palladiumnanoparticles can decrease the effective dielectric constantof the polymer matrix for the lightly infused samples.Chemical interactions between the fluorinated polymermatrix and the palladium nanoparticles, as well as nearsurface percolation phenomena make it difficult for current

effective medium theories to simply model the behavior ofthese materials systems.

REFERENCES

[1] R. Vaia, The AMPTIAC Newsletter, Vol.6, #1,[2] R.A. Vaia, J.W. Lee, W.Click, G. Price, C.S. Wang,

“Polymeric based conducting nanocomposites:controlling the distribution of the conducting phase,”Mat. Res. Soc. Symp. Vol. 519 1998.

[3] L.L. Beecroft, C.K. Ober, “Nanocomposite materialsfor optical applications,” Chem. Mater. 1997.

[4] P.M. Ajayan, L.S. Schadler, P.V. Braun,Nanocomposite Science Technology, Wiley Verlag,2003.

[5] A.B. Mayer, “Formation of noble metal nanoparticleswithin a polymer matrix: nanoparticle features andoverall morphologies,” Materials Science andEngineering,1998.

[6] K.Akamatsu, S.Deki, “Nanoscale Metal ParticlesDispersed in Polymer Matrix,” NanostructuredMaterials, Vol.8, No.8 1997.

[7] S.A. Zavyalov, A.N. Pivkina, J. Schoonman,“Formation and characterization of metal-polymernanostructured composites,” Solid State Ionics, 1472002.

[8] T.S. Koloski, T.G. Vargo, Us Patent #6608129 2003.[9] K.C. See, J.B. Spicer, J. Brupbahcer, D.Zhang, T.G.

Vargo, “Modeling Interband Transitions in SilverNanoparticle-Fluoropolymer Composites,” J. Phys.Chem. B 2005, 109, 2693-2698.

[10] I.Fabman, O.Levi, S. Efrima, “Optical response ofconcentrated colloids of coinage metals in the near-ultravioley, visible and infrared regions,” J. Chem.Phys., 96(9),1992.

[11] S. Torquato, “Effective electrical conductivity of two-phase disordered composite media,” J. Appl. Phys., 581985.

[12] U. Kreibig, M. Vollmer, Optical properties of metalclusters, Springer, Berlin 1995.

[13] K. L. Kelly, E. Coronado, L.L.Zhao, G. Schatz, “Theoptical properties of metal nanoparticles: The influenceof size, shape, and dielectric environment,” J. Phys.Chem. B 2003.

[14] C. Bohren, D.R. Huffman, Absorption and Scattering ofLight By Small Particles, John Wiley and Sons, 1983.

[15] H. Scher, R. Zallen, Critical density in percolationpreocesses, J.Chem.Phys.Vol.53, 1970.

[16] M.J. Powell, “Site percolation in randomly packedspheres,” Physical Review B, 20(10) 1979.

[17] D.Lide, Handbook of Chemistry and Physics, CRCPress, 2002.

[18] I. Lindseth, A.Bardal, R. Spooren, “Reflectancemeasurements of aluminum surfaces using integratingspheres,” Optics and Lasers in Engineering. 32, 2000.

792 NSTI-Nanotech 2006, www.nsti.org, ISBN 0-9767985-6-5 Vol. 1, 2006