Presentation on theme: "The Challenge of Nanomaterials: Routes to reliable materials? Prof Alun Vaughan October 2011."— Presentation transcript:
The Challenge of Nanomaterials: Routes to reliable materials? Prof Alun Vaughan October 2011
The Challenge of Nanomaterials: Routes to reliable materials? A random walk on the edge Prof Alun Vaughan October 2011
44 Motivation Increased performance Increased reliability Increased power density Increased functionality Reduced power losses Reduced environmental impact
55 The options New polymers Add something else –A filler (micro, meso, nano) –A polymer (immiscible, miscible) –Small molecules Which way to go?
66 Overview Why nano? What is the interphase? How much interphase? What is required for miscibility? How can we modify the interface?
Fillers: Micro, meso, nano?
88 Size matters DC breakdown data from 10% BN in epoxy Strength increases as the particle size is reduced Strength of the unfilled system ~165 kV mm -1 Thomas Andritsch, PhD Thesis, 2010 Nano-structuring the epoxy improves performance
99 Publications Search ISI Web of Knowledge using terms poly* AND nanocomposite* First paper published in 1986 Period of rapid exponential growth Plateau?
10 Projects Improved combinations of properties
11 Filler chemistry – TiO 2 DC breakdown data for TiO 2 in epoxy Strength decreases with nanoparticle inclusion Strength of the unfilled system ~320 kV mm -1 J Keith Nelson and John C Fothergill, Nanotechnology 15 (2004) 586–595 Nano-structuring the epoxy degrades performance
12 Filler chemistry – TiO 2 and Al 2 O 3 Addition of micro-sized filler is bad news Addition of even 0.1% of nanofiller is bad news Epoxy / TiO 2 Epoxy / Al 2 O 3 S.Singha, M.J.Thomas, IEEE Trans DEI 2008, 15, 12
13 Filler chemistry – SiO 2 and BN AC breakdown data for SiO 2 and BN in epoxy Strength of SiO 2 largely independent nanoparticle inclusion Strength increases with BN inclusion What’s the key feature of nanocomposites? Silica benign, meso-scopic BN good, even at “high” loadings
14 The nature of the beast Nanoparticle size/distribution/aspect ratio Nanoparticle chemistry/impurities Nanoparticle structure/crystallography Nanoparticle surface chemistry Interactions with matrix material – stoichiometry Interactions with matrix material – molecular mobility ( ┴ ) Charges/ions/polarisation Matrix morphology Aggregation/percolation Which of these factors are important?
What is the interphase?
16 The multilayer model T.Tanaka, IEEE Trans DEI 2005, 12, 914
17 Hard evidence – NMR Theory T 1 relaxation involves redistributing the populations of the nuclear spin states in order to reach the thermal equilibrium distribution. By definition this is not energy conserving. Moreover, spontaneous emission is negligibly slow at NMR frequencies. Hence truly isolated nuclear spins would show negligible rates of T 1 relaxation. However, a variety of relaxation mechanisms allow nuclear spins to exchange energy with their surroundings, the lattice, allowing the spin populations to equilibrate. The fact that T 1 relaxation involves an interaction with the surroundings is the origin of the alternative description, spin-lattice relaxation.
18 Hard evidence – NMR 1 NMR  considers the utility of NMR as a potential on-line screening tool for characterizing dispersion in nanocomposites. The rationale behind the approach is that paramagnetic Fe 3+ ions present in MMT as impurities will affect the proton longitudinal relaxation time in the polymer, a parameter termed T 1 H. In the case of protons located within about 1 nm of the MMT surface, T 1 H will be reduced directly, while so-called spin-diffusion results in this mechanism propagating into the bulk. Since the measured value of T 1 H will depends upon on the concentration of Fe 3+ ions in the system and their proximity to the polymer, the better the MMT dispersion, the greater the reduction in T 1 H compared with the value determined from the polymer alone.  J.W. Gilman, S. Bourbigot, J.R. Shields, M. Nyden, T. Kashiwagi, R.D. Davis, D.L. Vanderhart, W. Demory, C.A. Wilkie, A.B. Morgan, J. Harris, R.E. Lyon, “High Throughput Methods For Polymer Nanocomposites Research: Extrusion, NMR Characterization And Flammability Property Screening”, J. Mater. Sci. 38 (2003) 4451–4460.
19 Hard evidence – NMR 2 NMR  used NMR spectroscopy to study nanocomposites based upon styrene-butadiene rubber (SBR) and titania. Although 13 C NMR results revealed significant shifts in peak positions, which have been taken to indicate interactions between nanoparticles and polymer chains, spin lattice relaxation experiments suggest that the molecular mobility in both systems is equivalent.  T.M. Arantes, K.V. Leao, M.I.B. Tavares, A.G. Ferreira, E. Longo, E.R. Camargo, “NMR study of styrene- butadiene rubber (SBR) and TiO2 nanocomposites”, Polymer Testing 28 (2009) 490–494
20 Hard evidence – ESR Theory
21 Hard evidence – ESR 1 ESR  studied nanocomposites of poly(methyl acrylate) (PMA) and synthetic fluoromica, in which the PMA had been modified to include a so-called spin label. That is, a stable free radical, commonly nitroxide, which is introduced into a material that does not have an intrinsic paramagnetic response. This work showed that, in exfoliated systems, the mobility of PMA chains is reduced due to the interactions with the nanofiller. The thickness of the rigid interface region was estimated to be in the range 5-15 nm. In intercalated materials similar results were obtained, in that a fraction of constrained chain segments were detected at the clay interface together with another with a higher mobility.  Yohei Miwa, Andrew R. Drews, and Shulamith Schlick, “Detection of the Direct Effect of Clay on Polymer Dynamics: The Case of Spin-Labeled Poly(methyl acrylate)/Clay Nanocomposites Studied by ESR, XRD, and DSC”, Macromolecules 2006, 39,
22 The interphase The “interphase” corresponds to an intermediate region where material properties are representative of neither phase A nor phase B “A frequent situation in nanodielectric systems is one in which the surface or at least a part of the surface of particle A becomes effectively charged and the surrounding phase B responds by establishing a screening counter charge confronting the charge on A.” T.J.Lewis, IEEE Trans DEI 2004, 11, 739
23 How much interphase? “Interface properties become increasingly prominent if phase A is a particle of finite size and surrounded by B with the AB interface between them …the total interface contribution can become very significant as the particle diameter is reduced.” T J Lewis, Interfaces: nanometric dielectrics, J. Phys. D: Appl. Phys. 38 (2005) 202–212 T.J.Lewis, IEEE Trans DEI 2004, 11, 739
How much interphase?
25 More than two phases How does the fraction of interphase i vary with filler loading p ?
26 2-D – the effect of symmetry 43 I II Area of matrix phase
27 Interphase S. Raetzke and J. Kindersberger, Role of Interphase on the Resistance to High- voltage Arcing, on Tracking and Erosion of Silicone/SiO2 Nanocomposites, IEEE Trans. DEI 17, 2010, Three-fold (solid line) and four-fold (dashed line) 20 nm particles
28 Interphase The form of behaviour is independent of symmetry or dimensionality At low filler loading levels, the area fraction of interphase material increases linearly, according to the relationship: where n indicates the dimensionality of the model (here n = 2) and x ai and x ap represent the area fractions of interphase and particles respectively. This is independent of symmetry and corresponds to the regime before overlap of neighbouring interphase regions. At high filler loading levels, x ai varies with x ap according to: This is independent of symmetry, dimensionality, or the value chosen for the interphase thickness and corresponds to the regime where all of the area not occupied by the particles themselves corresponds to interphase material. Three-fold (solid line) and four-fold (dashed line) 20 nm particles
29 Lattice Consider adding the (R + 1)th nanoparticle. The (R + 1)th nanoparticle cannot occupy a cell that is already occupied by a nanoparticle. The (R + 1)th nanoparticle can occupy any one of the unoccupied (N – R) cells that were, previously, either interphase or matrix. The probability of it occupying an interphase cell can therefore be written P I (R), where: and I(R) represent the number of interphase cells present prior to the introduction of the (R + 1)th nanoparticle.
30 Lattice The inclusion of the (R + 1)th nanoparticle will convert neighbouring, previously matrix cells, into interphase cells. The coordination number, K n, specifies the number of interphase cells per nanoparticle in the limit R 0. At higher fill fractions, there will be a finite probability of each of these K n interphase cells coinciding with a cell that was not previously of matrix character. The effective number of additional interphase-type cells induced by the addition of the (R + 1)th nanoparticle can then be written: where M(R) represent the number of matrix cells present prior to the introduction of the (R + 1)th nanoparticle. Thus, the effective number of interphase cells after addition of the (R + 1)th nanoparticle, I(R+1) can be written:
33 Quench a polymer from a temperature T 1 to another temperature T 2, where T 1 > T g > T 2 The initial glassy state will depend upon both T 1 and T 2 The Gibbs-DiMarzio theory for a polymer AAAAAAAAAAAAAAAA : At some temperature, the distribution of free volume in the system is such that molecular motion is no longer possible within the time scale of the measurement. Free volume is envisaged as being dynamically created and destroyed locally through the cooperative motion of chain segments. Depends upon local bond conformations and “broken” A-A inter/intra molecular bonds. The nature of the glass transition
34 Where a polymer is close to a second medium, we need to consider both polymer – polymer (A-A) and polymer – medium (A-B) interactions This can affect molecular configurations and mobility and, consequently, the measured glass transition Consider polyethylene glycol (low molar mass PEO) confined within porous silica Confined polyethylene glycol “These results clearly indicate that confined PG exhibits longer relaxation times compared to the bulk dynamics. This finite size effect increases as the temperature is lowered and thus implies a considerable retardation in molecular mobility for confined polyethylene glycol near T g.” J.Schuller, Y.B.Melnichenko, B.Yu, R.Richert, E.W.Ficher, 1994 Dielectric studies of the glass transition in porous media Phys. Rev. Lett –7
35 Consider toluene in porous silica Two processes can affect the measured T g A decrease in T g can occur with decreasing pore size as a result of the material vitrifying under conditions of constant volume (isochoric conditions); modelling indicates that this is an intrinsic size effect related to the influence of a negative hydrostatic pressure on glass formation Interactions with the pore walls tends to reduce inhibit molecular interactions and, hence, increase T g Contributing processes D.Morineau, Y.D.Xia, C.Alba-Simionesco, 2002 Finite-size and surface effects on the glass transition of liquid toluene confined in cylindrical mesopores J. Chem. Phys –72.
36 Consider solutions of polystyrene (PS) in ortho-terphenyl (o-TP) “Interestingly, the DSC thermograms for the o-TP or o-TP/PS solutions confined in the pore show what appear to be two glass transitions. One is at a higher temperature than the bulk state T g and the other is at a lower temperature.” Multiple T g s J.Y.Park, G.B.McKenna, 1999 Size and confinement effects on the glass transition behavior of polystyrene/o-terphenyl polymer solutions Phys. Rev. B –76
37 So … Ideas based upon interphases are very reasonable in nanocomposites and include ideas of molecular confinement The interphase is believed to constitute a substantial fraction of the matrix in nanocomposites Evidence from spectroscopy of molecular interactions T g is intrinsically linked to thermodynamic interactions and molecular confinement Porous systems have been extensively studied Strong T g effects have been reported and analysed in detail (theory)
38 So … Ideas based upon interphases are very reasonable in nanocomposites and include ideas of molecular confinement The interphase is believed to constitute a substantial fraction of the the matrix in nanocomposites Evidence from spectroscopy of molecular interactions T g is intrinsically linked to thermodynamic interactions and molecular confinement Porous systems have been extensively studied Strong T g effects have been reported and analysed in detail (theory) … how about for nanocomposites?
39 T g in epoxy/silca systems T g is strongly dependent upon resin stoichiometry in both unfilled and filled (5%) systems Tg is suppressed in nanocomposites of optimum stoichiometry The value of c p varies systematically with stoichiometry/filling All glass transitions are singular Width of T g is constant within experimental error The complete system is being affected
40 Plot of the real part of the permittivity against volume fraction of nanoparticles for a random 3-D simulation of an array of nanoparticles (diameter 20 nm). The interphase thickness t i = 20 nm (K 3 = 26) and interphase permittivity ε i ’ = 2.4 throughout; results for nanoparticle permittivity values of ε p ’ = 6 and ε p ’ = 10 are shown. The solid and long dashed lines correspond to the upper and lower Wiener bounds respectively and the intermediate Lichtenecker-Rother equation is indicated by the dash/dot/dot line. r ´ - varying particle permittivity
41 Plot of the real part of the permittivity against volume fraction of nanoparticles for a random 3-D simulation of an array of nanoparticles (diameter 20 nm, ε p ’ = 8) and an interphase thickness of 20 nm (K 3 = 26). Results are shown for interphase permittivities ε i ’ = 2 and ε i ’ = 2.8. The solid and long dashed lines correspond to the upper and lower Wiener bounds respectively and the intermediate Lichtenecker-Rother equation is indicated by the dash/dot/dot line. Varying interphase permittivity
42 Plot of the real part of the permittivity against volume fraction of nanoparticles for a random 3-D simulation of an array of nanoparticles (diameter 20 nm, ε p = 8). Results for an interphase permittivity ε i ’ = 2.4 and interphase thicknesses of 10 nm (K 3 = 7) and 40 nm (K 3 = 63) are shown. The solid and long dashed lines correspond to the upper and lower Wiener bounds respectively and the intermediate Lichtenecker-Rother equation is indicated by the dash/dot/dot line. Varying interphase thickness
43 MgO The complete system is being affected Effective particle permittivity? Thomas Andritsch, PhD Thesis, 2010
Thermodynamics of miscibility
46 A random model of a three phase system The extent to which different systems mix depends on the Gibbs free energy of the system, G In thermodynamic terms, two components will mix intimately provided this results in a reduction in the total free energy of the system: whereG 12 = Gibbs free energy of mixture G 1 = Gibbs free energy of component A G 2 = Gibbs free energy of component B If ΔG m is the Gibbs free energy of mixing and ΔG m < 0, mixing will be favoured thermodynamically: Miscibility
47 In general, the entropy term can be written: where: Entropy and enthalpy In general, the enthalpy term can be written:
48 Theory EntropyEnthalpy
50 Sol-gel chemistry The initial reaction is hydrolysis: Si(OR) 4 + H 2 O → HO-Si(OR) 3 + R-OH Depending on the amount of water and catalyst present, hydrolysis may proceed to completion, so that all of the OR groups are replaced by OH groups, as follows: Si(OR) H 2 O → Si(OH) R-OH SILANOL PRODUCTION Hydrolyzed molecules undergo condensation reactions to form siloxane bonds: (OR) 3 –Si-OH + HO–Si-(OR) 3 → [(OR) 3 Si–O–Si(OR) 3 ] + H-O-H or (OR) 3 –Si-OR + HO–Si-(OR) 3 → [(OR) 3 Si–O–Si(OR) 3 ] + R-OH POLYMERISATION Polymerisation therefore results in the formation of a 1, 2, or 3- dimensional network of siloxane [Si–O–Si] bonds accompanied by the production of H-O-H and R-O-H species. TEOS tetraethyl orthosilicate tetraethoxysilane
51 Surface functionalisation (Dow) A two percent silane solution can be prepared in the alcohol of choice and applied to the sample Particles, e.g., pigments and fillers, can be silylated by stirring them in a solution for two to three minutes and then decanting the solution. The particles can then be rinsed with alcohol. Cure of the silane layer is for 5-10 min at 110 o C or for 24 hr at ambient conditions. A 95% ethanol-5% water solution is adjusted to pH with acetic acid; silane is added with stirring to yield a 2-10% final concentration Silanetriols are most stable at pH 3-6, but condense rapidly at pH For less soluble silanes, 0.1% of a nonionic surfactant could be added and an emulsion rather than a solution is prepared. Stability of aqueous silane solutions varies from hours for the alkyl silanes to weeks for the aminosilanes. Poor solubility parameters limit the use of long chain alkyl and aromatic silanes by this method
52 Variants Alkyl substitutedEpoxy compatible
53 Another recipe Angew. Chem. Int. Ed. 2003, 42, 4326 –4331 The solvents were all pre-dried by using standard methods To dry calcined MCM-41 in THF (15 mL), dichlorodiphenylsilane (0.48 g, 0.19 mmol) was added and stirred at room temperature for 1 hour (Ph 2 SiCl 2 ) reacts with external Si-OH moieties and ensures all proceeding silane species reacts at the internal surface of MCM-41). Elemental analysis found (%): C4.70, H0.87, Br Si CP MAS NMR indicated significant loss of Si-OH moieties, indicating grafting had occurred
54 RAFT methods Colloidal silica particle suspension (48 mL of 30 wt % SiO 2 in MIBK, D 20 nm), active silane (1.7 mmol, 0.61 g), and dried THF (6 mL) were added to flask. The reaction mixture was heated at 85 °C under N 2 protection overnight and then cooled to room temperature. The reaction mixture was then precipitated into a large amount of hexane(500 mL). The particles were recovered by centrifugation at 3000 rpm for 15 min. The particles were repeatedly re-dissolved in 20 mL of acetone and reprecipitated in 200 mL of hexane.
55 Characterisation 1 “Characteristic absorption bands were clearly visible at cm -1 due to the carbonyl group and at and 688 cm -1 due to the phenyl ring.” Chunzhao Li and Brian C. Benicewicz, Macromolecules 2005, 38,
56 Characterisation 2 2-stage process – TEOS then trimethyethoxylsilane (TMES) Clear chemical effects relating to changes in surface chemistry can be seen in both the FTIR and 29 Si NMR spectra Feng-Hsi Huang, Chao-Ching Chang, Tai-Yueh Oyang, Ching-Chung Chen, Liao-Ping Cheng, J Nanopart Res (2011) 13:3885–3897
57 But what does the surface need to look like?
59 MY Conclusions I think the interfaces are key in both the science and technology of nanodielectrics I think we have ideas but, at present, we don’t have enough understanding of what an interface/interphase is – there are techniques out there that have been used successfully I think we don’t have enough understanding of how to characterise interfaces I fear that much of our current attempts equate to “fighting the thermodynamics” and, in ameliorating this, are other demons introduced? I think that much more systematic study is necessary I do not have the tools to do what is necessary – I need to collaborate
60 The new hall of the Tony Davies HV Lab. Thank you