Optical second-harmonic generation (SHG)is used as a noninvasive probe of the interfaces of Si nanocrystals (SiNCs)embedded in an SiO₂ matrix. We verify experimentally that the second-harmonic polarization P⁽²⁾ has a quadrupolar form proportional to (E ·∇) E as proposed in recent models based on a locally noncentrosymmetric dipolar polarization averaged over the spherical NC interface. A two-beam sum-frequency geometry is found to enhance this polarization dramatically compared to a single-beam SHG geometry, yielding strong signals useful for scanning, spectroscopy and real time monitoring. Using this two beam geometry, we have produced non-invasive two dimensional SHG maps with few-micron resolution of 1-micron-thick layers of Si-NCs (3 and 5 nm average diameter)produced by ion implanting Si into SiO₂. Samples were scanned over a 5mm x 5mm area with two non-collinear,orthogonally polarized,amplified Ti:S laser pulses (80 fs,810nm,100 μJ,1 kHz repetition rate) while detecting the generated SH signal in transmission. The SHG signal is sensitive to chemical modification of the Si/SiO₂ interface and to local gradients in nanoparticle density.

The growth of self-assembled alkylsiloxane monolayers on uniform and patterned silicon substrates has been investigated at room temperature using atomic force microscopy (AFM), contact angle measurements and quartz crystal microbalance (QCM) gravimetry. Immersion of oxidized silicon substrates in a millimolar solution of octadecyltrichlorosilane (OTS) results in the formation of ordered octadecylsiloxane islands with a height close to 2.5 nm. In the area between these islands an additional - presumably disordered - adsorbate layer with a height of about 0.6 nm can be identified. The overall uptake-curves show subtle but significant deviations from the generally assumed first-order Langmuir adsorption kinetics. A nearly perfect fit, however, can be achieved on the basis of a simple model considering the adsorption of initially disordered species which subsequently transform into ordered islands. In this model, the disordered species are believed to occupy a larger surface area per entity and hence prevent adsorption of further molecules before rearrangement takes place. In contrast to oxidized silicon substrates, H-terminated areas on silicon substrates appear to remain uncoated after immersion into an OTS solution. Considering these results, a laser direct writing technique has been used in order to create arbitrarily patterned silicon substrates which expose H-terminated as well as oxidized areas. Starting with a uniformly H-terminated silicon surface this technique allows for writing oxide lines with a lateral resolution arround 500 nm suitable for the selective coating with an alkylsiloxane monolayer.

The ability to create an ensemble of mono-dispersed nanostructures is an important step towards the realization nanotechnology. The discovery of surface-magic-clusters (SMC), i. e. clusters exhibiting enhanced stability at certain sizes on a particular surface, has opened up the possibility of exploiting SMC formation for the growth of identical nanostructures on surfaces. Recently, it has been demonstrated that, under a well controlled deposition condition, several group III elements such as gallium can induce almost exclusive formation of SMC on the 7x7-reconstructed Si(111) surface, leading to the complete filling of the 7x7 half unitcells and the creation of unprecedented two-dimensional lattices of SMC. Based on scanning tunneling microscopy imaging studies, structure models for the SMC are proposed. Ab inito calculations of the model clusters are conducted and the results are compared with experiments.

The utilization of self-assembly mechanisms for the controlled deposition of nanoparticles at surfaces and interfaces recently has gained increasing popularity. A variety of methods, ranging from the use of purely physical phenomena to the application of chemical functionalization of the particles and/or the surface, have been proposed for the fabrication of two-dimensional mesoscopic structures based on nanoparticle assemblies. Potential applications are found in chemical and biological sensing, photonics, mesoscopic optics, and mesoscale electronics. Here, we present our recent results on the controlled deposition of monodisperse polystyrene (PS) latex particles onto chemically modified surfaces by use of small organic molecules added in proper amounts to the suspensions. In particular, the role of entropic forces in screening chemical selectivity for surface adsorption is elucidated. Thereby, a route for the controlled deposition of the PS particles onto carboxyl-functionalized surface areas utilizing carbodiimide chemistry is developed.

In this paper we discuss selected equilibrium and dynamic properties of adsorption layers of soluble surfactants. The surface state has been investigated by nonlinear optical techniques based on second order χ⁽²⁾ effects which exhibit a high surface specificity and suppress bulk contributions. The surface tension isotherm σ(c )of the homologous series of n -alkyldimethylphosphine (n =8 −12) can be described by Frumki ’s equation of state which yields the surface interaction parameter, surface coverage and the corresponding area per molecule A . The comparison of the surface tension σ at a given area per molecule A reveals a strong alternation within the homologous series. Odd C_2n±1 layers show a lower surface tension than the adjacent even members C_2n of the homologous series. This effect is also present at low surface coverage (A =1.4nm²)and cannot be attributed to a differences in the chain-packing within a crystalline state. Infrared-Visible Sum-Frequency Generation Spectroscopy (SFGS)has been used to monitor the orientation and chain order within the aliphatic tail. SFGS spectra have been recorded for different chain lengths and at different areas per molecule. The analysis of the spectra yields an order parameter G which is proportional to the number of gauche defects within the aliphatic tail. The odd-even effect in the surface tension turned out to be accompanied by an odd-even effect in the order parameter G. The data suggest that an ordered structure has a bigger impact on the surface tension than an unordered structure. The odd-even effect is also observed in the orientation of the terminating methyl group as retrieved by polarization dependent SFGS measurements. The data shed some light in the relation between molecular and macroscopic properties. Furthermore surface dilatational viscoelastic properties of a fluorinated amphiphile have been measured by a novel version of the oscillating bubble. The oscillating bubble method generates a non-equilibrium state by a harmonic compression and expansion of the surface layer formed at the tip of a capillary. The surface state is monitored by Surface Second Harmonic Generation (SHG).This technique is highly surface specific and discriminates between monolayer and subsurface coverage. Our set-up allows to measure the monolayer coverage under dynamic conditions and to relate this to surface dilatational viscosity and elasticity. For a purely elastic surface layer the prediction of the Lucassen van den Temple model (LvdT)are fulfilled.

The study of nanometer-thick molecular thin films deposited on a solid surface, due to recent technology applications, has become an important subject. Effective tools for unraveling the intrinsic structure within the molecular films and their growth mechanism, however, are still in the searching. Consequently, little is known about the structure and the most important factors controlling deposition of thin molecular films. This paper summarizes the demonstration showing that the nonlinear optical phenomenon- second harmonic generation- because of its symmetry properties can be used effectively to characterize the structure within the molecular films. Experiments show that disorder-order phase transitions, glass transitions, crystallization kinetics and nucleation processes, and interfacial molecular structure within the thin molecular films can be characterized. The nonlinear optical studies have revealed the mechanisms and established the most important criteria for the deposition and growth of ultrathin molecular films.

Holographically-formed Polymer Dispersed Liquid Crystals (H-PDLCs) are electro-optic devices containing alternating planar layers of polymer and liquid crystal droplets. Here, we investigate the effects of patterned polymer layers on the liquid crystal alignment. Three types of polymers were investigated: urethane resin, Poly Tetra Fluoro Ethylene (PTFE) and Poly Methyl Methacrylate (PMMA). Nanopatterns of the form of square grids were formed on the polymer using the MTS Nano-Indenter in the scratch mode. Liquid crystal deposition on these structures showed unique alignment effects at the polymer interface. The polymers were examined under the Atomic Force Microscope and alignment of the liquid crystals was observed by Polarizing Optical Microscopy.

The transient photoconductivity of dye-sensitized nanocrystalline colloidal TiO₂ has been measured time-resolved THz spectroscopy (TRTS), a non-contact electrical probe with sub-picosecond temporal resolution. The photoconductivity deviates strongly from Drude behavior and is explained by disorder-induced carrier localization and/or backscattering of the photogenerated carriers. In addition, the carriers are found to thermally equilibrate with the lattice in roughly 300 femtoseconds. The size-dependent photoconductivity of CdSe nanoparticles ranging from 2.54 nm up to >25 nm has also been measured using TRTS. The measured change in the frequency-dependent optical density and change in phase of the transmitted THz pulse fall into three distinct groupings as a function of size and can be classified for diameters smaller than the Bohr exciton radius, diameters greater than the Bohr exciton radius but smaller than the bulk mean free path, and diameters greater than the bulk mean free path. The underlying cause of the grouping is a size-dependent mobility (or carrier scattering rate).

Ag(I) doped ZnSe nanoparticles were synthesized using molecular cluster precursors. In the emission spectrum at 390 excitation, three emission bands, centered at 432 nm, 517 nm, and 484 nm, respectively were observed. The 432 nm and 517 nm bands can be assigned to ZnSe band-edge emission and donor-accepter emission from the vacancies and trap states in the ZnSe lattice to the Ag(I) dopant, respectively. Similarly the 484 nm band could be the result of ZnSe trap state or vacancies to Ag(I) acceptor emission or simply ZnSe trap state emission. X-ray Absorption Fine Structure (XAFS) data were collected at the Ag K-edge of the Ag(I) doped ZnSe nanoparticles. From these data it was concluded that the Ag(I) dopant occupied a variety of different environments in the ZnSe lattice.

In this paper the dynamical response of cylindrical nanorods to ultrafast laser-induced heating is examined. Theoretical analysis predicts that both extensional and breathing vibrational modes of the rods should be excited by laser-induced heating. Analytical formulas for the frequencies of these modes are derived assuming that the length of the rods is much greater than their radii. Because the frequency of the fundamental extensional mode is much lower than that of the breathing mode, the extensional mode will dominate the response for a real experiment, i.e., for a finite-time heating/expansion process. The results of this model are compared to data from transient absorption experiments performed on gold nanorods with average aspect ratios (length / width) between 2.1 and 5.5, and widths on the order of 10-20 nm. The transient absorption traces show pronounced modulations with a period between 45 and 70 ps, which are only observed when the probe laser is tuned to the longitudinal plasmon band of the sample. The measured periods are in good agreement with the expected period for the extensional modes of the rods. The actual value of the measured period depends on the specific sample and probe laser wavelength. This occurs because the samples are polydisperse, and different length rods absorb in different regions of the spectrum. For rods with widths greater than 20 nm, the breathing mode can also be observed and, again, the measured periods are in good agreement with the theoretical calculations. The breathing mode is not observed for the thinner rods (~10 nm width) because in this case the period is comparable to the timescale for lattice heating in the experiment.

Heterodimeric electon-donor/electron-acceptor charge-transfer complexes chemisorbed onto Au(111) by attachment of the electron-donor to the surface have been characterized by scanning tunneling microscopy and Kelvin probe experiments. Conductance measurements exhibit nearly Ohmic I(V) responses at low bias. The electrical properties of the charge-transfer complex are vastly different than those of the electron-donor alone which exhibits insulating behavior at low bias. In an extension of this work, strategies are being developed for attachment of charge-transfer complexes to semiconducting or insulating surfaces. Fabrication of nanoscale molecular electronic devices is being investigated by attaching one component of a charge-transfer complex to a silicon surface by chemically directed self-assembly. The single component-functionalized surface is then used as a substrate on which the second component of the charge-transfer complex is deposited by the atomic force microscopy method, dip-pen nanolithography (DPN). Derivatives of hexamethylbenze (electron-donor) with terminal olefins attached to crystalline silicon surfaces via hydrosilylation form monolayer-functionalized silicon surfaces that are expected to have insulating properties. Well-defined features can be “drawn” onto the donor-functionalized surfaces by DPN using tetracyanoethylene (electron-acceptor) as the "ink." The resulting charge-transfer complex nanostructures have conducting properties suitable for device function and are flanked by an insulating monolayer, thus creating "wires" made from charge-transfer complexes.

How an electron crosses a metal-molecule interface has been a longstanding question in many disciplines. The recent surge of interests in molecular electronics has renewed the need for quantitative answers to this question. In molecule-based conventional electronic devices, such as organic light emitting diodes, the metal-molecule interface determines the charge injection efficiency. The importance of the interface only increases as device dimension shrinks to the scale of a single or a small group of molecules, i.e., molecular electronics. This paper takes an experimentalist's view and discusses recent progress in understanding electron transport at metal-molecule interfaces using two-photon photoemission (2PPE) spectroscopy, which provides quantitative information on: the alignment of unoccupied molecular orbitals to the metal Fermi level, the localized nature of interfacial dipoles, the strength of interfacial electronic coupling, and the dynamics of the transient electronic state at the interface.

Hot electron injection from the aromatic chromophore perylene into TiO₂ was measured with transient absorption signals for different rigid anchor-cum-spacer groups revealing 15 fs as the shortest and 4 ps as the longest injection time. The energetic position of the donor orbital of the chromophore with respect to the conduction band edge was determined at about 0.8 eV employing ultraviolet photoelectron spectroscopy (UPS)and simple absorption spectroscopy.It is not clear in the case of rutile or anatase TiO₂ whether unoccupied surface states are involved in the electron injection process as acceptor states. Since the surface reconstruction of TiO₂ is difficult to control electron scattering between a well-defined surface state and isoenergetic unoccupied bulk states was studied with InP(100). Electron scattering was time-resolved employing two-color two-photon-photoemission (2PPE). Scattering from isoenergetic bulk states to the empty C₁ surface state was found to occur with a 35 fs time constant, and the reverse process showed a time constant in the range of 200 fs. The latter was controlled by energy relaxation in bulk states, i.e. via the emission of longitudinal optical phonons in InP. In general, the injection of a hot electron from a molecular donor into electronic states of a semiconductor as to be distinguished from consecutive electron scattering processes between surface states and bulk states. Distinguishing between the different processes may become difficult, however,if the electronic interaction becomes large for a small chromophore directly attached to the semiconductor.

The electron transfer (ET) from organic dye molecules to semiconductor-colloidal systems is characterized by a special energetic situation with a charge transfer reaction from a system of discrete donor levels to a continuum of acceptor states. If these systems show a strong electronic coupling they are amongst the fastest known ET systems with transfer times of less than 10 fs. In the first part a detailed discussion of the direct observation of an ET reaction with a time constant of about 6 fs will be given, with an accompanying argumentation concerning possible artifacts or other interfering signal contributions. In a second part we will try to give a simple picture for the scenario of such superfast ET reactions and one main focus will be the discussion of electronic dephasing and its consequences for the ET reaction. The actual ET process can be understood as a kind of dispersion process of the initially located electron into the colloid representing a real motion of charge density from the alizarin to the colloid.

The photoinduced electron transfer (ET) from a molecular electron donor to the TiO₂ semiconductor acceptor triggering Gratzel solar cells and other photochemical applications is investigated. The reported simulations reproduce the experimentally observed ET time-scale, establish the reaction mechanism, and provide a detailed picture of the ET process. The electronic structure of the chromophore-semiconductor system is simulated by density functional theory (DFT). Ab initio molecular dynamics (MD), including non-adiabatic (NA)transitions between electronic states, NAMD, is used to follow the ET reaction in real-time and at the molecular level. The simulation indicates that thermally driven adiabatic ET s dominant at room temperature. Vibrational motions of the chromophores induce oscillations of the photoexcited state energy that drives the photoexcited state in and out of the TiO₂ conduction band. Two distinct types of ET events are observed depending on the initial conditions. At low initial energies the photoexcited state is well localized on the chromophore, and an activation is required for ET, with comparable contributions from both the adiabatic and NA mechanisms. At high initial energies the photoexcited state is already substantially delocalized into the TiO₂ substrate. The remaining fraction of the ET process occurs rapidly and by the adiabatic mechanism.

Electron transfer (ET) dynamics between molecular adsorbates and semiconductor nanoparticles has been a subject of intense recent interest because of relevance to many applications of nanomaterials, such as dye-sensitized solar cells, molecular electronics and sensors. However, it is still unclear how the charge transfer rate depends on the properties of molecules and semiconductors. In this paper we examine electron injection from Ru and Re polypyridyl complexes to metal oxide (TiO₂, SnO₂ and ZnO) nanocrystalline thin films. Adsorbates with different energetics and electronic coupling are compared to identify molecular properties that influence ET dynamics. Different semiconductor nanomaterials are compared to understand the dependence on conduction band composition and energetics. ET dynamics were found to be biphasic consisting of ultrafast (<100fs) and slower components, with varying partitioning between them and rates of slow components. These kinetics can be well described by a two-state injection model, which includes injection from both unthermalized and thermalized excited states and competition between electron injection and intramolecular relaxation from the unthermalized state. The dependence of ET rates on various molecular and semiconductor properties is also discussed.

Two-photon photoemission of thiolate/Ag(111), nitrile/Ag(111), and alcohol/Ag(111) interfaces elucidates electron solvation and localization in two dimensions. For low coverages of thiolates on Ag(111), the occupied (HOMO) and unoccupied (LUMO) electronic states of the sulfer-silver bond are localized due to the lattice gas structure of the adsorbate. As the coverage saturates and the adsorbate-adsorbate nearest neighbor distance decreases, the HOMO and LUMO delocalize across many adsorbate molecules. Alcohol- and nitrile-covered Ag(111) surfaces solvate excess image potential state (IPS) electrons. In the case of alcohol-covered surfaces, this solvation is due to a shift in the local workfunction of the surface. For two-monolayer coverages of nitriles/Ag(111), localization accompanies solvation of the IPS. The size of the localized electron can be estimated by Fourier transformation of the wavefunction from momentum- to position-space. The IPS electron localizes to 15 ± 4 angstroms full-width at half maximum in the plane of the surface, i.e., to a single lattice site.

Technical developments and applications in infrared near-field microscopy for chemical imaging are discussed. A method to amplify the weak scattered signal in an apertureless microscope is demonstrated.

Quasi one-dimensional nanostructures are unique probes of cavity quantum electrodynamics because they are capable of exhibiting photonic and/or electronic confinement in two dimensions. The near-cylindrical geometry and sharp end facets of zinc oxide (ZnO) nanowires enable the realization of active nanoscale optical cavities that exhibit UV/blue photoluminescence (PL) waveguiding and lasing action at room temperature under appropriate optical pumping conditions. Study of individual nanostructures is crucial for isolating geometry-dependent effects, and here it is achieved through both near- and far-field microscopies. The polarization of the emitted PL or lasing from individual nanostructures characterizes the coupling of the spontaneous emission to cavity modes, depending both on the wavelength of the emitted light and the nature of the emitting species (i.e., excitons and intrinsic defects in various charge states). In addition, the spectral evolution of the lasing/PL as a function of the pump fluence indicates both exciton and electron-hole plasma dynamics. Variations of size, geometry, and material on the prototypical cylindrical ZnO nanowire lead to further observation of unique photonic and/or carrier confinement effects in novel nanostructures.

Recently, nanoparticles have become the platform for many sensing schemes. In particular, the utilization of the optical response of nanoparticles to changes in their nanoenvironment has served as a signal transduction mechanism for these sensing events. For example, silver nanoparticle arrays synthesized using nanosphere lithography have served as an ultrasensitive detection platform for small molecules, proteins, and antibodies with the detection limit of 60,000 and less than 25 molecules/nanoparticle for hexadecanethiol and antibodies, respectively. While this approach is low cost and highly portable, one limitation of the array platform is that the signal arises from approximately 1x10⁶ nanoparticles. A method to improve the overall number of molecules detected would be to decrease the number of nanoparticles probed. Recently, single nanoparticle sensing has been accomplished using dark-field microscopy. A 40 nm shift in the localized surface plasmon resonance induced from less than 60,000 small-molecule adsorbates has been monitored from a single Ag nanoparticle. Additionally, streptavidin sensing has also been demonstrated using a single Ag nanoparticle. Detection platforms based on nanoparticle arrays and single nanoparticles will be discussed and compared.

We present in this article some studies of the chemical reactivity of free metal clusters (~8-50 atoms) investigated at single-collision-like conditions in a molecular beam experiment. A beam of clusters is generated with a pulsed laser vaporization source and after expansion into vacuum the cluster beam passes collision cells, in which the clusters can make one or a few collisions with reactive gas molecules. Pure clusters and reaction products are detected with laser ionization and mass spectrometry. A strong size dependence in the reaction probability of N₂ with tungsten clusters is observed. When the temperature of the cluster source is lowered from room temperature to 80 K the reactivity increases strongly and N2 adsorbs in a weakly bound molecular state, whereas only a strongly bound dissociative state is stable at room temperature. The reactivity of platinum clusters with O₂ is much less size dependent and the reaction probability is high on all investigated sizes. If the Pt_nO_m products pass a second cell containing H₂(D₂) the number of adsorbed oxygen atoms decreases with increasing H₂ pressure. This is explained by formation of water molecules in a catalytic reaction on the surface of the Pt clusters.

The study of the adsorption and growth of metals on ceramics is a rapidly growing area, as these interactions are key to understanding many materials and processes used in modern technology. In particular, oxide-supported catalysts have been extensively studied, due to their widespread industrial applications. Knowledge of the role played by the underlying metal oxide in the reactivity of the metal catalyst can give insights into the design of more effective catalysts. Here, we use density functional theory (DFT) to investigate the adsorption of CO onto 1) bulk Pt, 2) Pt thin layers supported on an alpha-alumina surface and 3) Pt nanoparticles on alpha-alumina. Our results show strong binding for CO molecules on the surfaces of both the thin Pt layers and the 3-atom nanoparticles supported on alumina substrates. This enhanced binding can possibly lead to more reactive catalysts. Further calculations on reaction products are needed to determine the effectiveness of these new systems.

Scanning tunneling microscopy is utilized to investigate the structural changes of AgO chains on clean and carbidic-carbon containing Ag(110) surfaces under UV photoirradiation and CO exposure. Although AgO chains are arranged with the (2x1)structure on both of the surfaces, AgO chains are bundled to make the (2x1) bands on the C-containing surface, whereas they make much larger domains on the entire surface of clean Ag(110). The photo-induced elimination of O in AgO chains ccurs only on the C-containing surface. Kinetics of oxygen elimination by CO exposure are very different between the two surfaces. Oxygen coverage decreases steadily on the C-containing surface with CO exposure, whereas the reaction is accelerated in the lower O coverage range where AgO chains with (nx1)(n≥4) configurations show significant structural fluctuation. Comparison between the two surfaces and simulations based on the Ising model indicate that the acceleration of the reaction originates from the dynamical formation of active O adatoms by fluctuation of AgO chains.

Langmuir monolayer at the air/water interface is considered ideal model for studying 2-dimensional phase behaviors. In this work, the molecular detail on phase transition and orientational order in Langmuir monolayer is studied with simultaneous measurements using surface Second Harmonic Generation (SHG) and surface pressure. With the general methodology for SHG orientational analysis developed recently in our group, we studied the 4'-n-octyl-4-cyanobiphenyl (8CB) monolayer at the air/water interface. We were able to determine both the orientational angle and angular distribution, and also to determine the orientational angular changes in the liquid phase of the 8CB monolayer. These new data pointed to a novel phenomenon which can be attribute to the domain interaction driven orientation distribution narrowing (DIDODIN) mechanism during the orientational phase transition, which implies the phase transition is second-order. The same phenomenon is also observed for a quite different molecular system, Parc18. It implies this phenomenon can be general for Langmuir monolayers. Our findings demonstrated that quantitative orientational analysis is capable of determining the molecular interactions at interfaces in surprising details, which are responsible for the behaviors of the phase transition and orientational order in the Langmuir monolayer and molecular films.

Polymer/TaS₂ layered nanocomposites have been synthesized by using the exfoliation-adsorption technique. Single crystals of layered transition-metal dichalcogenides of 1T-TaS₂, 2H-TaS₂, and 4Hb-TaS₂ were used as host materials. Poly(ethylene oxide) (PEO) and poly(ethylenimine) (PEI) were adopted as guest intercalants. As the exfoliation-adsorption method may require complicated procedures, optimum conditions to synthesize nanocomposites were estimated for each set of TaS₂ polytypes and polymers. They have been characterized by powder X-ray diffraction (XRD) and electrical dc resistivity measurements. XRD patterns showed that all samples of the polymer/TaS₂ layered nanocomposites contain organic polymer between all individual TaS₂ sheets. Although 1T-, 2H-, and 4Hb-TaS₂ polytypes are well known to show quite different temperature dependences of resistivity, the resistivities of all nanocomposites show similar temperature dependences.

It is suggested by this investigation, that light emission from Porous Silicon is due to the recombination of excitons generated in different locations. The majority of excitons generate from the nano-size crystalline silicone. However, excitons generated from amorphous silicon/SiO2 interface are also contributing to the whole excitons population in the material. Further, it is suggested that all excitons are confined and recombine in amorphous region.

Lyotripic liquid crystals form highly regular porous matrices with aqueous channels on the nanometer length scale. We have used the cubic phases formed with water by either an amphiphilc block-copolymer (Pluronic F127) or by a lipid (monoolein) for electrophoretic separation of DNA and other biomolecules. Our goal is to use the well-defined pores and the amphiphilic environment to obtain new separation motifs compared to conventional matrices, and to exploit the well-known phase diagrams of these two systems to optimise applications. The Pluronic crystal consists of close-packed micelles and its main advantage is that the cubic phase melts below 10°C, and we show that the separated DNA can be recovered in a biologically active state in preparative applications. Our mechanistic studies revealed that double-stranded DNA undergoes a highly non-conventional (non-reptative) mode of migration, with the helix axis perpendicular to the field direction because the DNA migrates in the grain boundaries of the polycrystalline samples. In contrast to the Pluronic case, the monoolein cubic crystal is bicontinuous, and a main advantage is that it is in equilibrium with a water-rich phase. We exploited this phase-behaviour in the useful sub-marine mode of analytical electrophoresis. The migration of oligonucleotides in the monoolein is strongly retarded compared to free solution and conventional gels, to an extent which is consistent with that migration indeed occurs through the nm-pores. We demonstrate separation of oligonucleotides based on size, and on different types of secondary structure of the same oligonucleotide size, such as the double-stranded, single-stranded and hairpin forms.