Report

Porous Semiconductor Chalcogenide Aerogels

See allHide authors and affiliations

Science  21 Jan 2005:
Vol. 307, Issue 5708, pp. 397-400
DOI: 10.1126/science.1104226

Abstract

Chalcogenide aerogels based entirely on semiconducting II-VI or IV-VI frameworks have been prepared from a general strategy that involves oxidative aggregation of metal chalcogenide nanoparticle building blocks followed by supercritical solvent removal. The resultant materials are mesoporous, exhibit high surface areas, can be prepared as monoliths, and demonstrate the characteristic quantum-confined optical properties of their nanoparticle components. These materials can be synthesized from a variety of building blocks by chemical or photochemical oxidation, and the properties can be further tuned by heat treatment. Aerogel formation represents a powerful yet facile method for metal chalcogenide nanoparticle assembly and the creation of mesoporous semiconductors.

Porous solids, such as zeolites, ordered mesoporous materials (M41S type), aerogels, and photonic crystals, find applications in such technologies as catalysis, sorption, filtration, thermal and acoustic insulation, and optical switches. Their distinctive properties are a function of the high internal surface area; the size, shape, and degree of interconnectedness of the pores; and the chemical nature of the framework. The pore structure can act as either a thoroughfare or a trap, or can even modulate the optical properties (as in photonic crystals), whereas the chemical interface is where catalysis, sensing, and sorption take place. The vast majority of porous solids are oxide based, thereby limiting the range of chemical functionality and physical properties that can be engineered into these materials. Recently, there has been sustained effort to prepare metal chalcogenide (sulfide, selenide, and telluride) analogs that are semiconducting substitutes for traditional oxide materials, the majority of which are electrically insulating.

There has been considerable success with the creation of microporous (<2 nm pores) zeolitic chalcogenide structures (13), as well as macroporous (>50 nm) materials created by sphere-templating strategies (photonic crystals) (46). In contrast, attempts to prepare similarly ordered mesoporous (2 to 50 nm) materials of metal chalcogenides have met with less success. Efforts have largely emphasized surfactant templating methodologies, originally developed for M41S-type materials, that employ metal chalcogenide building blocks and transition metal ion linkers. The resulting materials demonstrate long-range ordering of the surfactant/chalcogenide framework; however, attempts to remove the surfactants to access the pores invariably result in structure collapse (712).

We have developed a strategy for the production of mesoporous nanostructured metal chalcogenides based on the assembly of discrete nanoparticles to produce aerogel-type frameworks. Aerogels are highly porous architectures that arise when the solvent in a wet polymeric gel is replaced by air while the structural integrity of the framework is maintained (13). They are defined by a nanonetwork of particles, each with a low degree of connectivity, resulting in porosity that is intrinsic to an aerogel (regardless of whether a monolith is produced) and not to a precipitate. Aerogels are typically produced by supercritical drying because other routine techniques for drying gels (e.g., vacuum extraction or heating) result in collapse of the pore structure due to capillary forces and, ultimately, in the formation of dense aggregates (xerogels). Unlike the mesostructured materials reported to date, there is no long-range order in aerogels; however, the range of pore sizes and interconnectedness should be optimal for properties dependent on facile transport of small molecules through the interior space, such as photocatalysis and sensing (14). Such properties have only begun to be exploited in aerogels and are usually achieved through the incorporation of conductive or fluorescent moieties into conventional silica aerogels (15, 16). In contrast, the semiconducting and luminescent nature of the aerogels presented herein is a consequence of the intrinsic properties of the metal chalcogenide aerogel framework.

The methodology for chalcogenide aerogel formation is simple, comprising three steps: (i) Nanoparticle formation/thiolate capping, (ii) gelation through controlled surface-group loss, and (iii) supercritical CO2 drying to maintain the pore architecture (1719). Both room-temperature reversemicellar strategies (CdS, ZnS, PbS, and CdSe) and high-temperature arrested-precipitation techniques (CdSe) can be employed for nanoparticle production. The controlled loss of surface groups to reveal reactive sites for nanoparticle condensation is achieved by chemical oxidation of thiolate capping groups (17) or, in the case of CdSe, by photooxidation of the capping groups (20). This step is pivotal for the preparation of porous gel frameworks, because rapid surface-group loss leads to the formation of dense aggregates and precipitation. The resultant gels are exchanged multiple times with acetone to remove the oxidized disulfide by-product and then are dried at ∼40°C by using supercritical CO2. For comparison, samples were also permitted to air dry on the benchtop, producing xerogels.

The chalcogenide aerogels appear to be morphologically similar to base-catalyzed silica aerogels, as depicted for CdS and CdSe (Fig. 1, A and B). The nanoparticle building blocks that make up the pearl-necklace morphology are visible, as is the presence of mesopores (2 to 50 nm). Unlike traditional silica aerogels, the building blocks appear to be crystalline, as evidenced by the presence of lattice fringes (Fig. 1C), which correspond to the (102) planes of the hexagonal wurtzite (CdS) structure. This is also in contrast to what is observed in the surfactant-templated mesostructured materials where the framework itself was amorphous, despite the presence of long-range order due to the surfactant substructure. Electron diffraction patterns of CdS aerogels confirm the crystallinity, revealing four rings. These could likewise be indexed to the wurtzite structure (fig. S1).

Fig. 1.

Electron microscope images of (A) CdS aerogels and (B) CdSe aerogels, illustrating the colloidal nature of the aerogel aggregates. The presence of mesopores (2 to 50 nm) can be clearly seen. (C) A high-magnification image of a CdS aerogel heated in vacuo at 100°C, illustrating the crystalline nature of the nanoparticle architecture. The lattice fringe separation is 2.5 Å, corresponding to the (102) reflection of hexagonal CdS. (D) A photograph provides a comparison of a wet CdS gel (center) with an unwashed CdS xerogel (left) and a monolithic CdS aerogel (right). In contrast to xerogel formation, minimal volume loss is observed in the conversion of wet gels to aerogels. The scale is in millimeters.

Notably, monoliths can be achieved for CdS and ZnS with bulk densities as low as 0.07 g/cm3 and 0.35 g/cm3, respectively, representing 1.4 to 8.7% of the density of a single crystal (CdS, 4.83 g/cm3;ZnS, 4.04g/cm3). Figure 1D shows a representative CdS aerogel monolith (right) compared with the wet gel (center). Only a small volume loss is observed upon aerogel formation, in contrast to a sample dried on the benchtop (xerogel) (Fig. 1D, left). If the wet gels are washed before benchtop drying, xerogel monoliths cannot be obtained at all, because fragmentation occurs upon drying.

An assessment of the porosity of the chalcogenide aerogels and xerogels is achieved by using N2 adsorption/desorption analysis (Table 1). All samples exhibit a type IV isotherm with H3-type hysteresis loop attributed to an interconnected mesoporous system with a broad pore-size distribution (fig. S2) (19, 21, 22). The small upward hysteresis exhibited by the isotherms suggests a cylindrical pore geometry (23). Poresize distribution plots of the aerogels confirm the presence of a range of pores extending from the meso to the macro regime, with averages of 15 to 45 nm (fig. S2 and Table 1). The xerogels (washed) have a considerably narrower size distribution, with average pore diameters in the lower mesoporous range (CdS is 5 to 7 nm for a xerogel versus 29 to 30 nm for an aerogel).

Table 1.

Surface area and porosity data for metal chalcogenide xerogels and aerogels. All samples were treated with H2O2 to induce gelation, with the exception of CdSe, which was treated with tetranitromethane. Data were acquired on a Micromeritics (Norcross, GA) ASAP 2010 Surface Area Analyzer. Samples were degassed for a minimum of 24 hours at 100°C before analysis. In an effort to extract the full pore volume of the aerogels, long equilibrium intervals were used and granular specimens were evaluated in lieu of monoliths (32). Surface areas were computed with the Brunauer, Emmett, and Teller (BET) multimolecular layer adsorption model, and average pore size and cumulative pore volumes were computed with the Barrett, Joyner, and Halenda (BJH) model. The range of values reflects data from two or more samples prepared and measured under similar conditions.

Metal chalcogenide xerogel/aerogel BET surface area (m2/g) Mean BET surface area (m2/g) BJH adsorption average pore diameter (nm) BJH adsorption cumulative pore volume (cm3/g)
CdS xerogel 38-55 47 5-7 0.05-0.11
CdS aerogel 239-250 245 29-30 1.90-2.02
CdSe aerogel 128-161 143 16-29 0.53-0.98
ZnS aerogel 182-202 192 15-30 0.40-0.86
PbS aerogel 119-141 130 21-45 0.79-0.94

The surface areas achieved for the chalcogenide aerogels range from 120 to 250 m2/g and are comparable to those obtained for many oxide-based aerogels. For example, aerogels of vanadia (24) are reported to have surface areas of 150 to 280 m2/g, whereas those of manganese oxide (25) have surface areas of up to 210 m2/g. Silica sets the bench-mark, with surface areas of up to 1600 m2/g (600 m2/g is a typical value) (13). When compared on a per mole basis, our CdS aerogels with mean surface area of 245 m2/g are equivalent to a silica aerogel of surface area 590 m2/g. In contrast, the washed xerogels exhibit considerably lower surface area (47 m2/g), not unlike values obtained for precipitated CdS nanoparticles (56 m2/g) (26).

Despite these aerogels having a three-dimensional connected chalcogenide network, they nevertheless retain the optical features of the nanoparticle building blocks. As illustrated in Fig. 2, all of the aerogel materials examined exhibit a sharp absorption onset at an energy far greater than the bandgap for bulk solids (Fig. 2, tabular inset), consistent with retention of quantum-confinement effects in the solids (19). This strongly suggests that the nanoparticle chromophore remains effectively isolated, which can be attributed to the fractal connectivity of the network or to the presence of grain interfaces. These chromophores can be ripened by heating, resulting in a systematic red shift of the absorption onset that is nearly linear with temperature (Fig. 3). This also correlates with a growth in the average crystallite size, as probed by x-ray powder diffraction (Fig. 3). Thus, the sharp absorption onset in these materials can effectively be tuned. In addition to the growth of particles, there is a systematic transformation from the cubic crystal structure (as observed for as-prepared aerogels produced from room-temperature reverse-micelle nanoparticles) to the more thermodynamically stable hexagonal modification (dominant phase at 500°C) (fig. S3). Even at conditions where the x-ray data suggest a dominant cubic structure (100°C), nanocrystallites with hexagonal features can clearly be seen with transmission electron microscopy (TEM), and indexing of lattice fringes and electron diffraction patterns are consistent with some degree of cubic-to-hexagonal transformation (Fig. 1 and fig. S1).

Fig. 2.

Diffuse reflectance data for as-prepared aerogels of PbS, CdSe, CdS, and ZnS and comparative values for bulk crystalline samples. Data were acquired on a Shimadzu (Columbia, MD) model UV-3101PC double-beam, double-monochromator spectrophotometer equipped with an integrating sphere, using BaSO4 as a 100% reflectance standard. The bandgaps of the samples were estimated from the low-energy onset in the absorbance data (converted from reflectance).

Fig. 3.

Influence of temperature on crystallite size of the primary particle components in CdS aerogels (open circles, x-ray powder diffraction data) and on absorption onset (filled triangles, diffuse reflectance data).

Photoluminescence studies on as-prepared CdS aerogels reveal only broad, trap-state emission near 600 nm (fig. S4). Upon heating at 100°C, some of the band-edge emission apparent in the nanoparticle precursors is recovered. This is coincident with the beginning of the transformation from cubic to hexagonal and an increase in crystallinity as evidenced by x-ray powder diffraction. It is well documented that the highly crystalline hexagonal nanoparticles produced from high-temperature (∼200°C to 300°C) routes exhibit greater photoluminescence quantum yields than the cubic nanoparticles produced at room temperature with reverse-micellar strategies (27). Thus, we sought to test whether nanoparticles produced from high-temperature routes would give rise to as-prepared aerogels demonstrating strong band-edge photoluminescence.

CdSe nanoparticles were prepared in trioctylphosphine oxide at elevated temperatures and capped with mercaptoundecanoic acid (MUA) to form hexagonal CdSe nanoparticles of ∼5.1 nm in diameter, as estimated from the absorbance onset according to the mass approximation model (20, 28, 29). The CdSe nanoparticles exhibit a sharp band-edge emission near 540 nm, along with a broad shoulder centered near 610 nm, attributable to trap states near the surface (Fig. 4). These nanoparticles transform into opaque orange gels upon standard treatment with tetranitromethane. However, gels were also observed to form in alcohol solutions after standing in ambient light over several days, likely as a result of photooxidation of the MUA capping groups from the surface, representing an alternate method of condensation (20). Aerogels formed from supercritical drying of these gels demonstrate slightly lower surface areas than those prepared from reverse-micellar strategies (30). However, the sharp band-edge emission characteristic of the quantum-confined nanoparticles is retained, with no further processing required (19). The appreciable change in the chromophore environment is evidenced by the shift in the broad trap-state emission peak to near 650 nm (Fig. 4), which suggests very different surface characteristics after condensation and drying than is present in the pristine CdSe nanoparticles in solution.

Fig. 4.

Photoluminescence data for MUA-capped CdSe nanoparticles (open triangles) and an as-prepared aerogel (filled squares), acquired by using an excitation wavelength of 480 nm. Data were collected at 77 K on a SPEX Fluorolog model spectrometer (Jobin Yvon Horiba, Edison, NJ) with 1681 Spex 0.22 m excitation module and 0.34 m emission module. Powdered aerogel samples were placed in evacuated quartz tubes for measurement.

Together, these data suggest that the controlled removal of surface groups under conditions that lead to the formation of lacunar aggregates and eventually gelation is quite general for metal chalcogenide nanoparticles. This is consistent with a previously reported study by Kotov in which CdSe nanoparticle samples stripped of excess thioglycolic acid spontaneously assembled first into chainlike aggregates and subsequently into wires (31). Likewise, Peng found that photooxidation of surface thiolate ligands from CdSe nanoparticles resulted in aggregate formation, and this was reversible upon introduction of more capping agent (20). We find that by adjusting the conditions under which capping groups are removed, robust gels can be produced, and the characteristic pore structure of these colloidal gels can be maintained by supercritical drying. The generality of this method should lend itself to a number of new aerogel materials if the surface chemistry can be appropriately tailored. Because the resulting aerogels retain the photophysical properties of the quantum-confined building blocks, gelation and aerogel formation represents an excellent strategy for the assembly of nanoparticles from solution into solid state devices where exploitation of size-dependent properties is desired.

The aerogels presented here based on PbS, CdSe, CdS, and ZnS cover the optical spectrum from the infrared through the ultraviolet, and the optical features of each material can be effectively “tuned” over a substantial range by adjusting the heating profile employed. The aerogel structure effectively preserves the integrity of the quantum dots by locking them into a network while providing a pore structure through which chemical species can be introduced, either as analytes or as secondary components for composite formation. Current efforts are devoted to preparing these materials in thin-film form and evaluating their potential for photovoltaic and sensing applications.

Supporting Online Material

www.sciencemag.org/cgi/content/full/307/5708/397/DC1

Materials and Methods

Figs. S1 to S4

References and Notes

References and Notes

View Abstract

Navigate This Article