3478
Langmuir 2007, 23, 3478-3484
Automated Preparation Method for Colloidal Crystal Arrays of Monodisperse and Binary Colloid Mixtures by Contact Printing with a Pintool Plotter Klaus Burkert,*,† Thomas Neumann,† Jianjun Wang,‡ Ulrich Jonas,‡ Wolfgang Knoll,‡ and Holger Ottleben§ Graffinity Pharmaceuticals GmbH, Im Neuenheimer Feld 518, 69120 Heidelberg, Germany, and Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany ReceiVed October 25, 2006. In Final Form: December 8, 2006 Photonic crystals and photonic band gap materials with periodic variation of the dielectric constant in the submicrometer range exhibit unique optical properties such as opalescence, optical stop bands, and photonic band gaps. As such, they represent attractive materials for the active elements in sensor arrays. Colloidal crystals, which are 3D gratings leading to Bragg diffraction, are one potential precursor of such optical materials. They have gained particular interest in many technological areas as a result of their specific properties and ease of fabrication. Although basic techniques for the preparation of regular patterns of colloidal crystals on structured substrates by self-assembly of mesoscopic particles are known, the efficient fabrication of colloidal crystal arrays by simple contact printing has not yet been reported. In this article, we present a spotting technique used to produce a microarray comprising up to 9600 single addressable sensor fields of colloidal crystal structures with dimensions down to 100 µm on a microfabricated substrate in different formats. Both monodisperse colloidal crystals and binary colloidal crystal systems were prepared by contact printing of polystyrene particles in aqueous suspension. The array morphology was characterized by optical light microscopy and scanning electron microscopy, which revealed regularly ordered crystalline structures for both systems. In the case of binary crystals, the influence of the concentration ratio of the large and small particles in the printing suspension on the obtained crystal structure was investigated. The optical properties of the colloidal crystal arrays were characterized by reflection spectroscopy. To examine the stop bands of the colloidal crystal arrays in a high-throughput fashion, an optical setup based on a CCD camera was realized that allowed the simultaneous readout of all of the reflection spectra of several thousand sensor fields per array in parallel. In agreement with Bragg’s relation, the investigated arrays exhibited strong opalescence and stop bands in the expected wavelength range, confirming the successful formation of highly ordered colloidal crystals. Furthermore, a narrow distribution of wavelength-dependent stop bands across the sensor array was achieved, demonstrating the capability of producing highly reproducible crystal spots by the contact printing method with a pintool plotter.
Introduction The need for high-throughput identification of drug candidates led to the emergence of diverse new sensor technologies suitable for the parallel detection of protein-ligand interactions. For this purpose, array technologies are widely accepted in the pharma industry, and novel array platforms are under constant development. Strong efforts were directed to the evaluation of photonic materials as the sensitive matrix in optical sensor arrangements, but the fabrication of these materials is still lacking in parallelization.1-4 For such sensor applications, it would be of great interest to combine the optical response of photonic materials to changes in its refractive index with the possibility of detecting thousands of probes simultaneously. However, the reproducible arraying of photonic materials was found to be a very difficult objective, taken into account that the miniaturization of the sensor is essential, in particular, to reduce the consumption of analytes * Corresponding author. E-mail:
[email protected]. Fax: +49 6221 6510-111. Tel: +49 6221 6510-139. † Graffinity Pharmaceuticals GmbH. ‡ Max Planck Institute for Polymer Research. § Present address: Boston Consulting, Dirckensstrasse 41, 10178 Berlin, Germany. (1) Alexeev, V. L.; Sharma, A. C.; Goponenko, A. V.; Das, S.; Lednev, I. K.; Wilcox, C. S.; Finegold, D. N.; Asher, S. A. Anal. Chem. 2003, 75, 2316-2323. (2) Asher, S. A.; Sharma, A. C.; Gopenenko; A. V.; Ward, M. M. Anal. Chem. 2003, 75, 167-1683. (3) Asher, S. A. Alexeev, V. L.; Goponenko, A. V.; Sharma, A. C.; Lednev, I. K.; Wilcox, C. S.; Finegold, D. N. J. Am. Chem. Soc. 2003, 125, 3322-3329. (4) Reese, C. E.; Asher, S. A. Anal. Chem. 2003, 75, 3915-3918.
in affinity screening experiments of rare protein targets. A particularly interesting class of such optical materials represents colloidal crystals, which are 3D Bragg gratings and show opalescence with optical stop bands. Various patterning methods are known that can facilitate the precise deposition of colloidal crystals on defined sensor fields on the basis of electrostatic forces, differences in wettability between sensor fields and their surroundings, and the assembly of colloidal crystals under spatial confinement using microstructured substrates or templates.5 Notably, neither of those approaches allows efficient addressing of individual sensor fields as is possible by an ink-jet arrayer6 or a contact printing process using pintools, which is used for the fabrication of various colloidal biochip arrays.7 In this article, we describe the application of the contact printing (spotting) technique with a pintool plotter to produce a microarray sensor setup comprising up to 9600 singly addressable sensor fields of colloidal crystal structures on a microfabricated substrate. Here, spotting is referred to as the transfer of a small volume of liquid from a reservoir onto a surface by means of stainless steel pins facilitated by adhesion forces between the liquid and the pins during the uptake of liquid from the reservoir and between (5) del Campo, A.; Duwez, A.-S.; Fustin, C.-A.; Jonas, U.; Dekker, M. Encyclopedia of Nanoscience and Nanotechnology; Schwarz, J. A., Contescu, C., Putyera, K., Eds.; Marcel Dekker: New York, 2004; pp 725-738. (6) Park, J.; Moon, J.; Shin, H.; Wang, D.; Park, M. J. Colloid Interface Sci. 2006, 298, 713-719. (7) Borchers, K.; Weber, A.; Brunner, H.; Tovar, G. Anal. Bioanal. Chem. 2005, 383, 738-746.
10.1021/la063122z CCC: $37.00 © 2007 American Chemical Society Published on Web 02/02/2007
Preparation Method for Colloidal Crystal Arrays
Langmuir, Vol. 23, No. 6, 2007 3479
Table 1. Array Formats and Pins Used
format 1 format 2 format 3
array dimension size/mm2
number of fields
spot diameter d/µm
spacing ∆/µm
dPin/um
number of pins in pinhead
7687 (71.65 × 107.27) 472 (26.69 × 17.70) 480 (26.88 × 17.88)
9216 1536 9600
400 280 110
750/1125 562 225
350 100 100
384 96 96
the liquid and the sensor surface upon contact of the wetted pins with the substrate. As a result, liquid droplets settle on defined sensor positions, where they are allowed to dry under controlled humidity conditions. In the case of the transfer of monodisperse polystyrole beads dispersed in a specially developed spotting solution, highly reproducible volumes of the particle suspension were transferred. During drying under controlled conditions, the particles in the droplets form periodic structures by spontaneously ordering in a self-assembly process, thus leading to highly ordered, crystalline aggregates at the initial droplet positions. When the diameter of the used particles was chosen to be in the range of the wavelength of visible light, the spotted crystal structures exhibited strong opalescence upon visual inspection. This optical phenomenon is based on the interaction of light with the periodical arrangement of the refractive index contrast between the particles nsphere and the cavities nmedium in the crystalline assemblies. According to Bragg’s relation (eq 1) for colloidal systems, the stop band wavelength λ at maximum reflection strongly depends on the angle of incidence θ of the illuminating light, the distance dhkl between the crystal planes, the layer period m, and the effective refractive index neff. The latter can be calculated from the filling factor f, which in turn describes the ratio between particles to medium in the cavities for a given colloid crystal.8
λ)
2dhkl n2 - sin2 θ m x eff
n2eff ) fn2sphere + (1 - f)n2medium
(1)
In the case of the matching Bragg condition, total reflection of the incident light occurs at distinct wavelengths, and the optical properties of the colloidal crystal material can therefore be characterized by both reflection and transmission spectroscopy. To examine in parallel all stop bands of a plotted colloidal crystal array, we devised the following optical setup. In reflection mode, the crystal array was illuminated by monochromatic light while the intensity of the reflected light was recorded by a highly sensitive CCD camera under defined angles of incidence and reflection (Figure 3 for the schematic setup). Recorded CCD camera pictures were subsequently subjected to automated grayscale analysis to obtain reflection spectra of all individual fields per array in a highly parallel fashion. This imaging approach required optically well-separated areas to allow precise recognition of a single field by the in-house-developed analysis software. Therefore, an array was designed in which the individual fields were defined by a titanium microstructure, enhancing the contrast between the glass fields (no titanium) and their surrounding (protecting titanium layer). The developed methods allow the automated parallel fabrication as well as the automated parallel readout of the colloidal crystal arrays. Experimental Section Polystyrol Particle Suspensions. Suspensions of monodisperse polystyrene (PS) particles in water were synthesized by emulsion polymerization. The styrene (St) monomer (distilled under reduced (8) Kopnov, F.; Lirtsman, V.; Davidov, D. Synth. Met. 2003, 137, 993-995.
pressure before use), the initiators ammonium peroxodisulfate (APS) and potassium peroxodisulfate (KPS), acrylic acid (AA), sodium chloride, and sodium dodecyl sulfate (SDS) were purchased from Aldrich, and deionized water was obtained from a Milli-Q system (Millipore GmbH, Eschborn, Germany). The reactor, a stirred 250 mL flask with a nitrogen inlet, was immersed in a thermostatted water/glycol bath, and agitation was controlled using a stirrer fitted with a tachometer. The reactor was charged with deionized water (with salt and emulgator, when appropriate) and purged with nitrogen for 20 min, and then the initiator and monomer mixture were added over 5 min. The reaction was run for 24 h at 75 °C under stirring. After polymerization, the colloidal particles were purified from large agglomerates by filtration through a standard paper filter. Six circles of centrifugation and redispersion were followed to remove lowmolecular-weight impurities. The average particle size and distribution were measured by dynamic light scattering with a Zetasizer 3000HS (Malvern Instruments Ltd., Malvern, U.K.) and confirmed by SEM. For the 420 nm particles, we used 300 g of H2O, 15.4 g of St, 0.139 g of APS (in 10 mL of H2O), 0.15 g of AA, 0.118 g of sodium chloride, a stirring speed of 300 rpm; the polydispersity index (PDI) was 0.011. For the 95 nm particles, we used 300 g of H2O, 16.7 g of St, 0.36 g of KPS (in 10 mL of H2O), 0.15 g of AA, 0.47 SDS, and a stirring speed of 540 rpm; the PDI was 0.07. Microarray Substrates. The massive parallel readout of the crystal array required a highly precise and reproducible process for microarray fabrication as described in the following text. For the spotting experiments, three different array formats (Table 1) were investigated to increase the array density. The highest density could be obtained with fields of diameter d ) 110 um and a spacing between the fields of ∆ ) 225 µm, which led to an array with 9600 single fields on a standard microscopy glass slide (26 × 76 mm2). The structured substrates were fabricated by photolithography and high-vacuum evaporation. After intense ultrasonic rinsing with ultrapure water and detergent solution (deconex, Bohrer Chemie, Zuchwil, Switzerland), glass slides (D263, Irlbacher, Scho¨nsee, Germany) were covered with a UV-light-sensitive resist (ARP5350, Allresist, Berlin, Germany). An array pattern was produced by illuminating the slide with a UV-light source (Hg) through a highprecision mask (Mask Aligner MA8, Karl Suss, Munich, Germany), followed by removal of the solubilized resist film. A 2000 Å titanium layer was deposited between the resist structures by means of electronbeam evaporation under high vacuum (Vacuum Coating System Classic 590, Pfeiffer Vacuum Technology, Asslar, Germany) and finally structured by a lift-off procedure of the remaining resist regions. In addition to the clear-cut sensor fields used for spotting the colloidal crystals, a set of fiducial fields were included for precise camera-controlled alignment of the pintool spotting head. Chemical Surface Modification. Chemical modification of substrate surfaces is known as an effective way to facilitate the preparation of laterally patterned colloidal crystals9 and was considered here to enhance the quality of pintool-spotted arrays. A significant difference in wetting behavior between the sensor fields (hydrophilic and wetting) and the surrounding titanium matrix (hydrophobized and dewetting) helps to confine dispersion droplets onto the sensor fields. It was shown that hexadecylphosphonic acid (HDPA) selectively adsorbed from solution onto the oxidized titanium surface and enhanced its hydrophobicity, forcing the liquid droplets (9) Fustin, C. A.; Glasser, G.; Spiess, H. W.; Jonas, U. AdV. Mater. 2003, 15, 1025.
3480 Langmuir, Vol. 23, No. 6, 2007
Burkert et al.
Figure 1. (a) High-precision portal robot used for spotting colloid suspensions onto prestructured microarray slides. To control the humidity during the spotting and drying process, the system is encapsulated in a housing equipped with an ultrasonic air humidifier. (b) Pinhead comprising 96 steel pins with a diameter of d ) 100 µm. The pinhead is automatically adjusted with reference to fiducial fields on the substrate by means of a telecentric microscopy CCD camera imaging system. to settle solely on the glass sensor fields without wetting the surrounding titanium. Spotting of Colloidal Crystal Arrays. Transfer of the particle suspension onto the predefined array fields was achieved by a contact printing process wherein individual pins, composed of a spotting head and mounted onto a portal robot (SPI, Oppenheim, Germany), were inked by dipping into suspension reservoirs (i.e., microwell plates) and positioned on the adjusted array substrate in parallel. The spotting suspension was homogenized by ultrasonic treatment before use. However, during spotting (∼1 h), shaking or mixing of the suspensions was not necessary because sedimentation effects were negligible. The placed volume of the suspension depends on the steel pin geometry so that each of the 96 pins with a diameter of dpin ) 100 µm (formats 2 and 3) transported picoliter volumes of suspension per spotting cycle. To obtain reproducible transfer, the suspension properties had to be adjusted in terms of the weight concentration of particles and its viscosity. Highly concentrated suspensions were difficult to print because of the strong capillary forces between particles. It was found that a solution of 0.5% (w/v) particles in water in the presence of 10% ethylene glycol resulted in a reproducible size of colloidal crystal spheres across the array after drying of the sensor chips. During the drying period, the aqueous suspension medium was allowed to evaporate slowly, and the individual particles were arranged in densely packed order. By increasing the relative humidity (70%) during spotting and drying, the evaporation kinetics could be controlled and reduced to a degree that was found optimal for the assembly of highly ordered crystals. For this purpose, the spotting robot was encased in a housing equipped with an ultrasonic air humidifier (Figure 1a). Several parameters influencing the crystal growth on the array (e.g., particle concentration, particle size distribution, and nature of spotting medium) were optimized to obtain homogeneous and reproducible crystal arrays without undesired effects such as the formation of amorphous material or insufficient suspension loading onto the sensor fields, which could interfere with the self-assembly of the particles. In particular for the spotting of binary colloidal crystals, as discussed further below, phase separation between differently sized particles could be an issue. Taking into account the miniaturization of the sensor, extreme requirements for the precision of the three-axis portal robot system and the positioning of the pintool had to be met. Therefore, a spotting system was constructed that allowed us to adjust the pinhead automatically with reference to the relative position of the fiducials on the substrate by means of a visual system containing a telecentric microscope CCD camera (Figure 1b). Furthermore, the linear stages in use were selected for a positioning accuracy below 10 µm. Binary Colloidal Crystals. Several approaches for producing laterally patterned structures of monodisperse colloidal crystals have been described.5,10-12 However, none of these techniques allowed individual sensor fields to be addressed, as is possible by the contact
printing method described here. Moreover, to our knowledge, the preparation of extended arrays of mixed colloidal crystals composed of two or more particle species has not been presented. For the preparation of such binary crystals, mixed suspensions containing particles of two different sizes with diameters of Lsmall ) 95 nm and Llarge ) 420 or 540 nm were used. As can be calculated using basic crystallographic models, a maximum diameter ratio of γmax ) Lsmall/ Llarge ≈ 0.225 was expected to be optimal for obtaining crystal-like structures by co-crystallization. Furthermore, the concentration ratios of the small and large particles, ν ) csmall/clarge, was an important parameter. In experimental series, both parameters were varied systematically, and their effect on the crystal structure and geometry was studied. For this, dilution series of particle suspensions with varying particle concentrations were prepared in microwell plates and spotted onto the prestructured substrates. The absolute concentration (solid content) of the small particles was varied from 0.056 and 0.106 g mL-1, and that of the large colloids was varied from 0.26 to 0.56 g mL-1. Parallel Array Readout Setup. The optical properties of the spotted colloidal crystal array were analyzed using the following readout setup. A white-light source was connected to a monochromator (Cornerstone 130, Oriel Instruments, Stratford, CT) followed by a polarizer and expansion optics in order to illuminate the entire array and enable its parallel readout. The sensor mounting was based on a goniometer (BGM, Newport, Darmstadt, Germany), and the detection unit contained a high-quantum-efficiency CCD camera (Sensicam, PCO, Kehlheim, Germany) and a focusing achromatic lens. The setup, as schematically illustrated in Figure 2, allowed the recording of reflection spectra of the crystal arrays over a given wavelength range under defined angles of incidence and reflection. These angles were controlled by adjusting two tilted mirrors and varying the goniometer position. In scanning mode, CCD camera images of the reflected light were taken over a wavelength range in 1 nm steps, and the recorded pictures were subsequently subjected to grayscale analysis in order to obtain a full reflection spectrum and the reflection maxima of all spots per array. By defining integration areas used by the in-house-developed image analysis software, thousands of reflection spectra per array could be analyzed in a parallel fashion.
Results and Discussion By visual inspection of the spotted arrays at varying observation angles under illumination with white light, strong opalescence was found. The iridescent color seen over the whole array surface is caused by Bragg diffraction of visible light at the crystal planes (10) Van Blaaderen, A. et al. Science 2002, 296, 106. (11) Ozin, G. A. et al. AdV. Mater. 2003, 15, 75. (12) Wang, D. et al. AdV. Mater. 2004, 16, 244.
Preparation Method for Colloidal Crystal Arrays
Langmuir, Vol. 23, No. 6, 2007 3481
in the colloid aggregates, which is a clear indicator of the successful formation of highly ordered colloidal crystals. The photographs in Figure 3 show the various substrates of the different formats introduced in Table 1, which were obtained by spotting dispersions of particles with diameter L ) 420 nm (with 10% ethylene glycol in the spotting solution). These results show that the opalescent behavior is essentially independent of the array format (number, diameter, and separation of sensor spots). To verify the high structural order within the spotted crystals, individual array spots were analyzed both by optical light microscopy and scanning electron microscopy. The optical micrographs in Figure 4 show sections of spotted arrays at different magnifications. The regular arrangement demonstrates the high precision with which the crystal spots could be deposited on the substrate. The position at each colloidal crystal spot is in perfect registry with the pattern of the underlying titanium structure, with the dark circles representing the holes in the surrounding titanium layer. From these micrographs, it is apparent that the colloid assembling on the array spots is independent of the spot diameter from 400 µm down to about 100 µm. At scanning electron microscopy (SEM) magnification, regularly ordered domains of the spotted crystals were recognizable in all formats. In Figure 5a and b, SEM pictures of format 3 (110 µm spot diameter) are shown as an example. A stepwise vertical increase in the number of colloid layers from the periphery to the center of the crystal is revealed. This terrace structure is assumed to result from the hemispherical shape of the meniscus of the suspension droplet during drying, acting as a self-removing template in crystal formation (Figure 6). Highly ordered crystal domains were found to be separated by cracks (Figure 5b), which are typically observed in colloidal crystals prepared by evaporation or vertical transfer. Such cracks are thought to be caused by drying, desolvation, and shrinking of the crystal. All of these findings clearly demonstrate the formation of colloidal crystals by the contact printing process, and the observed effects are in full agreement with the characteristics previously reported for other colloidal crystal structures13-16.
Binary Mixed Crystals. As known from atomic crystals, highly ordered structures composed of two differently sized constituents do exist as binary crystals. Such binary crystals are also possible for colloidal particles, and some examples have been described in the literature.17-19 In the fcc structure commonly formed by colloidal particles during evaporation of the suspending liquid, two types of interstitial sites are present, a tetrahedral and an octahedral site. From geometric considerations, the maximum size ratio of a small particle, which can be accommodated inside these interstitial sites without disrupting the fcc lattice, is γmax ) Lsmall/Llarge ≈ 0.225. The preparation of binary colloidal crystal arrays by contact printing was attempted by spotting mixed suspensions containing large (Llarge ) 420 or 540 nm) and small (Lsmall ) 95 nm) PS particles (size ratio γ ) Lsmall/Llarge ≈ 0.226 or 0.176) at varying concentration ratios. The SEM picture of the binary colloidal crystal (format 3) that was spotted from a particle mixture (95 nm/0.056 g mL-1 and 420 nm/0.56 g mL-1) in Figure 7a reveals regularly ordered structures and a similar terrace-shaped structure observed for monodisperse crystal spots. Only a minor phase separation of the two differently sized particles localized at the edges of the spots was observed. Probably the smaller particles accumulate at the crystal edge as a result of a better fit of the small particle layer to the contour of the meniscus in the drying liquid droplet. Also, as similarly found with monodisperse colloid crystals, the edges of the single-crystal layers show crystal packing with square symmetry (possibly the (100) face of fcc) whereas the inner regions pack with the common hexagonal symmetry of the (111) face of a fcc lattice. The concentration ratio ν had a strong impact on the resulting crystal geometry of the prepared mixed colloidal crystals. A concentration ratio of ν ) 0.10 (Figure 7a-c) led to structures with one smaller particle located in the interstitial surface sites formed by the crystal lattice of the larger particles (Figure 7b and c). An increased concentration of small particles (ν ) 0.41) resulted in the deposition of a larger number of small particles in the interstitials formed by the hexagonal ordering of the large particles (Figure 7d and e). In general, lower concentration ratios led to the formation of binary crystals with good hexagonal ordering of the large particles (Figure 7b), and increasing concentrations of the small particles impaired the crystal formation of the large spheres to some extent, as can be seen by the partial disorder in Figure 7d. In all cases, the distribution of the small particles was not homogeneous over a complete crystal spot, but the specific composition of the binary crystals varied between highly regular packed areas (Figure 7c and e) and less ordered regions with a varying ratio of small and large particles (e.g., between crystal domains and at layer edges). Similar observations were made with larger particles (540 nm) in combination with the small 95 nm colloids having a size ratio of γ ) 0.176, which also yielded well-ordered binary crystals at lower concentration ratios and less ordered or disordered mixed colloid films at higher concentration ratios. In summary, the structure and geometry of mixed colloidal crystals could be controlled to a large extent by the choice of particle ratios and concentrations. Parallel Array Readout. In addition to characterizing the crystal arrays by visual inspection and microscopy, their optical properties were examined by reflection spectroscopy using the
(13) Fustin, C. A.; Glasser, G.; Spiess, H. W.; Jonas, U. Langmuir 2004, 20, 9114-9123. (14) Goldenberg, L. M.; Wagner, J.; Stumpe, J.; Paulke, B. R.; Gornitz, E. Physica E 2003, 17, 433-435. (15) Nozar, P.; Diongi, C.; Migliori, A.; Calestani, G.; Cademartiri, L. Synth. Met. 2003, 667-670.
(16) Popov, O.; Lirtsman, V.; Kopnov, F.; Davidov, D.; Saraidarov, T.; Reisfeld, R. Synth. Met. 2003, 139, 643-647. (17) Kaplan, P. D.; Rouke, J. L.; Yodh, A. G.; Pine, D. J. Phys. ReV. Lett. 1994, 72, 582. (18) Schofield, A. B. Phys. ReV. E 2001, 64, 051403. (19) Wette, P.; Schoepe, H. J.; Palberge, T. J. Chem. Phys. 2005, 122, 144901.
Figure 2. Schematic drawing of the reflection readout setup consisting of a white-light source, monochromator, polarizer, tilted mirrors, goniometer, collection lenses, and CCD camera. The crystal array chips are mounted onto the goniometer, and reflection spectra are recorded at defined incidence and reflection angles.
3482 Langmuir, Vol. 23, No. 6, 2007
Burkert et al.
Figure 3. Within all three formats, visual inspection of the spotted crystal arrays revealed strong opalescent behavior, corroborating the successful formation of highly ordered colloidal crystals. (a) Format 1, (b) format 2, and (c) format 3 of Table 1.
Figure 4. Light microscopy images of spotted crystal arrays of investigated formats at different magnifications. Crystal spots were deposited precisely on the glass fields with diameter d (appearing dark) as indicated for each format.
Figure 5. (a) SEM image of an individual crystal spot showing the terrace-shaped form of the colloid crystal. (b) Magnification of the crystal surface exhibits highly ordered parts that are separated by cracks, formed during drying, desolvation, and shrinking of the crystal.
Figure 6. Schematic representation of the formation process for the terrace-shaped crystal under the influence of the hemispherical liquid droplet.
setup described above with a 1536 sensor array (format 2). In this study, we used an angle of incidence of the illuminating light of θ ) 73° and varied the wavelength range from λ ) 650 to 800 nm. Reflection spectra of all array fields were recorded simultaneously in parallel detection mode and were visualized by plotting the intensity of the reflected light against the wavelength of the incident light for the selected wavelength range. To compare monodisperse and binary colloidal crystals in one experiment, an array was spotted with three different colloidal solutions, receiving n ) 512 sensors in each case. Figure
8 shows representative reflection spectra of one spot of the crystal array obtained by spotting a monodisperse colloid suspension of PS particles with dparticle ) 420 nm, a binary colloidal solution with 420/95 nm PS particles with ν ) 0.10, and a binary colloid solution of 420/95 nm particles with ν ) 0.41. Here, the intensity was normalized to the intensity of a control field without colloidal crystals in order to eliminate the background arising from the light source spectra, scattered light, and the sensitivity characteristics of the CCD that was used. The exact position of the maximum wavelength was simultaneously determined for all spots/array by an automated Gaussian fitting routine implemented in the image analysis software. To evaluate the homogeneity of the produced crystal spots across the entire array, the wavelength-dependent reflectivity maxima of all sensor fields were compared. The distribution of the reflective maximum wavelengths λstop of 512 spots is presented
Preparation Method for Colloidal Crystal Arrays
Langmuir, Vol. 23, No. 6, 2007 3483
Figure 7. SEM images of mixed colloidal crystals. (a) SEM exhibits the same terrace-shaped form as found for monodisperse colloidal crystals. (b) Binary colloidal crystal as in part a that was obtained by co-spotting a mixture of two differently sized particles at a weight concentration ratio of ν ) 0.10. (c) Enlarged view of the image in part b with individual small particles in the 3-fold hollow surface sites. (d) Binary colloidal crystal spotted at a weight concentration ratio of ν ) 0.41 with partial disorder. (e) Enlargement of an ordered domain in part d with many small particles separating the hexagonally packed large spheres.
Figure 8. Representative reflection spectra of one crystal spot and histogram plots of the wavelength-dependent reflectivity (stop band wavelength) for 512 spots demonstrating the distribution of the stop band position across the entire array using the parallel reflection spectroscopy setup. (a) Monodisperse colloids (420 nm), (b) binary colloidal mixture (420/95 nm) of ν ) 0.10, and (c) binary colloidal mixture (420/95 nm) of ν ) 0.41.
in the histogram plots in Figure 8. Spotted monodisperse colloid crystals showed a distinct maximum in the spectra, exhibiting a narrow band distribution of the stop band positions across the crystal array (mean wavelength of λstop.mean ) 719.9 ( 4.7 nm) and hence demonstrating excellent homogeneity. Also, the stop band positions of binary colloid crystals (420/ 95 nm) with a weight concentration ratio of ν ) 0.10 were reproducible, although the maximum intensity of the reflection was slightly reduced and the distribution of the stop band positions was widened (mean wavelength of λstop.mean ) 721.7 ( 9.3 nm) with respect to those of monodisperse crystals. Notably, a red shift of λstop.mean was found, in agreement with Bragg’s theory
for binary colloidal systems. Theoretically, the shift can be explained by an increase in neff for ideal binary crystals. At a higher relative concentration ratio (ν ) 0.41), the reproducibility of the spotted binary colloidal crystals was reduced. Although a reflection maximum was observed, the peak intensity was smaller, and the half-width of the peak was larger compared to that of crystals with ν ) 0.10, which can be explained by a higher degree of disorder in the binary colloidal crystal with ν ) 0.41. The enhanced phase separation of the particles may also have led to a broader distribution of the stop band positions (mean wavelength of λstop.mean ) 715.6 ( 20.2 nm). This result was in good agreement with the higher imminent disorder also
3484 Langmuir, Vol. 23, No. 6, 2007
recognized in the SEM pictures (Figure 7) for spotted binary colloidal crystals obtained by increasing the relative concentration ratio from ν ) 0.1 to 0.41.
Conclusions We described a spotting technology for the production of optically active arrays of monodisperse as well as mixed binary colloidal crystals. For this purpose, a pintool plotter consisting of a high-precision portal robot and steel pins was found to be suitable for the arraying of colloid suspension droplets onto prestructured substrate chips with high reproducibility. By this method, colloid crystal arrays could be fabricated in parallel in a simple automated fashion. To obtain a miniaturized sensor system, we applied three functional array formats, with the highest degree of miniaturization being reached by an array of 9600 addressable fields per microscope slide. The obtained crystal arrays showed strong opalescent behavior after drying under controlled environmental conditions, which indicated the successful formation of highly ordered colloidal crystals, and optical microscope and SEM images indeed revealed characteristic crystal features in the dried array spots from monodisperse particle suspensions as well as from binary mixtures. Notably, under appropriate conditions only negligible phase separation after crystallization was found for mixed colloidal crystal systems. In the latter case, the influence of the concentration ratio and particle diameter ratio on the crystal properties and geometries was examined and found to be crucial to the quality of the crystals. To obtain efficient readouts of the optical properties of the crystal arrays, a spectroscopic measurement setup by means of
Burkert et al.
CCD camera imaging was established to examine in parallel the stop band wavelength of all array spots simultaneously. Using this setup, we could measure the stop band wavelengths for the spotted monodisperse colloid crystals in parallel, further corroborating the formation of highly ordered optical structures with Bragg diffraction in the visible/near-infrared range of light. The results of such reflection measurements showed promising results not only in terms of the quality of detectable stop bands but also in the homogeneity of crystals across the entire spotted array. The quality of spotted binary colloidal crystal arrays was found to be highly impacted by an increasing relative weight particle concentration ratio, whereas for monodisperse colloidal crystals the narrow distribution of the stop band position across the sensor array demonstrated the capability of fabricating highly reproducible crystal spots by the developed contact printing method with the pintool plotter. The prepared colloidal crystal arrays form the basis for further developments toward a potential application as a protein array sensor. For such purpose, changes in the mean refractive index of the colloid crystals upon binding ligands from solution to proteins immobilized in the interstitial sites could be used to realize a biosensor for the label-free and highly parallel detection of protein-ligand interactions. Acknowledgment. This work has been supported by the German Ministry of Education and Research BMBF (grants 0312015A and B). We thank Gunnar Glasser for the help with the SEM images. LA063122Z
