1142 Christian Leiterer1 Tanja Deckert-Gaudig1,2 Prabha Singh1,2 Janina Wirth1 Volker Deckert1,2 Wolfgang Fritzsche1 1 Leibniz
Institute of Photonic Technology Jena (IPHT), Jena, Germany 2 Institute of Physical Chemistry and Abbe Center of Photonics,Friedrich-SchillerUniversity Jena, Jena, Germany
Received November 3, 2014 Revised February 15, 2015 Accepted February 20, 2015
Electrophoresis 2015, 36, 1142–1148
Research Article
Dielectrophoretic positioning of single nanoparticles on atomic force microscope tips for tip-enhanced Raman spectroscopy Tip-enhanced Raman spectroscopy, a combination of Raman spectroscopy and scanning probe microscopy, is a powerful technique to detect the vibrational fingerprint of molecules at the nanometer scale. A metal nanoparticle at the apex of an atomic force microscope tip leads to a large enhancement of the electromagnetic field when illuminated with an appropriate wavelength, resulting in an increased Raman signal. A controlled positioning of individual nanoparticles at the tip would improve the reproducibility of the probes and is quite demanding due to usually serial and labor-intensive approaches. In contrast to commonly used submicron manipulation techniques, dielectrophoresis allows a parallel and scalable production, and provides a novel approach toward reproducible and at the same time affordable tip-enhanced Raman spectroscopy tips. We demonstrate the successful positioning of an individual plasmonic nanoparticle on a commercial atomic force microscope tip by dielectrophoresis followed by experimental proof of the Raman signal enhancing capabilities of such tips. Keywords: Dielectrophoresis / Nanoparticle / Raman spectroscopy DOI 10.1002/elps.201400530
1 Introduction Raman spectroscopy is an established spectroscopic technique for specific material characterization and is applied in complex fields such as identification of microorganisms [1] and classification of tumor tissue [2]. However, its generally low sensitivity hampers a more general application in particular if small sample volumes should be addressed. The sensitivity can be increased by surface-enhanced Raman spectroscopy based on a local enhancement of the electromagnetic field at plasmonic nanostructures [3, 4]. Due to the near-field character of this effect, the recorded spectral information is dominated by contributions from molecules in the immediate vicinity (lower nanometer range) of the nanostructures. In order to map the lateral distribution of analytes, atomic force microscope (AFM) tips can be modified with plasmonic features (sharp edges, nanoparticles) and raster-scanned over the sample surface during Raman spectrum acquisition. This tip-enhanced Raman spectroscopy (TERS) [5, 6] technique allows the collection of laterally resolved spectral information with nanometer precision. The achieved lateral resolution is determined by step-size and nature of the utilized nanos-
Correspondence: Dr. Christian Leiterer, Leibniz Institute of Photonic Technology Jena (IPHT), Albert-Einstein-Straße 9, 07745 Jena, Germany E-mail: [email protected]
Abbreviations: DEP, dielectrophoresis; p-NTP, p-nitrothiophenol; TERS, tip-enhanced Raman spectroscopy C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
tructure and can reach the subnanometer scale [7]. For this purpose the generation of reproducible and affordable TERS tips remains an important issue. Cost-efficient TERS tip preparation includes silver or gold evaporation or sputtering [5,8] yielding a metal island film on the tip. While the general usability of such tips is reasonably good, nevertheless both methods intrinsically yield nanoparticles distributions on the apex that generally vary in terms of size and shape. Consequently, spectroscopic properties and also topography quality can vary considerably from tip to tip. Moreover, since the complete tip is covered with potentially field enhancing metal nanostructures, experiments in liquids are challenging due to the numerous enhancement sites that can lead to signals from undesired positions. To circumvent such difficulties, either tips with an appropriate physical or chemical coating are required or a single plasmonic nanoparticle (preferentially spectroscopically precharacterized) is exclusively attached to the tip apex. Such a nanoparticle could be wet-chemically synthesized, thus having the advantage of a highly crystalline structure compared to polycrystalline structures obtained from sputtering approaches [9]. It is known that highly crystalline particles work at higher plasmonic efficiency due to a well-defined charged distribution (less electron damping) within the particles [10]. Therefore, these particles would show stronger Raman enhancements resulting in a better S/N (signal-to-noise ratio) TERS imaging.
Colour Online: See the article online to view Figs. 1 and 3-5 in in colour
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 1142–1148
A direct growth of an isolated particle at an AFM probe by photocatalytic [11] or wet-chemical Ag reduction [12] lacks a possible precharacterization and size control. For presynthesized single nanostructures, only elaborate and therefore time-consuming (and quite costly) procedures have been described, such as nanomanipulation using a glue-coated tip [13], electron beam induced deposition-based welding of gold-capped Si-nanowires [14], or focused ion beam (FIB) based approaches [15]. All these procedures are serial by design and require extended manual fine-tuning to produce a single particle at a tip. Therefore, these techniques are difficult to adapt for batch processing as, e.g. established for standard AFM tips, and lack high reproducibility. Electrical field based manipulation methods such as dielectrophoresis (DEP) are known to enable the positioning of individual micro- and nanostructures [16–20] and have already been applied to the positioning of plasmonic nanostructures [21–26]. DEP manipulates dielectric particles by a nonuniform (AC) electric field. This method was introduced by Pohl in 1951 for the manipulation of cells and other micro sized dielectric materials and became known as DEP [27, 28]. Throughout the years this technique was adapted to manipulate also nanosized objects independent from their inherent conductivity. Recently, the applicability of DEP for the attachment of single micrometer sized Polystyrene beads and Ag wires to AFM tips has been reported [29, 30], with resulting tip diameters of 100–200 nm. In order to investigate DNA or protein strand-like samples, nanoparticles with a significantly smaller radius are required as signal enhancers to reduce the probed sample volume to as few molecules as possible. A scheme for the envisioned DEP-based TERS tip production is shown in Fig. 1. In the initial step an AFM probe is dipped in a solution of colloidal metal nanoparticles with desired plasmonic properties. For example, chemically synthesized nanoparticles with a plasmon resonance matching the excitation light source can be employed (Fig. 1A). It should be noted that in principle any kind of nanostructure (i.e. semiconductors or carbon nanostructures) can be used in this step. The only premise is a colloidal dispersion in a medium that allows the polarization of the nanostructures for a positive DEP process. In the next step, an alternating, nonuniform electric field (DEP) is used to force a nanostructure toward the AFM probe (Fig. 1B). After switching off the field the nanostructure will stick to the tip through strong adhesive forces at the nanoscale and the tip can be cleaned by a simple washing and drying procedure. In the presented study, TERS tips with individual silver nanoparticle attached on their apex are obtained (Fig. 1C). In the presented results chemically synthesized particles in typical citrate buffer solution was used. A reduction of the conductivity from the particle solution was not necessary. Nevertheless, it should be mentioned that high salt concentration can have strong influence on the DEP process influencing the polarizability of the particles, therefore the salt concentration should be kept at a necessary minimum when performing positive DEP. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Nanoanalysis
1143
Figure 1. DEP-based attachment of plasmonic nanoparticles for TERS tip production. (A) Immerse AFM tip with colloidal nanoparticle solution. (B) A nonuniform alternating electric field forces a particle to the tip. (C) Turning off the field yields a TERS tip with an individual nanoparticle.
The advantage of this method is the ability to work with submicron precision on all kinds of nanostructures when adjusting the field parameters accordingly. The method requires comparably inexpensive equipment and can be potentially used in parallel by a simple parallel connection. In general, it is a novel approach to modify AFM tips, making them attractive for different fields of science. The presented proof-of-principle TERS measurements with DEP-metallized AFM tips on a model analyte demonstrate the feasibility of the method. www.electrophoresis-journal.com
1144
C. Leiterer et al.
2 Materials and methods 2.1 Nanoparticles and AFM tips Silver and gold nanoparticles (British Biocell, UK) of diameters between 15 and 100 nm were used in a concentration of 1010 –1013 P/mL.
2.2 Nanoelectrodes The gold nanoelectrode arrays used for the preliminary experiments were structured by e-beam lithography, having a gap width of about 80 nm and a thickness of about 50 nm. They were described in detail in a previous publication [26].
2.3 Dielectrophoresis In order to apply the AC field required for DEP, a function generator (Agilent 33220A, Agilent Technologies, B¨oblingen, Germany) providing up to 20 Vpp at open circuit and up to 10 Vpp at 50 Ohm, providing a frequency of up to 20 MHz. All experiments were carried out at 1 MHz, the voltage was varied from 1 to 20 Vpp . For the DEP experiments, 100 L droplets of nanoparticle solution were deposited on the substrates for the procedure.
2.4 Spectral measurements (microscopy spectrometer) The scattering spectra of individual nanoparticles in air have been acquired using an optical Axio Imager microscope (C.-Zeiss Microimaging, G¨ottingen, Germany) in dark field configuration with a fiber-coupled spectrometer. The light source was a tungsten halogen lamp with a continuous spectrum and a color temperature of 3200 K. This light was passed through a pinhole and a multimode fiber to an Acton Research SpectraPro 2300i microspectrometer (Princeton Instruments, Trenton, NJ, USA) with a grating with 150 lines and a Peltier cooled CCD camera. The pinhole was coplanar to the tube lens of the optical microscope, i.e. the pinhole was in the magnified, real image plane, and had a diameter of 100 m. In a first step, metal nanoparticles were immobilized on glass substrates and then investigated with dark field illumination. After that, the measuring spot was positioned next to a particle of interest and the background signal was measured. Then the dark field scattering spectrum of the nanoparticle was recorded. The spectral characteristics of the light source and the dark current were also measured for spectral correction of the data. Spectra were collected by the CCD camera as spectrographic images of 1024 pixels × 256 pixels, in which the horizontal axis represents the wavelength, and the vertical axis represents scattering intensity. The wavelength scale
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 1142–1148
was calibrated using a mercury wavelength standard. “Origin 8G” software (OriginLab Corporation, Northampton, MA, USA) was used to process the spectral data by the following equation to get Ires : Ires = ( INP − IBG ) / ( ILS − INL ) INP represents the raw signal from the single nanoparticle, IBG refers to the background signal next to the particle, ILS to the spectrum of the light source, and INL to the noise level of the system. The final spectrum Ires was then smoothed with a 25-point FFT-Filter and normalized to 1 in respect to the maximum wavelength to determine the peak position.
2.5 TERS imaging A home built inverted confocal Raman setup (microscope: IX-71, Olympus, Hamburg, Germany, spectrometer: Acton Advanced SP2750 A, SI, Germany) was combined with an AFM (NanoWizard II, JPK Instruments, Berlin, Germany) to perform measurements in back-scattering geometry through the sample support. Illumination and collection of the TERS signals was done via an oil immersion microscope objective (40X, 1.35NA, Olympus). A HeNe laser = 532 nm, P = 400 W (HNL 150L, Thorlabs, Newton, NJ, USA) was used for excitation. Further information regarding the TERS experiment can be found in Rasmussen and Deckert [31]. For TERS experiments, glass slides covered with smooth gold nanoplates were prepared following the procedure described in Deckert-Gaudig and Deckert [32]. A self-assembled monolayer of p-nitrothiophenol (p-NTP) on such gold surfaces was generated by immersing a substrate in a 9 mM ethanolic p-NTP solution for 24 h, followed by washing with ethanol and drying under argon. After attaching the respective nanoparticles to a commercial AFM tip (NSG10, NT-MDT, Moscow, Russia), the TERS tip was positioned in the laser focus. The sample was then scanned beneath the tip and an appropriate gold nanoplate was selected from the topography image. Spectra were recorded as follows. Tip with 40 nm silver particles: point-to-point distance 3 nm, acquisition time (tacq ) = 2 s; tip with 60 nm silver particles: point-to-point distance 8 nm, tacq = 5 s.
3 Results and discussion Prior to the DEP-induced attachment of nanoparticles to AFM tips, some preliminary experiments were carried out to verify the technical feasibility of precise positioning. For this purpose, a plane-tapered electrode array was used to check the attachment of a few or even a single metal nanoparticles to a tip-shaped electrode. Solutions of colloidal gold nanoparticles of different sizes (15, 30, 100 nm) were tested regarding their suitability for DEP attachment. A selection of 50 simultaneously contacted electrode gaps (80–120 nm) was
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 1142–1148
Nanoanalysis
1145
Figure 2. Positioning of individual nanoparticles to nanoelectrodes: an electrode array of 50 electrode arrangements with 80 nm gap size was used for the DEP of colloidal gold nanoparticles with different diameters (see legend in the images). The low quality of the SEM figures is due to the fact that the resolution of the microscope is at its limit. Beside gaps bridged by particle chains, the attachment of just a single particle on each electrode could be realized. Details can be found in the following publication [26].
used to attach gold nanoparticles to the electrode via positive DEP. In Fig. 2, selected electrodes with successfully attached Au nanoparticles are shown. Individual nanoparticles of different sizes (15, 30, 100 nm) are used here. In the next step, Ag nanoparticles were positioned on a doped silicon AFM tip using DEP. Doped AFM tips were chosen to warrant sufficient conductivity, a prerequisite for a DEP suitable electrode. In order to monitor the deposition in real-time, a distance of about 20 m was chosen between the tip immersed into an Ag nanoparticle solution and the counterelectrode. Due to the enlarged distance (compared to the 80 nm gap size used in the previous experiment in Fig. 2), the voltage had to be increased to 20 Vpp to ensure a sufficiently high electrical field for Ag nanoparticle attraction. The process was recorded in real-time with an optical microscope using a water immersion objective to provide optimal resolution. Although metal nanoparticles are in principal detectable at single particle level using optical dark field microscopy developed already a century ago [33], this technique is not applicable here due to the large dark field (background)
Figure 3. DEP manipulation followed in real-time with bright field microscopy. Ag nanoparticlesare attached to an AFM tip using an AC field @20 Vpp, 1 MHz and are subsequently released. (A)power off (B) power on, particle chain assembled at the tip (C) power on, particles floating awayfrom the tip.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. (A) Scheme showing the stability of assembled nanoparticle structures before (left) and after the electrical field has been turned off (right). Top: In case of electrodes on a substrate, the substrate-nanoparticle interactions stabilize the resulting structure. Bottom: Without a stabilizing substrate, particles are usually released after the field has been turned off, only the very first particle at the electrode often remains. (B) Scanning electron micrograph images of a typical metal nanoparticle assembly (electrode on substrate, see (A) left) at the substrate surface (left: 15 nm gold; right: 60 nm gold). (C) Scanning electron micrograph images of an AFM tip (free electrode, see (A) bottom) with a single metal nanoparticle (40 nm) at the tip apex positioned by DEP (left: overview; right: zoom).
signal of the AFM tip itself. Consequently, the observed features originate from particle ensembles (e.g. chains). The image sequence in Fig. 3 shows particles being attracted by the AFM probe electrode, positioned and immobilized there at the tip apex (Fig. 3B), and released after the power has been turned off (Fig. 3C). The observed release after the field was turned off raised the question, whether the particles dragged to the electrode by the electric field will stay at the AFM tip after the field has been switched off. Experiments with spherical metal [25, 26] as well as with elongated semiconductor [34] nanostructures show that any particle, which is in sufficient contact with a surface (electrode or any other substrate) can reside after the electrical field has been switched off, and usually remains there even after removal of the solution which suggests that strong interfacial forces are present. Only structures such as nanoparticle chains formed in solution (see Fig. 4A)
www.electrophoresis-journal.com
1146
C. Leiterer et al.
without any connection to substrates will be released. This effect could be even helpful for the present work by self-limiting the number of particles to exactly one: exclusively the very first in line being in direct contact to the AFM tip electrode. This is difficult to observe with light microscopy, however, can be easily checked with subsequent SEM of the tips. With this method, silver nanoparticles can be successfully attached to an AFM tip permanently without further chemical modification of the tip surface. This is especially important for TERS measurements, since chemical modification of the tip surface, e.g. silanization, would yield unwanted additional TERS signals. In Fig. 4C, SEM images of a DEP-generated TERS tip are shown. It is evident that the attachment of a single Ag nanoparticle at the apex is possible. As other particles attach only at distance larger than the diffraction limited spot size an interference upon irradiation with laser light during the actual TERS experiment will not occur. An advantage of the DEP-induced TERS tip production compared to, e.g. silver evaporation is the possibility to select nanoparticles of distinct size, shape, and appropriate plasmonic properties. Those silver nanoparticles are commercially available or can be wet-chemically synthesized following established protocols. To obtain ideal TERS results, the particle resonance is a crucial parameter. A maximum enhancement can only be realized if the particle resonance matches the exciting laser wavelength. Decreasing the nanoparticle concentration further will allow the DEP-based method to trap just one single nanoparticle at the entire tip (see Fig. 4C). This way, also experiments in liquid environments become accessible without interfering signals from particles beyond the actual TERS tip. The particle resonance with respect to the laser wavelength can be easily characterized by spectral measurements of single particles on a surface comparable to an AFM tip (e.g. doped silicon). In Fig. 5A, the spectrum of a single silver (60 nm, black) and gold (45 nm, red) nanoparticle on a silicon surface is shown, demonstrating the wavelength-tuning capabilities using different materials. Using silver nanoparticle modified AFM tips for TERS, silver oxidation has to be considered since an oxide layer influences the plasmon resonance of the nanoparticle. Even a thin layer can cause a significant red shift when silver nanoparticle spectra are measured in air. Thin layers of oxidized silver would have a refractive index of up to 2.5 or even higher [35, 36], which is significantly higher than elemental silver. In order to characterize this spectral shift, time-dependent spectral investigations of silver nanoparticles were performed (see Fig. 5B). Since the scattering spectrum of the silver nanoparticle on the surface is fairly close to the absorption spectrum for particle diameters below 50 nm, such measurements can be used to select suitable particles for the TERS measurements. Storing fresh silver nanoparticle covered tips in air for several days can obviously cause a significant red shift compared to the initial spectral measurements. This effect could be minimized by using fresh TERS tips or store them in inert atmosphere, to prevent oxidation. Here, only fresh tips were used for the TERS experiments. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 1142–1148
Figure 5. (A) Spectroscopic characterization of single silver (60 nm, black) and gold nanoparticles (45 nm, red) as potential TERS probes; (B) Spectroscopic monitoring of a single silver nanoparticle over 6 days. The red shift is caused by an increasing oxidized silver layer surrounding the particle. The less-saturated graphs in the background represent the raw data.
3.1 TERS measurements To evaluate the activity of the DEP-prepared tips, TERS experiments using 532 nm laser excitation were performed. For the measurements, p-NTP was adsorbed on smooth gold nanocrystals. In this way, a p-NTP monolayer is obtained which is known to provide reproducible TERS spectra with a prevailing characteristic band at 1332 cm−1 (NO2 ) [37] and two further peaks around 1072 and 1560 cm−1 indicating phenyl ring vibrations. Figure 6A and B shows TERS spectra of the p-NTP sample measured with two different tips, both had only one single nanoparticle at the tip apex, which was confirmed by subsequent SEM imaging. Silver nanoparticles diameters were 40 and 60 nm, respectively. It is noteworthy that for both measurements the same sample was used and spectra were recorded along a single line on equidistant acquisition points (3 and 8 nm, respectively). Evidently, all spectra are consistent and dominated by the NO2 marker band, which nicely correlates with the literature. The band at 521 cm−1 is www.electrophoresis-journal.com
Nanoanalysis
Electrophoresis 2015, 36, 1142–1148
Figure 6. TERS spectra of p-nitrothiophenol (p-NTP) collected with two different tips on adjacent points. (A) 60 nm Ag nanoparticle at tip apex, tacq = 2 s, I(NO2) = 470 cts; (B) 40 nm Ag nanoparticle at tip apex, tacq = 5 s, I(NO2) = 160 cts. P = 400 W at 532 nm. Point-to-point distance is 3 nm (A) and 8 nm (B), respectively. For better visualization the spectra are plotted with an offset.
assigned to Si-Si modes of the TERS tip and can be regarded as an internal standard. As expected, the tip with attached 60 nm particles result in a three times higher Raman signal enhancement at 2.5 times shorter acquisition times compared to the tip with 40 nm particles due to the stronger localized surface plasmon effect from the bigger nanoparticles excited with 532 nm. The spectra plainly visualize that as expected the TERS signal enhancement depends on the probes geometrical factors (size). This influences the enhanced electromagnetic field generated by surface plasmons [38]. Consequently, the results demonstrate that active and specifically tailored TERS tips can be prepared by the DEP method described above. Most importantly in this case a single nanoparticle at the tip apex provides an excellent device for TERS. While higher enhancements can be achieved by clusters of particles, with respect to reproducibility, a single particle is the preferred geometry of a TERS probe, as this excludes complex interactions between many nanoparticles.
4 Concluding remarks In this contribution, a novel method to produce TERS active AFM tips is presented. Using positive DEP, silver nanoparticles from a colloidal solution can be positioned on the APEX of a commercial AFM tips. An advantage of this method is the attachment of only a few isolated particles with specific size and shape compared to the complex island films that are obtained by silver evaporation. It is demonstrated that it is also possible to immobilize a single nanoparticle at the tip apex. Due to better control which particles (material, shape, size) to select a more constant signal enhancement maybe achieved for future TERS measurement, which would also lead to more reproducible results in the acquisition of TERS spectra. From a colloidal solution of commercially available or home-made nanoparticles of distinct size, a spectroscopic characterization, in other words preselection can be done. The single particle absorption spectrum correlates with the C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1147
shape and size and consequently with the localized surface plasmon resonance. For maximum signal enhancement, this should preferably match (or be as close as possible to) the used laser wavelength. In this study, nanoparticles with a diameter of 40 and 60 nm silver were used for DEP positioning and TERS measurements were carried out with such tips. The single nanoparticles were firmly immobilized on the tip apex enabling TERS spectra acquisition of a p-NTP monolayer on consecutive points. With both tips consistent and reproducible spectra could be collected, which were in agreement with the literature. As expected, the tip equipped with 60 nm silver particles yielded spectra with better signalto-noise ratio than that with 40 nm particles because their plasmon enhancement is significantly stronger when excited with 532 nm. Storage of such tips in an argon atmosphere prevents oxidation and preserves their activity and does not change the plasmon resonance for days. We thank Antti-Pekka Eskelinen and P¨aivi T¨orm¨a for providing the nanoelectrode arrays and assistence with the respective experiments. We acknowledge Robert Kretschmer, Norbert Jahr, and Matthias Urban for valuable comments and discussion, and Franka Jahn for SEM imaging. Funding was provided by DFG (DEP4TERS, FR 1348/19-1) BMBF (NAWION FKZ: 16SV5386K, V4MNI014), TKWFK project Microplex (FKZ: PE113-1), DAAD project PPP Finland (FKZ: 50020468), and the state of Thuringia (FKZ: 2011 FE 9048; 2011 VF 0016). The authors have declared no conflict of interest.
5 References ¨ [1] Rosch, P., Schmitt, M., Kiefer, W., Popp, J., J. Mol. Str. 2003, 661–662, 363–369. [2] Pujary, P., Maheedhar, K., Krishna, C. M., Pujary, K., Pathol. Res. Int. 2011, 2011, 1–9. [3] Campion, A., Kambhampati, P., Chem. Soc. Rev. 1998, 27, 241. [4] Moskovits, M., Rev. Mod. Phys. 1985, 57, 783–826. ¨ [5] Stockle, R. M., Suh, Y. D., Deckert, V., Zenobi, R., Chem. Phys. Lett. 2000, 318, 131–136. [6] Bailo, E., Deckert, V., Chem. Soc. Rev. 2008, 37, 921. [7] Zhang, R., Zhang, Y., Dong, Z. C., Jiang, S., Zhang, C., Chen, L. G., Zhang, L., Liao, Y., Aizpurua, J., Luo, Y., Yang, J. L., Hou, J. G., Nature 2013, 498, 82–86. [8] Hayazawa, N., Inouye, Y., Sekkat, Z., Kawata, S., J. Chem. Phys. 2002, 117, 1296. [9] Ditlbacher, H., Hohenau, A., Wagner, D., Kreibig, U., Rogers, M., Hofer, F., Aussenegg, F. R., Krenn, J. R., Phys. Rev. Lett. 2005, 95, 257403. [10] Kusar, P., Gruber, C., Hohenau, A., Krenn, J. R., Nano Lett. 2012, 12, 661–665. [11] Okamoto, T., Yamaguchi, I., J. Microsc. 2001, 202, 100–103. [12] Wang, J. J., Saito, Y., Batchelder, D. N., Kirkham, J., Robinson, C., Smith, D. A., Appl. Phys. Lett. 2005, 86, 263111.
www.electrophoresis-journal.com
1148
C. Leiterer et al.
[13] Kalkbrenner, T., Ramstein, M., Mlynek, J., Sandoghdar, V., J. Microsc. 2001, 202, 72–76. ¨ G., Geiger, R., Schreiber, [14] Becker, M., Sivakov, V., Andra, J., Hoffmann, S., Michler, J., Milenin, A. P., Werner, P., Christiansen, S. H., Nano Lett. 2007, 7, 75–80. [15] Min, Q., Santos, M. J. L., Girotto, E. M., Brolo, A. G., Gordon, R., J. Phys. Chem. C 2008, 112, 15098–15101. [16] Pohl, H., Crane, J., Biophys. J. 1971, 11, 711–727.
Electrophoresis 2015, 36, 1142–1148
[27] Pohl, H. A., J. Appl. Phys. 1951, 22, 869–871. [28] Pohl, H. A., Dielectrophoresis: The Behavior of Neutral Matter in Nonuniform Electric Fields., Cambridge University Press, Cambridge, UK, 1978. [29] You, Y., Purnawirman, N. A., Hu, H., Kasim, J., Yang, H., Du, C., Yu, T., Shen, Z., J. Raman Spectrosc. 2010, 41, 1156–1162.
[17] Washizu, M., Suzuki, S., Kurosawa, O., Nishizaka, T., Shinohara, T., IEEE Trans. Ind. Appl. 1994, 30, 835–843.
[30] Jose, J., Kress, S., Barik, A., Otto, L. M., Shaver, J., Johnson, T. W., Lapin, Z. J., Bharadwaj, P., Novotny, L., Oh, S.-H., ACS Photonics 2014, 1, 464–470.
[18] Raychaudhuri, S., Dayeh, S. A., Wang, D., Yu, E. T., Nano Lett. 2009, 9, 2260–2266.
[31] Rasmussen, A., Deckert, V., J. Raman Spectrosc. 2006, 37, 311–317.
[19] Wolff, A., Leiterer, C., Csaki, A., Fritzsche, W., Front. Biosci. 2008, 13, 6834–6840.
[32] Deckert-Gaudig, T., Deckert, V., Small 2009, 5, 432–436.
[20] Pethig, R., Biomicrofluidics 2010, 4, 022811. [21] Barsotti, R. J., Vahey, M. D., Wartena, R., Chiang, Y.-M., Voldman, J., Stellacci, F., Small 2007, 3, 488–499. [22] Kumar, S., Yoon, S.-H., Kim, G.-H., Curr. Appl. Phys. 2009, 9, 101–103. [23] Bernard, L., Calame, M., Molen, S. J. van der, Liao, J., ¨ Schonenberger, C., Nanotechnology 2007, 18, 235202. [24] Zoy, A., Nassiopoulos, A. A., Nassiopoulou, A. G., Nanotechnology 2007, 18, 345608.
[33] Mappes, T., Jahr, N., Csaki, A., Vogler, N., Popp, J., Fritzsche, W., Angew. Chem. Int. Edit. 2012, 51, 11208–11212. [34] Leiterer, C., Broenstrup, G., Jahr, N., Urban, M., Arnold, C., Christiansen, S., Fritzsche, W., J Nanopart. Res. 2013, 15, 1–7. [35] Pettersson, L. A. A., Snyder, P. G., Thin Solid Films 1995, 270, 69–72. [36] Bennett, J. M., Stanford, J. L., Ashley, E. J., J. Optic. Soc. Amer. 1970, 60, 224.
[25] Kretschmer, R., Fritzsche, W., Langmuir 2004, 20, 11797–11801.
[37] Schrojenstein Lantman, E. M. van, Deckert-Gaudig, T., Mank, A. J. G., Deckert, V., Weckhuysen, B. M., Nature Nanotechnol. 2012, 7, 583–586.
[26] Leiterer, C., Berg, S., Eskelinen, A.-P., Csaki, A., Urban, ¨ a, ¨ P., Fritzsche, W., Sens. Actuators B-Chem. M., Torm 2013, 176, 368–373.
[38] Kelly, K. L., Coronado, E., Zhao, L. L., Schatz, G. C., J. Phys. Chem. B 2003, 107, 668–677.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
Institute of Photonic Technology Jena (IPHT), Jena, Germany 2 Institute of Physical Chemistry and Abbe Center of Photonics,Friedrich-SchillerUniversity Jena, Jena, Germany
Received November 3, 2014 Revised February 15, 2015 Accepted February 20, 2015
Electrophoresis 2015, 36, 1142–1148
Research Article
Dielectrophoretic positioning of single nanoparticles on atomic force microscope tips for tip-enhanced Raman spectroscopy Tip-enhanced Raman spectroscopy, a combination of Raman spectroscopy and scanning probe microscopy, is a powerful technique to detect the vibrational fingerprint of molecules at the nanometer scale. A metal nanoparticle at the apex of an atomic force microscope tip leads to a large enhancement of the electromagnetic field when illuminated with an appropriate wavelength, resulting in an increased Raman signal. A controlled positioning of individual nanoparticles at the tip would improve the reproducibility of the probes and is quite demanding due to usually serial and labor-intensive approaches. In contrast to commonly used submicron manipulation techniques, dielectrophoresis allows a parallel and scalable production, and provides a novel approach toward reproducible and at the same time affordable tip-enhanced Raman spectroscopy tips. We demonstrate the successful positioning of an individual plasmonic nanoparticle on a commercial atomic force microscope tip by dielectrophoresis followed by experimental proof of the Raman signal enhancing capabilities of such tips. Keywords: Dielectrophoresis / Nanoparticle / Raman spectroscopy DOI 10.1002/elps.201400530
1 Introduction Raman spectroscopy is an established spectroscopic technique for specific material characterization and is applied in complex fields such as identification of microorganisms [1] and classification of tumor tissue [2]. However, its generally low sensitivity hampers a more general application in particular if small sample volumes should be addressed. The sensitivity can be increased by surface-enhanced Raman spectroscopy based on a local enhancement of the electromagnetic field at plasmonic nanostructures [3, 4]. Due to the near-field character of this effect, the recorded spectral information is dominated by contributions from molecules in the immediate vicinity (lower nanometer range) of the nanostructures. In order to map the lateral distribution of analytes, atomic force microscope (AFM) tips can be modified with plasmonic features (sharp edges, nanoparticles) and raster-scanned over the sample surface during Raman spectrum acquisition. This tip-enhanced Raman spectroscopy (TERS) [5, 6] technique allows the collection of laterally resolved spectral information with nanometer precision. The achieved lateral resolution is determined by step-size and nature of the utilized nanos-
Correspondence: Dr. Christian Leiterer, Leibniz Institute of Photonic Technology Jena (IPHT), Albert-Einstein-Straße 9, 07745 Jena, Germany E-mail: [email protected]
Abbreviations: DEP, dielectrophoresis; p-NTP, p-nitrothiophenol; TERS, tip-enhanced Raman spectroscopy C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
tructure and can reach the subnanometer scale [7]. For this purpose the generation of reproducible and affordable TERS tips remains an important issue. Cost-efficient TERS tip preparation includes silver or gold evaporation or sputtering [5,8] yielding a metal island film on the tip. While the general usability of such tips is reasonably good, nevertheless both methods intrinsically yield nanoparticles distributions on the apex that generally vary in terms of size and shape. Consequently, spectroscopic properties and also topography quality can vary considerably from tip to tip. Moreover, since the complete tip is covered with potentially field enhancing metal nanostructures, experiments in liquids are challenging due to the numerous enhancement sites that can lead to signals from undesired positions. To circumvent such difficulties, either tips with an appropriate physical or chemical coating are required or a single plasmonic nanoparticle (preferentially spectroscopically precharacterized) is exclusively attached to the tip apex. Such a nanoparticle could be wet-chemically synthesized, thus having the advantage of a highly crystalline structure compared to polycrystalline structures obtained from sputtering approaches [9]. It is known that highly crystalline particles work at higher plasmonic efficiency due to a well-defined charged distribution (less electron damping) within the particles [10]. Therefore, these particles would show stronger Raman enhancements resulting in a better S/N (signal-to-noise ratio) TERS imaging.
Colour Online: See the article online to view Figs. 1 and 3-5 in in colour
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 1142–1148
A direct growth of an isolated particle at an AFM probe by photocatalytic [11] or wet-chemical Ag reduction [12] lacks a possible precharacterization and size control. For presynthesized single nanostructures, only elaborate and therefore time-consuming (and quite costly) procedures have been described, such as nanomanipulation using a glue-coated tip [13], electron beam induced deposition-based welding of gold-capped Si-nanowires [14], or focused ion beam (FIB) based approaches [15]. All these procedures are serial by design and require extended manual fine-tuning to produce a single particle at a tip. Therefore, these techniques are difficult to adapt for batch processing as, e.g. established for standard AFM tips, and lack high reproducibility. Electrical field based manipulation methods such as dielectrophoresis (DEP) are known to enable the positioning of individual micro- and nanostructures [16–20] and have already been applied to the positioning of plasmonic nanostructures [21–26]. DEP manipulates dielectric particles by a nonuniform (AC) electric field. This method was introduced by Pohl in 1951 for the manipulation of cells and other micro sized dielectric materials and became known as DEP [27, 28]. Throughout the years this technique was adapted to manipulate also nanosized objects independent from their inherent conductivity. Recently, the applicability of DEP for the attachment of single micrometer sized Polystyrene beads and Ag wires to AFM tips has been reported [29, 30], with resulting tip diameters of 100–200 nm. In order to investigate DNA or protein strand-like samples, nanoparticles with a significantly smaller radius are required as signal enhancers to reduce the probed sample volume to as few molecules as possible. A scheme for the envisioned DEP-based TERS tip production is shown in Fig. 1. In the initial step an AFM probe is dipped in a solution of colloidal metal nanoparticles with desired plasmonic properties. For example, chemically synthesized nanoparticles with a plasmon resonance matching the excitation light source can be employed (Fig. 1A). It should be noted that in principle any kind of nanostructure (i.e. semiconductors or carbon nanostructures) can be used in this step. The only premise is a colloidal dispersion in a medium that allows the polarization of the nanostructures for a positive DEP process. In the next step, an alternating, nonuniform electric field (DEP) is used to force a nanostructure toward the AFM probe (Fig. 1B). After switching off the field the nanostructure will stick to the tip through strong adhesive forces at the nanoscale and the tip can be cleaned by a simple washing and drying procedure. In the presented study, TERS tips with individual silver nanoparticle attached on their apex are obtained (Fig. 1C). In the presented results chemically synthesized particles in typical citrate buffer solution was used. A reduction of the conductivity from the particle solution was not necessary. Nevertheless, it should be mentioned that high salt concentration can have strong influence on the DEP process influencing the polarizability of the particles, therefore the salt concentration should be kept at a necessary minimum when performing positive DEP. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Nanoanalysis
1143
Figure 1. DEP-based attachment of plasmonic nanoparticles for TERS tip production. (A) Immerse AFM tip with colloidal nanoparticle solution. (B) A nonuniform alternating electric field forces a particle to the tip. (C) Turning off the field yields a TERS tip with an individual nanoparticle.
The advantage of this method is the ability to work with submicron precision on all kinds of nanostructures when adjusting the field parameters accordingly. The method requires comparably inexpensive equipment and can be potentially used in parallel by a simple parallel connection. In general, it is a novel approach to modify AFM tips, making them attractive for different fields of science. The presented proof-of-principle TERS measurements with DEP-metallized AFM tips on a model analyte demonstrate the feasibility of the method. www.electrophoresis-journal.com
1144
C. Leiterer et al.
2 Materials and methods 2.1 Nanoparticles and AFM tips Silver and gold nanoparticles (British Biocell, UK) of diameters between 15 and 100 nm were used in a concentration of 1010 –1013 P/mL.
2.2 Nanoelectrodes The gold nanoelectrode arrays used for the preliminary experiments were structured by e-beam lithography, having a gap width of about 80 nm and a thickness of about 50 nm. They were described in detail in a previous publication [26].
2.3 Dielectrophoresis In order to apply the AC field required for DEP, a function generator (Agilent 33220A, Agilent Technologies, B¨oblingen, Germany) providing up to 20 Vpp at open circuit and up to 10 Vpp at 50 Ohm, providing a frequency of up to 20 MHz. All experiments were carried out at 1 MHz, the voltage was varied from 1 to 20 Vpp . For the DEP experiments, 100 L droplets of nanoparticle solution were deposited on the substrates for the procedure.
2.4 Spectral measurements (microscopy spectrometer) The scattering spectra of individual nanoparticles in air have been acquired using an optical Axio Imager microscope (C.-Zeiss Microimaging, G¨ottingen, Germany) in dark field configuration with a fiber-coupled spectrometer. The light source was a tungsten halogen lamp with a continuous spectrum and a color temperature of 3200 K. This light was passed through a pinhole and a multimode fiber to an Acton Research SpectraPro 2300i microspectrometer (Princeton Instruments, Trenton, NJ, USA) with a grating with 150 lines and a Peltier cooled CCD camera. The pinhole was coplanar to the tube lens of the optical microscope, i.e. the pinhole was in the magnified, real image plane, and had a diameter of 100 m. In a first step, metal nanoparticles were immobilized on glass substrates and then investigated with dark field illumination. After that, the measuring spot was positioned next to a particle of interest and the background signal was measured. Then the dark field scattering spectrum of the nanoparticle was recorded. The spectral characteristics of the light source and the dark current were also measured for spectral correction of the data. Spectra were collected by the CCD camera as spectrographic images of 1024 pixels × 256 pixels, in which the horizontal axis represents the wavelength, and the vertical axis represents scattering intensity. The wavelength scale
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 1142–1148
was calibrated using a mercury wavelength standard. “Origin 8G” software (OriginLab Corporation, Northampton, MA, USA) was used to process the spectral data by the following equation to get Ires : Ires = ( INP − IBG ) / ( ILS − INL ) INP represents the raw signal from the single nanoparticle, IBG refers to the background signal next to the particle, ILS to the spectrum of the light source, and INL to the noise level of the system. The final spectrum Ires was then smoothed with a 25-point FFT-Filter and normalized to 1 in respect to the maximum wavelength to determine the peak position.
2.5 TERS imaging A home built inverted confocal Raman setup (microscope: IX-71, Olympus, Hamburg, Germany, spectrometer: Acton Advanced SP2750 A, SI, Germany) was combined with an AFM (NanoWizard II, JPK Instruments, Berlin, Germany) to perform measurements in back-scattering geometry through the sample support. Illumination and collection of the TERS signals was done via an oil immersion microscope objective (40X, 1.35NA, Olympus). A HeNe laser = 532 nm, P = 400 W (HNL 150L, Thorlabs, Newton, NJ, USA) was used for excitation. Further information regarding the TERS experiment can be found in Rasmussen and Deckert [31]. For TERS experiments, glass slides covered with smooth gold nanoplates were prepared following the procedure described in Deckert-Gaudig and Deckert [32]. A self-assembled monolayer of p-nitrothiophenol (p-NTP) on such gold surfaces was generated by immersing a substrate in a 9 mM ethanolic p-NTP solution for 24 h, followed by washing with ethanol and drying under argon. After attaching the respective nanoparticles to a commercial AFM tip (NSG10, NT-MDT, Moscow, Russia), the TERS tip was positioned in the laser focus. The sample was then scanned beneath the tip and an appropriate gold nanoplate was selected from the topography image. Spectra were recorded as follows. Tip with 40 nm silver particles: point-to-point distance 3 nm, acquisition time (tacq ) = 2 s; tip with 60 nm silver particles: point-to-point distance 8 nm, tacq = 5 s.
3 Results and discussion Prior to the DEP-induced attachment of nanoparticles to AFM tips, some preliminary experiments were carried out to verify the technical feasibility of precise positioning. For this purpose, a plane-tapered electrode array was used to check the attachment of a few or even a single metal nanoparticles to a tip-shaped electrode. Solutions of colloidal gold nanoparticles of different sizes (15, 30, 100 nm) were tested regarding their suitability for DEP attachment. A selection of 50 simultaneously contacted electrode gaps (80–120 nm) was
www.electrophoresis-journal.com
Electrophoresis 2015, 36, 1142–1148
Nanoanalysis
1145
Figure 2. Positioning of individual nanoparticles to nanoelectrodes: an electrode array of 50 electrode arrangements with 80 nm gap size was used for the DEP of colloidal gold nanoparticles with different diameters (see legend in the images). The low quality of the SEM figures is due to the fact that the resolution of the microscope is at its limit. Beside gaps bridged by particle chains, the attachment of just a single particle on each electrode could be realized. Details can be found in the following publication [26].
used to attach gold nanoparticles to the electrode via positive DEP. In Fig. 2, selected electrodes with successfully attached Au nanoparticles are shown. Individual nanoparticles of different sizes (15, 30, 100 nm) are used here. In the next step, Ag nanoparticles were positioned on a doped silicon AFM tip using DEP. Doped AFM tips were chosen to warrant sufficient conductivity, a prerequisite for a DEP suitable electrode. In order to monitor the deposition in real-time, a distance of about 20 m was chosen between the tip immersed into an Ag nanoparticle solution and the counterelectrode. Due to the enlarged distance (compared to the 80 nm gap size used in the previous experiment in Fig. 2), the voltage had to be increased to 20 Vpp to ensure a sufficiently high electrical field for Ag nanoparticle attraction. The process was recorded in real-time with an optical microscope using a water immersion objective to provide optimal resolution. Although metal nanoparticles are in principal detectable at single particle level using optical dark field microscopy developed already a century ago [33], this technique is not applicable here due to the large dark field (background)
Figure 3. DEP manipulation followed in real-time with bright field microscopy. Ag nanoparticlesare attached to an AFM tip using an AC field @20 Vpp, 1 MHz and are subsequently released. (A)power off (B) power on, particle chain assembled at the tip (C) power on, particles floating awayfrom the tip.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. (A) Scheme showing the stability of assembled nanoparticle structures before (left) and after the electrical field has been turned off (right). Top: In case of electrodes on a substrate, the substrate-nanoparticle interactions stabilize the resulting structure. Bottom: Without a stabilizing substrate, particles are usually released after the field has been turned off, only the very first particle at the electrode often remains. (B) Scanning electron micrograph images of a typical metal nanoparticle assembly (electrode on substrate, see (A) left) at the substrate surface (left: 15 nm gold; right: 60 nm gold). (C) Scanning electron micrograph images of an AFM tip (free electrode, see (A) bottom) with a single metal nanoparticle (40 nm) at the tip apex positioned by DEP (left: overview; right: zoom).
signal of the AFM tip itself. Consequently, the observed features originate from particle ensembles (e.g. chains). The image sequence in Fig. 3 shows particles being attracted by the AFM probe electrode, positioned and immobilized there at the tip apex (Fig. 3B), and released after the power has been turned off (Fig. 3C). The observed release after the field was turned off raised the question, whether the particles dragged to the electrode by the electric field will stay at the AFM tip after the field has been switched off. Experiments with spherical metal [25, 26] as well as with elongated semiconductor [34] nanostructures show that any particle, which is in sufficient contact with a surface (electrode or any other substrate) can reside after the electrical field has been switched off, and usually remains there even after removal of the solution which suggests that strong interfacial forces are present. Only structures such as nanoparticle chains formed in solution (see Fig. 4A)
www.electrophoresis-journal.com
1146
C. Leiterer et al.
without any connection to substrates will be released. This effect could be even helpful for the present work by self-limiting the number of particles to exactly one: exclusively the very first in line being in direct contact to the AFM tip electrode. This is difficult to observe with light microscopy, however, can be easily checked with subsequent SEM of the tips. With this method, silver nanoparticles can be successfully attached to an AFM tip permanently without further chemical modification of the tip surface. This is especially important for TERS measurements, since chemical modification of the tip surface, e.g. silanization, would yield unwanted additional TERS signals. In Fig. 4C, SEM images of a DEP-generated TERS tip are shown. It is evident that the attachment of a single Ag nanoparticle at the apex is possible. As other particles attach only at distance larger than the diffraction limited spot size an interference upon irradiation with laser light during the actual TERS experiment will not occur. An advantage of the DEP-induced TERS tip production compared to, e.g. silver evaporation is the possibility to select nanoparticles of distinct size, shape, and appropriate plasmonic properties. Those silver nanoparticles are commercially available or can be wet-chemically synthesized following established protocols. To obtain ideal TERS results, the particle resonance is a crucial parameter. A maximum enhancement can only be realized if the particle resonance matches the exciting laser wavelength. Decreasing the nanoparticle concentration further will allow the DEP-based method to trap just one single nanoparticle at the entire tip (see Fig. 4C). This way, also experiments in liquid environments become accessible without interfering signals from particles beyond the actual TERS tip. The particle resonance with respect to the laser wavelength can be easily characterized by spectral measurements of single particles on a surface comparable to an AFM tip (e.g. doped silicon). In Fig. 5A, the spectrum of a single silver (60 nm, black) and gold (45 nm, red) nanoparticle on a silicon surface is shown, demonstrating the wavelength-tuning capabilities using different materials. Using silver nanoparticle modified AFM tips for TERS, silver oxidation has to be considered since an oxide layer influences the plasmon resonance of the nanoparticle. Even a thin layer can cause a significant red shift when silver nanoparticle spectra are measured in air. Thin layers of oxidized silver would have a refractive index of up to 2.5 or even higher [35, 36], which is significantly higher than elemental silver. In order to characterize this spectral shift, time-dependent spectral investigations of silver nanoparticles were performed (see Fig. 5B). Since the scattering spectrum of the silver nanoparticle on the surface is fairly close to the absorption spectrum for particle diameters below 50 nm, such measurements can be used to select suitable particles for the TERS measurements. Storing fresh silver nanoparticle covered tips in air for several days can obviously cause a significant red shift compared to the initial spectral measurements. This effect could be minimized by using fresh TERS tips or store them in inert atmosphere, to prevent oxidation. Here, only fresh tips were used for the TERS experiments. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2015, 36, 1142–1148
Figure 5. (A) Spectroscopic characterization of single silver (60 nm, black) and gold nanoparticles (45 nm, red) as potential TERS probes; (B) Spectroscopic monitoring of a single silver nanoparticle over 6 days. The red shift is caused by an increasing oxidized silver layer surrounding the particle. The less-saturated graphs in the background represent the raw data.
3.1 TERS measurements To evaluate the activity of the DEP-prepared tips, TERS experiments using 532 nm laser excitation were performed. For the measurements, p-NTP was adsorbed on smooth gold nanocrystals. In this way, a p-NTP monolayer is obtained which is known to provide reproducible TERS spectra with a prevailing characteristic band at 1332 cm−1 (NO2 ) [37] and two further peaks around 1072 and 1560 cm−1 indicating phenyl ring vibrations. Figure 6A and B shows TERS spectra of the p-NTP sample measured with two different tips, both had only one single nanoparticle at the tip apex, which was confirmed by subsequent SEM imaging. Silver nanoparticles diameters were 40 and 60 nm, respectively. It is noteworthy that for both measurements the same sample was used and spectra were recorded along a single line on equidistant acquisition points (3 and 8 nm, respectively). Evidently, all spectra are consistent and dominated by the NO2 marker band, which nicely correlates with the literature. The band at 521 cm−1 is www.electrophoresis-journal.com
Nanoanalysis
Electrophoresis 2015, 36, 1142–1148
Figure 6. TERS spectra of p-nitrothiophenol (p-NTP) collected with two different tips on adjacent points. (A) 60 nm Ag nanoparticle at tip apex, tacq = 2 s, I(NO2) = 470 cts; (B) 40 nm Ag nanoparticle at tip apex, tacq = 5 s, I(NO2) = 160 cts. P = 400 W at 532 nm. Point-to-point distance is 3 nm (A) and 8 nm (B), respectively. For better visualization the spectra are plotted with an offset.
assigned to Si-Si modes of the TERS tip and can be regarded as an internal standard. As expected, the tip with attached 60 nm particles result in a three times higher Raman signal enhancement at 2.5 times shorter acquisition times compared to the tip with 40 nm particles due to the stronger localized surface plasmon effect from the bigger nanoparticles excited with 532 nm. The spectra plainly visualize that as expected the TERS signal enhancement depends on the probes geometrical factors (size). This influences the enhanced electromagnetic field generated by surface plasmons [38]. Consequently, the results demonstrate that active and specifically tailored TERS tips can be prepared by the DEP method described above. Most importantly in this case a single nanoparticle at the tip apex provides an excellent device for TERS. While higher enhancements can be achieved by clusters of particles, with respect to reproducibility, a single particle is the preferred geometry of a TERS probe, as this excludes complex interactions between many nanoparticles.
4 Concluding remarks In this contribution, a novel method to produce TERS active AFM tips is presented. Using positive DEP, silver nanoparticles from a colloidal solution can be positioned on the APEX of a commercial AFM tips. An advantage of this method is the attachment of only a few isolated particles with specific size and shape compared to the complex island films that are obtained by silver evaporation. It is demonstrated that it is also possible to immobilize a single nanoparticle at the tip apex. Due to better control which particles (material, shape, size) to select a more constant signal enhancement maybe achieved for future TERS measurement, which would also lead to more reproducible results in the acquisition of TERS spectra. From a colloidal solution of commercially available or home-made nanoparticles of distinct size, a spectroscopic characterization, in other words preselection can be done. The single particle absorption spectrum correlates with the C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1147
shape and size and consequently with the localized surface plasmon resonance. For maximum signal enhancement, this should preferably match (or be as close as possible to) the used laser wavelength. In this study, nanoparticles with a diameter of 40 and 60 nm silver were used for DEP positioning and TERS measurements were carried out with such tips. The single nanoparticles were firmly immobilized on the tip apex enabling TERS spectra acquisition of a p-NTP monolayer on consecutive points. With both tips consistent and reproducible spectra could be collected, which were in agreement with the literature. As expected, the tip equipped with 60 nm silver particles yielded spectra with better signalto-noise ratio than that with 40 nm particles because their plasmon enhancement is significantly stronger when excited with 532 nm. Storage of such tips in an argon atmosphere prevents oxidation and preserves their activity and does not change the plasmon resonance for days. We thank Antti-Pekka Eskelinen and P¨aivi T¨orm¨a for providing the nanoelectrode arrays and assistence with the respective experiments. We acknowledge Robert Kretschmer, Norbert Jahr, and Matthias Urban for valuable comments and discussion, and Franka Jahn for SEM imaging. Funding was provided by DFG (DEP4TERS, FR 1348/19-1) BMBF (NAWION FKZ: 16SV5386K, V4MNI014), TKWFK project Microplex (FKZ: PE113-1), DAAD project PPP Finland (FKZ: 50020468), and the state of Thuringia (FKZ: 2011 FE 9048; 2011 VF 0016). The authors have declared no conflict of interest.
5 References ¨ [1] Rosch, P., Schmitt, M., Kiefer, W., Popp, J., J. Mol. Str. 2003, 661–662, 363–369. [2] Pujary, P., Maheedhar, K., Krishna, C. M., Pujary, K., Pathol. Res. Int. 2011, 2011, 1–9. [3] Campion, A., Kambhampati, P., Chem. Soc. Rev. 1998, 27, 241. [4] Moskovits, M., Rev. Mod. Phys. 1985, 57, 783–826. ¨ [5] Stockle, R. M., Suh, Y. D., Deckert, V., Zenobi, R., Chem. Phys. Lett. 2000, 318, 131–136. [6] Bailo, E., Deckert, V., Chem. Soc. Rev. 2008, 37, 921. [7] Zhang, R., Zhang, Y., Dong, Z. C., Jiang, S., Zhang, C., Chen, L. G., Zhang, L., Liao, Y., Aizpurua, J., Luo, Y., Yang, J. L., Hou, J. G., Nature 2013, 498, 82–86. [8] Hayazawa, N., Inouye, Y., Sekkat, Z., Kawata, S., J. Chem. Phys. 2002, 117, 1296. [9] Ditlbacher, H., Hohenau, A., Wagner, D., Kreibig, U., Rogers, M., Hofer, F., Aussenegg, F. R., Krenn, J. R., Phys. Rev. Lett. 2005, 95, 257403. [10] Kusar, P., Gruber, C., Hohenau, A., Krenn, J. R., Nano Lett. 2012, 12, 661–665. [11] Okamoto, T., Yamaguchi, I., J. Microsc. 2001, 202, 100–103. [12] Wang, J. J., Saito, Y., Batchelder, D. N., Kirkham, J., Robinson, C., Smith, D. A., Appl. Phys. Lett. 2005, 86, 263111.
www.electrophoresis-journal.com
1148
C. Leiterer et al.
[13] Kalkbrenner, T., Ramstein, M., Mlynek, J., Sandoghdar, V., J. Microsc. 2001, 202, 72–76. ¨ G., Geiger, R., Schreiber, [14] Becker, M., Sivakov, V., Andra, J., Hoffmann, S., Michler, J., Milenin, A. P., Werner, P., Christiansen, S. H., Nano Lett. 2007, 7, 75–80. [15] Min, Q., Santos, M. J. L., Girotto, E. M., Brolo, A. G., Gordon, R., J. Phys. Chem. C 2008, 112, 15098–15101. [16] Pohl, H., Crane, J., Biophys. J. 1971, 11, 711–727.
Electrophoresis 2015, 36, 1142–1148
[27] Pohl, H. A., J. Appl. Phys. 1951, 22, 869–871. [28] Pohl, H. A., Dielectrophoresis: The Behavior of Neutral Matter in Nonuniform Electric Fields., Cambridge University Press, Cambridge, UK, 1978. [29] You, Y., Purnawirman, N. A., Hu, H., Kasim, J., Yang, H., Du, C., Yu, T., Shen, Z., J. Raman Spectrosc. 2010, 41, 1156–1162.
[17] Washizu, M., Suzuki, S., Kurosawa, O., Nishizaka, T., Shinohara, T., IEEE Trans. Ind. Appl. 1994, 30, 835–843.
[30] Jose, J., Kress, S., Barik, A., Otto, L. M., Shaver, J., Johnson, T. W., Lapin, Z. J., Bharadwaj, P., Novotny, L., Oh, S.-H., ACS Photonics 2014, 1, 464–470.
[18] Raychaudhuri, S., Dayeh, S. A., Wang, D., Yu, E. T., Nano Lett. 2009, 9, 2260–2266.
[31] Rasmussen, A., Deckert, V., J. Raman Spectrosc. 2006, 37, 311–317.
[19] Wolff, A., Leiterer, C., Csaki, A., Fritzsche, W., Front. Biosci. 2008, 13, 6834–6840.
[32] Deckert-Gaudig, T., Deckert, V., Small 2009, 5, 432–436.
[20] Pethig, R., Biomicrofluidics 2010, 4, 022811. [21] Barsotti, R. J., Vahey, M. D., Wartena, R., Chiang, Y.-M., Voldman, J., Stellacci, F., Small 2007, 3, 488–499. [22] Kumar, S., Yoon, S.-H., Kim, G.-H., Curr. Appl. Phys. 2009, 9, 101–103. [23] Bernard, L., Calame, M., Molen, S. J. van der, Liao, J., ¨ Schonenberger, C., Nanotechnology 2007, 18, 235202. [24] Zoy, A., Nassiopoulos, A. A., Nassiopoulou, A. G., Nanotechnology 2007, 18, 345608.
[33] Mappes, T., Jahr, N., Csaki, A., Vogler, N., Popp, J., Fritzsche, W., Angew. Chem. Int. Edit. 2012, 51, 11208–11212. [34] Leiterer, C., Broenstrup, G., Jahr, N., Urban, M., Arnold, C., Christiansen, S., Fritzsche, W., J Nanopart. Res. 2013, 15, 1–7. [35] Pettersson, L. A. A., Snyder, P. G., Thin Solid Films 1995, 270, 69–72. [36] Bennett, J. M., Stanford, J. L., Ashley, E. J., J. Optic. Soc. Amer. 1970, 60, 224.
[25] Kretschmer, R., Fritzsche, W., Langmuir 2004, 20, 11797–11801.
[37] Schrojenstein Lantman, E. M. van, Deckert-Gaudig, T., Mank, A. J. G., Deckert, V., Weckhuysen, B. M., Nature Nanotechnol. 2012, 7, 583–586.
[26] Leiterer, C., Berg, S., Eskelinen, A.-P., Csaki, A., Urban, ¨ a, ¨ P., Fritzsche, W., Sens. Actuators B-Chem. M., Torm 2013, 176, 368–373.
[38] Kelly, K. L., Coronado, E., Zhao, L. L., Schatz, G. C., J. Phys. Chem. B 2003, 107, 668–677.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
