A new method is demonstrated for optically trapping micron-sized absorbing particles in air and obtaining their single-particle Raman spectra. A 488-nm Gaussian beam from an Argon ion laser is transformed by conical lenses (axicons) and other optics into two counter-propagating hollow beams, which are then focused tightly to form hollow conical beams near the trapping region. The combination of the two coaxial conical beams, with focal points shifted relative to each other along the axis of the beams, generates a low-light-intensity biconical region totally enclosed by the high-intensity light at the surface of the bicone, which is a type of bottle beam. Particles within this region are trapped by the photophoretic forces that push particles toward the low-intensity center of this region. Raman spectra from individual trapped particles made from carbon nanotubes are measured. This trapping technique could lead to the development of an on-line real-time single-particle Raman spectrometer for characterization of absorbing aerosol particles.
©2012 Optical Society of America
Optical trapping and manipulation  of neutral micron-sized particles, nanoparticles, molecules, and atoms are powerful techniques used in aerosol science, chemistry, physics, biology, and interdisciplinary studies [1–20]. Combinations of trapping with other analytical techniques, e.g., Raman spectroscopy of particles in air [4–7] or in liquid [8–11], provide new ways to analyze molecular composition and to study time-varying phenomena, e.g., dynamic reactions in individual living cells [2,3,9–11]. In optical trapping, the main forces on a particle are: radiation pressure (often split into a “scattering” or axial component and a radial or “gradient” component); photophoretic (resulting from asymmetric heating of the fluid surrounding the particle); gravitational; drag; buoyancy; and electrostatic force. For particles in open air without an applied electronic field, the buoyancy and electrostatic forces (even on charged particles) are negligible, but the drag force can be particularly problematical because the viscosity of air is low and the air flows are turbulent. For the case of “suspending relatively transparent particles in relatively transparent media” , the gradient force can be stronger than the photophoretic force, and it attracts the particle towards to the strongest electric field, where the laser beam is focused. One type of radiation pressure trap for this non-absorbing case, commonly referred to as optical tweezers, uses a tightly focused beam to hold particles. Other trapping methods for this case have been reported [12–15], e.g., using longitudinal trapping , holographic methods , self-reconstructing beams , and axicon(s) .
For absorbing particles the photophoretic force can be orders of magnitude larger than the radiation pressure force [16–19]. Photophoretic forces push particles away from the high-intensity region, and so different light-intensity configurations are required for trapping absorbing particles stably [16–19]. A vertically oriented TEM01● (doughnut) beam generating a force that is counterbalanced gravity , several methods using vortex beams [17–19], and bottle beams generated by moiré methods , have all been used to demonstrate trapping and manipulating of single or multiple absorbing particles in air. For example, two co-rotating counter-propagating vortex beams, with their beam waists shifted relative to each other along the axis of the beams, generate a region of low intensity near the axis and away from the two beam waists that is surrounded in any direction by a high-intensity surface of light. Whenever a trapped particle moves in any direction away from the low-intensity region near the center of the trap, it eventually encounters the high intensity light, and the photophoretic force pushes the particle back toward the center of the trapping volume. The particle eventually reaches a stable position by balancing all the forces [17–19]. Carbon nano-foam or agglomerates of carbon nanoparticles were chosen by Shvedov et al [17–19] as nearly ideal samples for photophoretic trapping because of their low-bulk density (2−20 mg/cm3) and very low-thermal conductivity (0.0266 W/ mK). The low thermal conductivity helps the carbon nano-foam agglomerate particles to maintain a highly asymmetric surface temperature, which allows them to be trapped using laser powers as low as 0.3 mW ).
Raman spectra of particles trapped using optical tweezers, including spectra of optically trapped individual biological cells (blood, yeast and bacteria) in water, have been reported [8–11]. Acquisition times of 20 seconds were used to obtain Raman spectra of individual trapped Bacillus thuringiensis spores excited by a 785-nm, 30-mW beam focused to a spot of 1 to 2 μm diameter (roughly 1.5 × 106 W/cm2). Raman spectra of particles trapped using photophoretic forces have not been reported, so far as we know. Raman spectra of samples airborne particles collected onto surfaces have been shown to be useful for bioaerosol pathogen characterization .
In this letter, a new optical arrangement is described for trapping absorbing particles in air stably for several hours. Axicons and other optical components form two counter-propagating circular hollow laser beams from one 488-nm Gussian Ar-ion-laser beam. The two beams are focused to form hollow conical beams that overlap to produce a biconical region of low intensity surrounded in any direction by the high intensity regions of the hollow beams, i.e., a type of bottle beam. This trap is stable for absorbing particles. With this trap, Raman spectra can be readily measured from individual micron-sized clusters of multi-walled carbon nanotubes (MWCNTs) in air. To the best of our knowledge, this is the first report of measurements of single-particle Raman spectra of a trapped absorbing particle in air.
2. Experimental arrangement
Figure 1 shows the schematic of the experimental setup. The light source is a continuous-wave (cw) argon ion laser at 488 nm (Coherent Inc., Innova 300C FreD). It produces a linearly polarized Gaussian beam (TEM00). Its output power can be controlled by adjusting the plasma tube current. The laser beam power also can be adjusted using a neutral density filter. The laser beam is cleaned and re-collimated using a 2-mm-diameter diaphragm, two lenses (L1, f = 25 mm; L2, f = 50 mm), and a 400 μm diameter pinhole. The Gaussian beam is transformed into a hollow beam by passing it through three axicons (Thorlabs, cone angle 176°) and a spatial filter. The spatial filter is an optical window with a 5-mm diameter mirror at the center for blocking the remaining light along the optical axis. The first half-wave plate (WP) and polarizing beam splitter (PBS1 in Fig. 1) are used for fine adjustment of the laser intensity. The second combination of WP and PBS2 is used to separate the laser beam into two equal intensity beams propagating perpendicularly to each other, where one is polarized horizontally and the other polarized vertically. The different polarizations for these two beams are not required for generating the desired optical intensity distribution for trapping the particles, but are designed to make the two equal-intensity beams travel around the triangular path only once, then exit at PBS2, in order to minimize unwanted multiple reflections and interference patterns. The two beams are further made to overlap as they propagate in opposite directions, and (initially) to focus to the same point (equidistant from PBS2) by microscope objectives MO1 and MO2 (Nikon, ELWD 20 × , N.A. = 0.4). Each MO is mounted on a 2D tiltable frame supported by a 3D translation stage. By moving MO1 and MO2 towards each other at least 10 μm (50 μm for the measurements shown here) along the optical axis longitudinally, a hollow biconical region with two closed ends is formed for particle trapping (see the inset at the center of Fig. 1).
An additional cw diode laser (DL) at 635 nm is used to illuminate the trapped particle for alignment. MO3 (Creative Device, f = 20 mm, 50 × , N.A. = 0.42) collects both the Raman scattering signal and the elastic scattering from the trapped particle. The dichroic beam splitter (DBS, Semrock LPD01-488RU-25) reflects the 488-nm light and transmits the 494.3- to 756.4-nm band (>95%) for a 45° incident beam. Another long pass Raman edge filter (LP, Semrock LP02-488RE-25) further blocks the residue of elastic scattering from the laser with high transmission in the 491.2-1110.8 nm band (>95%). The scattering image of the trapped particle is monitored by a CCD camera (Pulnix TM-9701) for the 488-nm scattering, and by an Electron Multiplying CCD (EM-CCD, Princeton, ProEM 1600 × 200), operating in imaging mode, for the 635-nm scattering. The Raman signal is dispersed by a spectrograph (Acton, SP2300, grating 1200/mm, blaze 500 nm), and recorded by the EMCCD in spectroscopy mode.
3.1 Measurements of the hollow trapping beams
To verify that the conical beams have the desired intensity distributions transversely and longitudinally, the beam profiles in the vicinity of the trapping region were measured. Figure 2 shows one of the conical beams at three different locations. These images were obtained by directly exposing the beam to a CCD when the laser was at a very low power. These images indicate that the beam can be focused to a spot, with full-width at half-maximum (FWHM) around 50 μm (see top right). This spot provides the closed end along the axis of the biconical region. The beam profile exhibits relatively little azimuthal variation in intensity (angular variation around the axis). The thickness of the ring is about 100 μm FWHM around the ring. The intensity near the axis, away from the focal region, is very low (see bottom part). This ring is a slice through the high-intensity surface that forms the hollow biconical region. Originally, the hollow beams have a series of concentric rings, with rings further from the axis having lower intensities. Here the beam is spatially filtered to allow only the strongest ring to enter the aperture of the MOs. The total length of the round trip in the triangular trapping path is adjusted so that the weaker rings that are further from the optical axis are blocked by the apertures of MOs. This filtering of the beam is designed to reduce unnecessary interferences also.
3.2 Trapping of particles: smut spores, riboflavin, carbon nanotubes, nigrosin and carbon black
Using the system shown in Fig. 1, individual micron-sized particles or aggregates of particles of Johnson grass smut spores (an example fungal spore), riboflavin, and powders of carbon black, nigrosin, and carbon nanotubes were trapped in the open air for times up to hours. The particles were aerosolized by placing them on a microscope slide set under the trapping region, and blowing them into the air either with turbulent air flows from an ultrasonic needle, or with a tiny puff from a can of compressed air. Once the aerosolized particles enter into the trapping region, many particles fly towards the trap center and compete for a stable position. Typically only one trapped particle or agglomerate remains after a short time (seconds). Occasionally two or three particles are trapped near, but not touching, each other. We think the most likely explanation is that each of these repelling particles has a residual charge of the same sign, which causes them to repel each other and remain separated. If two particles having charges with opposite sign were trapped, they would probably agglomerate with each other. The elastic scattering light from these particles is strong enough to be visible to the naked eye. Figure 3 (Top) shows a single particle trapped in air by photophoretic forces, as captured by a video camera. Figure 3 (Bottom) shows the images recorded by the CCD for a 100-μm diameter fiber used for aligning the system (left panel), and the three trapped particles (with diameters around 5 μm, 20 μm, 35 μm, respectively) in the panels to the right. Smaller (~2 μm) and larger (~100 μm) particles have also been trapped.
We found that solid particles with a much higher density and thermal conductivity than the carbon nano-foam or agglomerates trapped and studied by Shvedov et al [17–19] can be readily captured and steadily trapped in air for long times (up to hrs) with sufficient laser power (30 mW or above). The thermal conductivity for carbon black is 6~174 W/mK , which is at least 100 to 1000 times higher than the carbon nano-foam. Agglomerates of carbon nanotubes, which have an even higher thermal conductivity (2000~6000 W/mK ), were also readily trapped. No thermal conductivity data was found for nigrosin, Johnson grass smut spores, or riboflavin. Higher laser powers (30 mW or higher) were required to capture and trap particles of riboflavin and Johnson grass smut spores. The absorption at 488 nm is probably lower for the smut spores and riboflavin than for the samples that are mostly elemental carbon.
The higher the laser power, the more easily particles can be trapped. The highest power used was 100 mW. Once a particle is trapped, it can remain trapped while the laser power is gradually decreased to some lower limit. The lowest power that was found to be sufficient to hold a carbon black particle for seconds was 1 mW (after it was initially captured with 20 mW).
Usually there is no noticeable movement or rotation of the trapped particle, as monitored by the CCD. Occasionally, a trapped particle leaves the trap. Other times, a trapped particle may drift away from the trapping position, and then bounce back quickly from the beam at the surface of the bicone, and stop where it was trapped previously. These movements often appear correlated with motion of persons in the room, which may increase the turbulent air currents. The drag force on small particles in air can be large, easily exceeding the gravitational force. This sensitivity to movement of persons suggests that particles probably could be trapped with even lower laser powers for long times in an air-tight chamber.
3.3 Measurements of the Raman spectra of trapped particles
Figure 4 shows the Raman scattering signal from a single trapped aggregate (~20 μm diameter) of multi-walled carbon nanotubes (MWCNTs) measured using different data acquisition conditions. The MWCNTs have an outside diameter less than 7 nm, an inside diameter approximately 2-5 nm, and a length approximately 10-30 μm (US Research Nanomaterials, Inc). The Raman signal is excited by the 488-nm beam that also traps the particle.
In Fig. 4(a) has the largest signal/noise (S/N) (30 s; gain × time = 60); 4(b) has the next largest S/N (10 s; gain × time = 50); and 4(c) has the smallest S/N (0.5 s; gain × time = 50). Even though the width of the entrance slit for the spectrograph in 4(b) is 100 μm, i.e., 1/5th of the 500 μm used in 4(a) and 4(c), the S/N in 4(b) is larger than in 4(c). The narrow slit width in (b) also improves the spectral resolution.
4.1 How can particles overcome the photophoretic energy barrier and enter the trap?
The intersection of the two hollow trapping beams forms a closed high-intensity axi-symmetric surface. Particles remain trapped because gravitational drag, and other forces are not sufficient to exceed the photophoretic trapping forces. How then do these particles have the energy to cross the photophoretic barrier and enter the trap?
The particles were aerosolized by turbulent airflows from an ultrasonic needle or from a can of compressed air. These rapid airflows initially generate high drag forces on the aerosol particles, which accelerate the particles. The combination of high momentum and high drag forces could exceed the barrier posed by the trap. Once the generation of these turbulent airflows stops, the airflow and particle velocities decrease and can change direction, as do the drag forces on a particle. Therefore, the velocities and drag forces that were able to push a particle into the trapping region may change direction and/or decrease in magnitude sufficiently that they are not able to push the particle out of the trap.
4.2 Possible approach to trapping, measuring and releasing for a continuously operating aerosol sampling system
The strong and stable trapping demonstrated here, along with the ability to measure Raman spectra of the trapped particles, suggests that the method might be developed into a continuous sampling system for measuring Raman spectra of absorbing atmospheric particles, similar to other continuous sampling single-particle characterization methods, e.g., those based on laser-induced-fluorescence (LIF)  or mass spectrometry (MS). However, in contrast to those online techniques, where the spectrum is recorded within a few μs or less, each 1-μm diameter particle would likely need to be held briefly for 100’s of ms to many seconds (depending upon the material and excitation wavelength) while its Raman spectrum is measured. Such a system would still be termed “continuous” or “online”, but the peak sample rate would be lower than the MS or LIF systems.
The envisioned sampling system would operate in a trap-measure-release mode. The particles to be sampled would be drawn into the trapping region in an airflow (ideally laminar) with a steady flow rate. The trapping laser beam would be unblocked (or gated-on) to trap the target particle. The Raman spectrum of the trapped particle would be measured. After its Raman measurement was completed, the particle would be released by blocking (or gating-off) the trapping laser beam. The laser beam then would be unblocked again to trap the next particle and the process would be repeated. To detect a particle approaching the trap and indicate when to unblock the laser beam, the elastic scattering from particles immediately upstream from the trapping region could be used, as is common in single-aerosol-particle LIF sampling systems. The combination of two different-wavelength crossed-beam diode lasers and associated photodetectors with bandpass filters detects the arriving particle.
One advantage of such a photophoretic trap-and-release system for sampling atmospheric aerosols is that once the trap is “on,” other particles that are sufficiently absorbing to be trapped, at the steady airflow rate through the trap, are prevented from entering the trap. This feature provides a way to make a trap-measure-release system that measures one particle at a time, without other particles entering the trap during the acquisition of the spectrum of the particle. On the other hand, in tweezers traps other particles in the airstream may be added continually to a trapped particle as its spectrum is being measured.
A potentially useful feature of such a trapping system is that if any feature of the particle (e.g., the Raman emission at some wavelength) meets some predefined criteria, the particle can be held longer for additional measurements, e.g., for measuring the Raman spectrum again with longer acquisitions time to improve S/N.
4.3 Illumination intensities, acquisition times, and potential for heating the particles
Raman spectra from individual bacterial spores or cells, which are held using a tweezers trap, can be measured with adequate S/N when the particle is illuminated with intensities in the range of 1 MW/cm2 for times of 16 to 60 seconds [8,9]. These times and large intensities are required because of: i) the extremely small cross sections for Raman scattering from bacteria excited in the near-IR; ii) the spread of the Raman emission over a broad range of vibrational frequencies, resulting from the many different molecular species in biological cells; iii) the desire to minimize photodamage; and, iv) the small sizes of the single bacteria studied (approximately 0.8 to 2 μm diameter for equivalent volume spheres).
The photophoretic trapping system and measuring system for absorbing particles, described in this paper, differs in several ways as follows.
- a) The illumination intensity on the particle is estimated to be roughly 104 times smaller than that used for Raman measurements of bacteria trapped with laser tweezers [8,9]. The intensity illuminating the particles in Fig. 4 is estimated to be roughly 100 W/cm2 (when the laser power was 30 mW). The intensity was kept low partly because of our interest in a system that does not require high power. When the laser power was increased to 100 mW, the trap was more able to catch less absorbing particles and would generate stronger Raman signals. The laser intensity was also kept low partly to avoid overheating the particles, although we did not notice evidence of overheating. Using the analysis of Ref , we estimate an increase in equilibrium temperature of 84 C for a cw 30-mW beam illuminating a solid 20-μm diameter particle that absorbs all the light impinging directly on it (i.e., for an absorption efficiency of 1.0). A typical diameter of the atmospheric particles we desire to measure eventually is around 2 μm. We estimate the increase in equilibrium temperature for a highly absorbing 2 μm particle to be 8 C. Increasing the illumination intensity on the particle could reduce the acquisition times required for Raman spectra. This increase could be achieved by increasing the laser power. It could also be achieved by adjusting the optics to increase the intensity at the center of the beam, so long as the trapping intensity forming the biconical trap was also increased. Efficient trapping does not require that the intensity at the center of the trap be very small. Trapping requires that intensity barrier preventing the particles from leaving the trap be sufficiently high above the intensity (whatever it is) at the intensity minimum of the trapping region.
- b) Because the laser wavelength must be absorbed well by at least one of the species of molecules in the particle, it is likely that at least one of the species will exhibit a resonantly enhanced Raman cross section. Resonance Raman cross sections can be orders of magnitude larger than those measured far from resonance. In laser-tweezers trapping the excitation wavelength for the Raman emission cannot be near a resonance of a molecule that contributes a significant fraction of the mass of the particle. If it did, the strong photophoretic forces would push the particle away from trap. Also note that in Fig. 4 the diameters of the MWCNT particles have diameters roughly 20 times larger than the bacterial spores measured in the near-IR [8,9]. The absorption by these 20-μm diameter MWCNT particles would be roughly 400 times larger than a 1 μm MWCNT particle. Although Raman cross sections of selected individual carbon nanotubes, excited near resonance with an ideal orientation, can be as large as 10−22 cm2/sr , it is not clear how such this number applies to the case studied here, where there is a large dispersion in nanotube diameters, orientations, and numbers of walls.
- c) The Raman signal will be generated near the surface of the large MWCNT particles studied here, because the illuminating light is highly absorbed near the surface, and cannot penetrate far into the particle. In the transparent bacterial particles studied with Raman excited by wavelengths in the near IR [8–11] molecules throughout the volume should be able to contribute to the signal.
- d) A high fraction of the Raman emitted by a nanotube (but not necessarily by the particle composed of many nanotubes) will be absorbed before it is able to exit the particle. This absorption should occur because the MWCNT material is also highly absorbing at the Raman emission wavelengths. This reabsorption of emission would not be significant in the case of the low-absorbing particles studied with laser tweezers in the near IR [8,9]. This reabsorption of emission would not be the case for materials where the absorption of the excitation wavelength is high (needed for photophoretic trapping with excitation at the trapping wavelength), but where the absorption of light at the wavelengths of the Raman emission is relatively low.
4.4 Avoiding fluorescence and selecting an excitation wavelength
Fluorescence cross sections are typically far larger (e.g., 104 × ) than the Raman cross sections for bioaerosols and for many polycyclic aromatic hydrocarbons excited at UV and visible wavelengths. For most biological aerosols, the fluorescence can swamp the Raman signal. One way to circumvent this problem is to first photobleach the fluorescence, and then measure the Raman spectrum. This method is employed in an instrument that uses 532-nm light to excite Raman spectra of particles collected on a substrate [21,26]. Although this approach works well for that application, the required photobleaching times are longer than 5 minutes . However, it is not suitable for a continuously sampling trap-and-measure single-particle Raman spectrometer.
Another approach to reducing fluorescence-Raman overlap is to illuminate the particles at wavelengths shorter than about 245 nm. At these wavelengths the fluorescence from biological materials is shifted sufficiently far from the laser wavelength, and the Raman emission is sufficiently close to the laser wavelength, that the Raman spectra are measurable without fluorescence interference and without photobleaching.
Other benefits of exciting the Raman emission at wavelengths below 245 nm are: a) the Raman cross sections increase by (1/λ)4, (where λ is the excitation wavelength), and that is a factor of 81 increase for decreasing the excitation wavelength from 732 nm to 244 nm; b) many of the biological molecules of interest absorb strongly at these wavelengths and the Raman emission can be resonantly enhanced; and, c) the particles should be easier to trap photophoretically because they are more absorbing.
4.5 Trapping and exciting Raman emission with different wavelength lasers could afford increased flexibility
In the experiments in this paper, one laser is used to both trap the particle and generate the Raman emission. Non-absorbing particles would not be trapped in the “trapping region.” Increased flexibility in trapping and measuring could be afforded by using two lasers. The 635-nm diode laser in Fig. 1, noted as DL, which in the experiments here was not used to excite Raman or to trap particles, could be replaced at that position by a second laser of any wavelength, including the wavelength of the trapping laser. The second laser could be focused tightly or weakly depending on the application. Two of the possible uses of the two-laser system are as follows.
- a) It could be used for trapping both absorbing particles and non-absorbing particles. Absorbing particles would be trapped using photophoretic forces as described above. Particles with little or no absorption would be trapped using the radiation pressure forces generated by tightly focusing the second laser beam to form a laser-tweezers trap. The wavelength of the second laser beam would be chosen to be one that was not absorbed well by the particle.
- b) It could be used to controllably excite the Raman emission at whatever wavelength is desired. The first laser would be used for trapping as described above. If the first laser, with a wavelength chosen to trap the desired absorbing particles, generates too much fluorescence in the Raman emission region, the second laser may be chosen to avoid fluorescence interference, e.g., at wavelengths longer than 700 nm where the fluorescence interference is much weaker. If the second laser is at a wavelength that is not absorbed by the particles of interest, it could be focused tightly to the center of the trapping region and could also contribute to trapping the particles. If the second beam were at a wavelength that is absorbed by the particles of interest, its intensity could be kept low enough that it does not push particles out of the trap. Once a particle was trapped, the intensity of the second beam could be increased so that the Raman spectrum could be measured more rapidly. One application of such an approach could be to use relatively inexpensive light, e.g., at 400 nm or longer, to trap the particles, but use the expensive 244-nm light to excite the Raman emission in a region where the fluorescence problem can be avoided and where the Raman cross sections are relatively large. Another application might be to generate both the trapping and the Raman-excitation beam from one laser (e.g., at 244 nm), using a beamsplitter and a way to control the amplitude of the second beam (the Raman-excitation beam).
This research was supported by the U.S. Army Research Laboratory (ARL) Director’s Research Initiative (DRI) project (FY10 CIS-05).
References and links
1. P. A. Ashkin, “Acceleration and trapping of particles by radiation pressure,” Phys. Rev. Lett. 24(4), 156–159 (1970). [CrossRef]
4. R. E. Preston, T. R. Lettieri, and H. G. Semerjian, “Characterization of single leviatated droplets by Raman spectroscopy,” Langmuir 1(3), 365–367 (1985). [CrossRef]
6. A. Biswas, H. Latifi, R. L. Armstrong, and R. G. Pinnick, “Double-resonance stimulated Raman scattering from optically levitated glycerol droplets,” Phys. Rev. A 40(12), 7413–7416 (1989). [CrossRef] [PubMed]
7. J. B. Wills, K. J. Knox, and J. P. Reid, “Optical control and characterization of aerosol,” Chem. Phys. Lett. 481(4-6), 153–165 (2009). [CrossRef]
8. C. Xie and Y. Q. Li, “Confocal micro-Raman spectroscopy of single biological cells using optical trapping and shifted excitation difference techniques,” J. Appl. Phys. 93(5), 2982–2986 (2003). [CrossRef]
9. D. Chen, S. S. Huang, and Y. Q. Li, “Real-time detection of kinetic germination and heterogeneity of single Bacillus spores by laser tweezers Raman spectroscopy,” Anal. Chem. 78(19), 6936–6941 (2006). [CrossRef] [PubMed]
10. P. F. Zhang, L. B. Kong, P. Setlow, and Y. Q. Li, “Multiple-trap laser tweezers Raman spectroscopy for simultaneous monitoring of the biological dynamics of multiple individual cells,” Opt. Lett. 35(20), 3321–3323 (2010). [CrossRef] [PubMed]
11. L. B. Kong, P. F. Zhang, G. W. Wang, J. Yu, P. Setlow, and Y. Q. Li, “Characterization of bacterial spore germination using phase-contrast and fluorescence microscopy, Raman spectroscopy and optical tweezers,” Nat. Protoc. 6(5), 625–639 (2011). [CrossRef] [PubMed]
14. V. Garcés-Chávez, D. McGloin, H. Melville, W. Sibbett, and K. Dholakia, “Simultaneous micromanipulation in multiple planes using a self-reconstructing light beam,” Nature 419(6903), 145–147 (2002). [CrossRef] [PubMed]
15. B. Shao, S. C. Esener, J. M. Nascimento, M. W. Berns, E. L. Botvinick, and M. Ozkan, “Size tunable three-dimensional annular laser trap based on axicons,” Opt. Lett. 31(22), 3375–3377 (2006). [CrossRef] [PubMed]
16. M. Lewittes, S. Arnold, and G. Oster, “Radiometric levitation of micron sized spheres,” Appl. Phys. Lett. 40(6), 455–457 (1982). [CrossRef]
18. A. S. Desyatnikov, V. G. Shvedov, A. V. Rode, W. Krolikowski, and Y. S. Kivshar, “Photophoretic manipulation of absorbing aerosol particles with vortex beams: theory versus experiment,” Opt. Express 17(10), 8201–8211 (2009). [CrossRef] [PubMed]
19. V. G. Shvedov, C. Hnatovsky, A. V. Rode, and W. Krolikowski, “Robust trapping and manipulation of airborne particles with a bottle beam,” Opt. Express 19(18), 17350–17356 (2011). [CrossRef] [PubMed]
20. P. Zhang, Z. Zhang, J. Prakash, S. Huang, D. Hernandez, M. Salazar, D. N. Christodoulides, and Z. Chen, “Trapping and transporting aerosols with a single optical bottle beam generated by moiré techniques,” Opt. Lett. 36(8), 1491–1493 (2011). [CrossRef] [PubMed]
21. K. S. Kalasinsky, T. Hadfield, A. A. Shea, V. F. Kalasinsky, M. P. Nelson, J. Neiss, A. J. Drauch, G. S. Vanni, and P. J. Treado, “Raman chemical imaging spectroscopy reagentless detection and identification of pathogens: signature development and evaluation,” Anal. Chem. 79(7), 2658–2673 (2007). [CrossRef] [PubMed]
22. Z. Han and A. Fina, “Thermal conductivity of carbon nanotubes and their polymer nanocomposites: A review,” Prog. Polym. Sci. 36(7), 914–944 (2011). [CrossRef]
23. Y. L. Pan, R. G. Pinnick, S. C. Hill, J. M. Rosen, and R. K. Chang, “Single-particle laser-induced fluorescence spectra of biological and other organic-carbon aerosols in the atmosphere: measurements at New Haven, Connecticut, and Las Cruces, New Mexico,” J. Geophys. Res. 112(D24), D24S19 (2007). [CrossRef]
24. J. R. Finke, C. L. Jeffrey, and R. E. Spjut, “Measurement of the emissivity of small particles at elevated temperatures,” Opt. Eng. 27, 684–690 (1988).
25. J. E. Bohn, P. G. Etchegoin, E. C. Le Ru, R. Xiang, S. Chiashi, and S. Maruyama, “Estimating the Raman cross sections of single carbon nanotubes,” ACS Nano 4(6), 3466–3470 (2010). [CrossRef] [PubMed]
26. J. Guicheteau, S. Christesen, D. Emge, and A. Tripathi, “Bacterial mixture identification using Raman and surface-enhanced Raman chemical imaging,” J. Raman Spectrosc. 41(12), 1632–1637 (2010). [CrossRef]