Optical guiding of microscopic particles is studied and compared for femtosecond and continuous wave Bessel light beams. We confirm that optical guiding is an average power effect and observe no difference in the guiding velocities for non-fluorescing polymer microspheres. Additionally, we observe second harmonic generation of guided KTP particles. This observation opens up the prospect for optical identification of guided cells for sorting purposes.
©2004 Optical Society of America
All optical transport of microscopic particles was first reported by Ashkin in 1970  when he demonstrated that the radiation pressure of light combined with refraction through transparent media were sufficient to manoeuvre and guide microscopic objects. In the Mie regime, we can consider the refraction of light through the particle as localising it in two dimensions orthogonal to the beam axis with the backscattered component creating a radiation pressure that propels the particle along the beam axis. This work was a precursor to the technique of optical tweezing which uses a single laser beam focused by a high NA microscope objective to trap a particle in three dimensions . Both optical guiding and optical tweezing are topics of particular current interest which are further enriched with advanced applications in biology and colloidal physics. Some of the excitement has been fuelled by the ability of novel light patterns and hologram generated arrays to enable new scientific research with optical traps. In this context, Bessel beams have been shown to offer the capacity for extended guiding of particles that exceed the guiding distance of Gaussian beams. . They can be described as propagation invariant solutions of the scalar Helmholtz equation and they manifest themselves as a set of concentric rings . Additionally, their reconstruction ability is of use in creating multiply aligned traps .
To the best of our knowledge all previous work in optical guiding has been performed with continuous wave light beams. However, there has been impressive progress in developing compact ultrashort pulse lasers that operate in the femtosecond regime. They now offer average cw powers in the region 100mW-1W with correspondingly high peak powers. Important issues arise as to whether it is the peak power of the guide light or the average power that dominates the dynamics of particle transport. As a potential benefit, ultrashort pulse lasers can, by the process of frequency conversion, give access to regions of the spectrum that are difficult to obtain using cw sources. Perhaps a more compelling reason for using ultrashort pulse lasers lies in the potential to exploit simultaneous guiding and two-photon studies on the same particle. Such nonlinear processes are dramatically enhanced in the presence of high intensity ultrashort light pulses. Notably second harmonic generation and two-photon fluorescence of biological samples can give key information in suitable samples. Thus, by combining such properties with optical guiding there can be important and novel implications in respect of cell sorting  and diagnosis for example. We note that optical tweezing of Rayleigh particles using pulsed lasers has already been observed .
In this paper we describe the assessment and comparison of the guiding of microscopic particles in cw and femtosecond Bessel light beams. We record particle velocities and observe that it is the average power that dictates the optical guiding process. Finally, we present results relating to the simultaneous second harmonic generation and optical guiding for microscopic fragments of a nonlinear crystal. This offers a potentially exciting future mechanism with which to distinguish and diagnose differences in guided cells or chromosomes, for example.
The experimental set up is illustrated in Fig. 1. The laser source was a Ti:Sapphire laser pumped by a Spectra Physics 2040 argon-ion laser. This laser provided up to 1.2W of output power that was tuned between 790nm and 850nm for these assessments. The pulse duration, measured using an intensity autocorrelation, (Fig. 2) was found to be 110fs at the 800nm centre wavelength. With the bandwidth of 9nm and assuming sech2 pulse profiles, the implied duration-bandwidth product was 0.46. The pulse repetition frequency of the laser in modelocked operation was 80MHz.
The Ti:Sapphire laser could be readily switched between modelocked and continuous wave (cw) operation simply by a transient interruption of the intracavity beam. The change between modelocked and cw operation gave only a small (3%) change in laser performance (spot size, pointing etc.) enabling experiments in the two regimes to be conducted without the need for any time-consuming experimental realignments. Care was also taken to ensure that when comparisons were made between modelocked and cw operation that the centre wavelength of the modelocked pulses corresponded to the cw operating wavelength of the laser.
The TEM00 laser output was passed through a x4 telescope arrangement to illuminate fully the 10 axicon used to generate the Bessel beam. This process produced a Bessel beam with a central beam diameter of 30µm (measured using a video footage analysed with LabView IMAQ software). The beam was found to have a diffraction-free propagation distance of over 25cm. For the guiding experiments, a smaller central beam was required, so a x10 demagnifying telescope was used after the axicon giving a central beam diameter of 3–4µm and a propagation distance of approximately 3mm. The Bessel beam then passed through a 2mm×2mm×20mm cuvette (length×width×height). This cuvette contained 1 micron diameter polymer spheres (Duke Scientific 4000 series, n=1.56 at 589nm) in a distilled water suspension. It was also confirmed that the propagating laser beam did not suffer any distortion in transiting this medium .
In practice, the absorption coefficient of water in the wavelength range investigated resulted in strong heating effects that adversely affected the guiding behaviour of the spheres. The water was therefore replaced with a 10:1 mixture of water and D2O which had a reduced absorption coefficient around 800nm. The side of the cuvette was imaged onto the CCD array of a video camera using a microscope objective and the output viewed on a TV monitor and recorded on VHS video. Using this arrangement, a significant length of the propagating laser beam in the cuvette could be imaged. Subsequent analysis undertaken on a frame-by-frame basis then enabled the velocity of the guided particles to be deduced.
Data were taken for 20 microspheres within the sample. The guiding velocities obtained are plotted in Fig. 3 for 1µm spheres for a variety of input powers from the Ti:Sapphire laser. It can be seen that at the power levels investigated, the average guiding velocity does not change significantly for the operational regime change from cw to modelocked.
The duration and bandwidth of the modelocked laser pulses were also monitored following their exit from the axicon. The pulses were observed to have broadened from 95fs to 130fs, while the bandwidth remained unchanged at 9nm. [Measurement of the phase of the pulse was beyond the scope of this experiment but this will be considered in subsequent evaluations]. For further comparison with a Gaussian beam, the axicon could be removed and replaced with a 5cm focal length lens to facilitate Gaussian guiding of the particles. This lens produced a 3µm spot size with a confocal parameter of 70µm. The beam was focused through the cuvette and data analysed in a similar manner as to the Bessel beam. A measurement of guiding velocities at 150mW showed that the cw average velocity (18.5µm/s standard deviation 1.8) did not differ significantly from the modelocked case (17.6µm/s standard deviation 1.4). The Gaussian guiding measurements were taken at lower input power values than for the Bessel guided particles in recognition that the power quoted for the Bessel beam applies to the entire series of rings. [The power in the central spot is the total power divided by the number of rings.] The assessment for this Gaussian-beam guiding was repeated with larger spot sizes and confocal parameters for a range of polymer sphere diameters. In all instances of Gaussian guiding no significant difference in guiding velocity between the case for cw and mode-locked case was observed.
We also investigated the difference in heating in relation to the modelocked and cw radiation. This was analysed by examining the behaviour and speed of polymer spheres caught in convective currents in the cuvette. In this case, the spheres could be clearly observed moving upwards in the cuvette following some initial settling. With the cuvette containing water only (no D2O), the average speed of the upwards-travelling spheres was measured for both cw and modelocked pulses. The centre wavelength of the modelocked pulses for this was 819nm so the cw wavelength was adjusted to cover the 800nm to 850nm range. The bandwidth of the modelocked pulses in this case was measured to be 12nm. The average convection velocities for a variety of particle sizes are shown in Fig. 4. From the plot, it can be seen that the heating for the modelocked laser beam is comparable with that of cw operation. The change in convection velocity over the wavelength range is consistent with the rapid change in water absorption .
To show the utility of using femtosecond pulses for optical guiding, a further experimental evaluation was conducted to demonstrate the combination of single-photon (guiding) and two-photon processes arising from the high peak intensity of the ultrashort pulses. The polymer microspheres were replaced with KTiOPO4 (KTP) crystallites. This sample that was produced by grinding up a piece of nonlinear crystal contained an assortment of irregularly sized KTP fragments in the 1–10µm range. In spite of the non-symmetric nature of these fragments, strong guiding was obtained using the Bessel beam. When the aperture of the CCD camera was covered with a filter that transmitted <0.1% around 800nm whilst transmitting strongly in the blue, a clear frequency-doubled signal could be observed from the CCD camera, orthogonal to the Bessel beam propagation direction. If the laser was operated in the cw regime, the guiding continued, but no second harmonic signal could be observed on the CCD camera with the filter in place. This was a clear demonstration of simultaneous guiding and second harmonic light generation by the KTP crystal fragments. As expected, the intensity of second harmonic light generated by the KTP was observed to change as the particle tumbled and changed its orientation within the guide beam. We believe that this intensity variation is caused by the phase matching angle required for efficient non-linear generation in KTP. As the guided particles reached the end of the cuvette cluster of KTP fragments was formed. The image of this cluster was analysed and the second harmonic intensity generated was indirectly measured by considering the pixel intensity on the CCD.
The intensity of second harmonic power generated in a nonlinear optical process, PSH is expected to be a quadratic function of the fundamental power, , so a plot of log(PSH) versus log(PIN) should reveal a straight line gradient of 2. The pixel intensity of the particle agglomeration was then measured as a function of the fundamental power of the guiding beam from the laser. The results obtained are illustrated as a log-log plot. This gives the straight line expected with a calculated gradient of 1.99, in good agreement with the theoretically expected result providing further evidence of two-photon-based nonlinear optical processes taking place. As far as we know, this is the first demonstration of nonlinear frequency conversion in optically guided microscopic objects.
We have performed optical guiding of microscopic particles in the Mie regime using cw and femtosecond-pulse Bessel light beams. Our experimental assessments have confirmed that optical guiding is an average power effect as equivalent particle velocities were observed for both regimes of laser operation. We have also established the effect of wavelength on bulk heating of the medium as an important consideration in guiding and that it is an average power effect with no discernible difference between the cw and femtosecond laser beams. Second harmonic generation has been observed in guided microscopic particles and this provides an exciting prospect for exploiting simultaneously one-photon and two-photon effects.
Two-photon fluorescence is now established as a key tool for diagnosis in biology. By combining this with optical guiding as presented here introduces a practical mechanism for determining differences between guided particles and could be of particular usefulness in cell sorting. As a specific example, two-photon processes can lead to protein expression within the cell of interest that can subsequently be deflected by additional laser sources to a collection chamber. Femtosecond sources are becoming ever more compact and practical and their use for optical guiding may have potential applications in biophotonics and cell sorting.
We thank B. Agate, I. Cormack and B. Stormont for help in the initial part of the experiment. Kishan Dholakia thanks D. McGloin and E.M. Wright for useful discussions and funding support from the UK Engineering and Physical Sciences Research Council and the Scottish Higher Education Funding Council is gratefully acknowledged.
References and links
1. A. Ashkin, “Acceleration and trapping of microscopic particles,” Phys. Rev. Lett. 24, 1570 (1970)
3. J. Arlt, V. Garcés-Chávez, W. Sibbett, and K. Dholakia. “Optical micromanipulation using a Bessel light beam”, Opt. Commun. 197, 239 (2001) [CrossRef]
4. J. Durnin, “Exact solutions for nondiffracting beams .I. The scalar theory,” J. Opt. Soc. Am. A 4, 651–654 (1987) [CrossRef]
5. V. Garcés-Chávez, D. McGloin, H. Melville, W. Sibbett, and K. Dholakia, “Simultanoeus micromanipulation multiple planes using a self reconstructing light beam”, Nature 419, 145 (2002) [CrossRef] [PubMed]
6. T.N. Buican, M.J. Smyth, H.A. Crissman, G.C. Salzman, C.C. Stewar t, and J.C. Martin, “Automated single-cell manipulation and sorting by li ght trapping,” Appl. Opt. 26, 5311–16 (1987). [CrossRef] [PubMed]
7. L. Malmqvist and H.M. Hertz, “Second harmonic generation in optically trapped nonlinear particles with pulsed lasers”, Appl. Opt. 34, 3393 (1995) [CrossRef]
8. Q. Feng, J.V. Moloney, A.C. Newell, E.M. Wright, K. Cook, P.K. Kennedy, D.X. Hammer, B.A. Rockwell, and C.R. Thompson, “Theory and simulation on the threshold of water breakdown induced by focused ultrashort pulses”, IEEE J. Quantum. Electronics , 33, 127 (1997). [CrossRef]