CN100411497C - Method and apparatus for optical separation - Google Patents

Abstract

Static arrays of optical traps can be used to sort microscopic objects with exponential sensitivity to size. Such optical fractionation relies on competition between an externally applied force and the moving objects' differing affinities for optical gradient traps. In a reverse fractionation method, objects that are more strongly influenced by the traps tend to become kinetically locked in to the array and are systematically deflected back into an input flow. In a thermal ratcheting method, patterns are spaced to allow particle diffusion, thus providing the opportunity for forward or reverse movement through the patterns. Unlike other sorting techniques, optical fractionation can operate continuously and can be continuously optimized. The exponential sensitivity arises quite generally from the particle size dependence of the potential wells' apparent widths.

Description

Method and apparatus for optical separation

technical field

The present invention generally relates to embodiments of methods and systems for classifying microscopic particles. More specifically, the present invention is aimed at classifying tiny objects such as macromolecules, biomolecules, nanoclusters, colloidal particles, and biological cells using holographic optical tweezers technology.

Background technique

Optical tweezers use optical gradient forces to trap tiny, usually micron-scale, volumes of matter along two and three dimensions into wells. A holographic form of optical tweezers capable of building a multitude of optical tweezers from a single laser beam using computer-generated diffractive optical units. These tweezers can be arranged in any desired structural form according to actual requirements.

While some systems are known that move particles precisely and with relatively high confidence, typical systems require individually projected holograms for each individual step of particle motion. Computing multiple holograms is time-consuming and requires considerable computational effort. Furthermore, the computer addressable projection systems required to implement the computer-generated optical tweezers described above, or other dynamic optical tweezers systems, such as scanning optical tweezers, appear to be prohibitively expensive.

Contents of the invention

The operation of many technologically and commercially important systems arises through the classical transfer of topography that modulates potential energy. One method of utilizing these operations is optical separation. Optical separation methods enable the sorting of a population of tiny objects, continuously (over a given period of time), into separate components according to their differing abilities to pass through an array of optical traps. Specifically, an object driven by an external force, such as viscous drag in a flowing liquid, encounters an array of optical traps whose axes of symmetry are oriented at an angle relative to the driving force. In general, these wells can be created using holographic optical tweezers techniques. Objects that are more strongly affected by the wells of potential energy created by these wells tend to jump from well to well, thereby deflecting the direction of the driving force. Other objects, more strongly affected by the driving force, or less strongly affected by the optical trap, pass through the array without deflection. Depending on the configuration of the well, the invention can be used to deflect different components by different amounts. In some cases it may be preferable to use the clean separation of the two-component embodiments described above. However, it is also within the scope of the present invention to select multiple components for collection. For example, in one embodiment inhomogeneous samples may be fanned out in a continuously distributed direction range of the "optical chromatography" method. Deflected and undeflected components can be collected separately.

In general, the non-uniform input sample and output components are dispersed in the liquid flowing through the channel. In a preferred embodiment, the channels take the form of a so-called H-shaped cross, where the two outputs, one containing the input mixed sample and the other containing only the background liquid, are separated before being split into two output channels. Bring together to flow side by side for a set distance. If the channel is small enough, the Reynold number of the flowing liquid is small enough that the two fluids do not mix, but flow side by side in a laminar fashion. As a result, objects in the input fluid generally do not cross the boundary between fluids and enter the buffer channel, except occasionally or due to diffusion.

One aspect of the invention involves optical separation using an array of discrete optical traps to sequentially classify microscopic objects based on their relative affinities for optical traps and competition with an applied force. An unwanted component that is more diffuse or mobile than a desired component. However, another aspect of the invention relates to the "inverse" optical separation method. In the opposite optical separation method, the desired component is more diffuse or mobile than the unwanted component.

Another aspect of the invention relates to a modification of the technique known as optical peristalsis, in which tiny objects are deterministically displaced by a sequence of projected light trap patterns. The difference between optical peristalsis and the disclosed optical thermal ratchet technology qualitatively confers new capabilities to the system and method, including unlimited bi-directional pumping through the so-called flow reversal effect, and the ability to sort inhomogeneous samples. new possibilities.

Description of drawings

Figure 1 depicts the optical separation method in which the microfluidic H-shaped intersection includes a first fluid containing a heterogeneous sample to be sorted, and a second fluid consisting of a background or buffer liquid;

Figure 2 depicts the opposite optical separation method, where the microfluidic H-shaped intersection includes two flowing liquids, one containing the heterogeneous sample to be separated and the other containing only the buffer solution;

Figure 3A shows a schematic side view of the optical separation method; Figure 3B shows a top view of the optical separation method of Figure 3A;

Fig. 4 has drawn the optical separation of large quartz spheres from small quartz spheres; Fig. 4A has drawn the representative track of every 1/60 second measurement of 0.79 micron radius sphere; Fig. 4B has drawn the trajectory of 0.5 micron radius sphere obtained simultaneously; Fig. 4C is the time-averaged areal density of 0.79 micron radius spheres relative to their mean areal density; and Figure 4D is the time-averaged areal density of 0.50 micron radius spheres relative to their mean areal density;

Figure 5 shows the spatially resolved quality of separation obtained for a single optical trap line;

Figure 6A depicts a prior art optical peristalsis method in which an optical trap pattern orients an object; Figure 6B illustrates replacing the optical trap pattern with another optical trap pattern displaced by a certain distance; Figure 6C illustrates yet another displaced optical trap pattern; well pattern; finally, Figure 6D depicts the completion of one cycle of optical peristalsis; and

Figure 7 shows numerical solutions to the equations of motion for an optical thermal ratchet embodiment demonstrating flow reversal.

Detailed ways

The present invention relates to methods and apparatus for optical separation. One aspect involves optical separation using an array of discrete optical traps to sequentially sort microscopic objects based on their relative affinities for the optical traps and competition for an applied force. Another aspect of the invention relates to "inverse" optical separation methods. A third aspect of the present invention involves the use of "ratcheting" optical separation techniques.

To study modulated migration, a model system has been developed in which individual colloidal spheres are driven through a regular array of potential wells created with discrete optical tweezers while their motion is analyzed with a digital video microscope. Experiments on this system show that as the array is rotated relative to the driving force, the driven particles trace out a kinematically locked Devil ladder system. In each of these states, the trajectory of the particle follows a symmetrically chosen direction through the lattice of the well, independent of the orientation of the array, and thus is deflected laterally away from the driving force. This deflection can be predicted to provide the basis for a sequential separation technique in which a selected population is deflected by the well array while the rest of the sample passes unimpeded. This method gives a practical demonstration of the optical separation method and also demonstrates that the resolution of the optical separation method can be exponentially dependent on the particle size. Therefore, this method gives a sensitivity unmatched by any previously reported classification technique.

One can demonstrate a conceptual form of optical separation by using an array of discrete optical traps to sequentially sort microscopic objects according to their relative affinities for the optical traps and for competition with an external applied force. This demonstration utilizes the trajectories of two different sized colloidal quartz spheres dispersed in water passing through a linear array of optical tweezers aligned at an angle to the water flow. The colloidal dispersion of the flow was confined to a 4 mm x 0.7 mm x 40 μm glass channel formed by bonding the edge of the cover slip to the microscope slide. The pressure differential applied across the channel produces an approximately constant Poisseuille flow of about 60 μm/sec for several minutes. Samples include spheres with a = 0.79 μm radius (Duke Scientific Duke Scientific Corporation, 2463 Faber Place Palo Alto, California 94303, Lot No. 24169) and spheres with a = 0.5 μm radius (Duke Scientific Lot No. 19057), both of which can be used Conventional bright-field microscopy and digital video analysis, tracking with 30nm precision in the plane every 1/60th of a second. Furthermore, these spheres can be reliably distinguished based on their appearance, providing an ideal model system whose microscopic response to optical separation can be monitored in real time. Typical trajectories of large and small balls are shown in Figures 4A and 4B, respectively.

The quartz spheres are roughly twice as dense as water and thus settle in a single layer just above the glass wall below the channel, with the smaller spheres floating higher because they are lighter. Given the Poisseuille streamlines in the channel, the smaller ball moves slightly faster with an average velocity of u _s =17±9 μm/sec compared to u _b =13±2 μm/sec for the larger ball. The viscous resistance F ₁ =γu of the static ball is characterized by the drag coefficient γ, which is related to the radius a of the ball and the neighborhood of the interface with the ball. The overall drag coefficients can be estimated from their diffusivity D using the Einstein-Smoluchowsky relationship D=k _b T/γ, where K _B T is the thermal energy fraction at temperature T. The diffusivity, in turn, can be measured from the fluctuation of the transverse velocity in the traces such as those shown in Figures 4A and 4B. More generally, the applied force _F1 can be provided by processes such as electrophoresis, electroosmosis, magnetophoresis, or gravitational sedimentation.

The optical trap shown above was created using the dynamic holographic optical tweezers technique. Twelve discrete optical tweezers, each generated by converging a 532 nm laser at 1.7 ± 0.8 mW, were arranged in a line at θ = 12.0° ± 0.5° to the channel axis, with a center-to-center spacing of b = 3.6 ± 0.1 μm. Each well can be roughly simulated by a Gauss potential well, and the depth V _o and width ε _T of the potential well are both related to the radius a of the sphere.

移动。只要外加的力F₁足够大，光阱只把一个粒子偏折，使它离开它的轨迹。如果偏折是小的，那么粒子将继续向下游移动，离开阱的线，且可以说已经从阱的线逃逸。相反，每一阱可以足够强，把粒子偏折进它邻近的影响力范畴。在此情形下，粒子将从一个阱到另一个阱，并有效地被阵列捕获。这就是运动学锁定迁移的机理。选择的偏折角θ，要接近该锁定迁移的最大偏折。捕获粒子的轨迹对逃逸粒子的轨迹的相对偏折，是用光学分离法分类的基础。偏折和不偏折的成分，可以分别地收集，而图1示意画出该过程。If it is not for the optical trap, then when the particle driven by the external force F ₁ passes through the viscous liquid, it will be at the average speed

move. As long as the applied force _F1 is large enough, the light trap will deflect only one particle out of its trajectory. If the deflection is small, the particle will continue to move downstream, off the line of the trap, and can be said to have escaped from the line of the trap. Instead, each well can be strong enough to deflect particles into its neighboring fields of influence. In this case, particles will travel from one well to another and be effectively trapped by the array. This is the mechanism of kinematic lock migration. The deflection angle θ is chosen to be close to the maximum deflection of the locked migration. The relative deflection of the trajectory of the trapped particle to that of the escaping particle is the basis for classification by optical separation. Deflected and undeflected components can be collected separately, and Figure 1 schematically depicts this process.

Given the well geometry, the laser power can be set between the empirically determined large and small sphere escape thresholds. The trajectories in Figures 4A and 4B show that, under these conditions, larger spheres are systematically deflected by the well array, while smaller spheres are not. The result is that the small balls flow unhindered into the shadows created in the distribution of the large balls, where they can be collected. Instead, the deflected large spheres are concentrated into a small area at one end of the optical trap array, where they can be collected separately. Because the purification of the small spheres and the concentration of the large spheres result from the lateral deflection of the larger components, the optical separation process can be performed continuously, thereby outperforming batch-mode techniques such as gel electrophoresis.

为中心的面积区域

中，描绘球的时间平均面密度n，其中对每一总体都以平均的时间平均面密度n_o归一化。球对阱的相对亲和力，可作如下估计：大球在阱中比在整个流体中的可能性粗略地大18倍，而小球的可能性只大3倍。给定球的相对速度，那么这些比值，与较大球临时停在局部的势极小中，而较小球被简单地减速，是一致的。This qualitative explanation of only a small number of trajectories is more convincing by considering the statistics of thousands of trajectories collected in Figure 4C and 4D. Here we are using

centered area

, depicting the time-averaged areal density n of the ball , where each population is normalized by the average time-averaged areal density n _o . The relative affinity of the spheres to the well can be estimated as follows: large spheres are roughly 18 times more likely to be in the well than in the fluid as a whole, while small spheres are only 3 times more likely. Given the relative velocities of the balls, these ratios are consistent with the fact that the larger ball stops temporarily in the local minimum of potential, while the smaller ball is simply decelerated.

The resulting separation quality can be estimated by measuring the relative bulk concentration as a function of position in the fluid:

$\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Figures 5A and 5B show that the figure of merit of the above formula reaches a maximum value of one unit in the region containing only large balls, and is negative one unit in the region containing only small balls. Transecting a segment of the fluid along line A in Figure 5A before the trap array, reveals a fully mixed sample, Q(y) = 0, as indicated by the small circles in Figure 5B. A similar segment along line B after the trap array, drawn as a larger circle in Figure 5B, roughly shows a 40 percent purification for both large and small spheres. Many backgrounds contribute to collisions in the well array, which can cause large balls to escape. Collision-induced escape is evident in the macrosphere concentration profile downstream of the trap array in Figure 4C, with the probability of collision and escape gradually increasing as macrospheres saturate at the downstream end of the trap array. The most effective way to avoid such collisions is to project several parallel well lines. Under the present experimental conditions, as few as three lines can provide a substantially perfect separation, in denser suspensions more lines are required.

The data in Figures 4A, 4B and 5A, 5B demonstrate that a discrete array of optical traps can continuously separate spheres according to their size. Considering the physical conditions that cause one kind of particle to escape from the optical trap array and another to be trapped can give the basis for optimizing the optical separation method.

For simplicity, the analysis only has the effect of two discrete optical traps centered at x=±b/2 with particles near their midpoint x=0. The total potential energy of the particle is

$\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; b</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; b</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

The point through which the particle escapes is

$<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}$

Here, the y-component of the total force is equal to zero. Particles are likely to escape close to x = 0 because the force of trapping is weakest and at y = σ the force of separation is greatest. In this case, the maximum achievable deflection that still allows the capture trajectory is given by

$\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; \&ap;</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 8</span>}{\mathrm{<span\; class="notranslate">\; \&zeta;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Here, the relative well strength $f\left(a\right)=(2/\sqrt{e})V_o/{\mathrm{<figure-callout\; id="1"\; label="V"\; filenames="C20048001339500221.png,C20048001339500241.png"\; state="\{\{state\}\}">V</figure-callout>}}_1$ It is related to the properties of particle materials, including their size, but not to the form of well construction. Here V ₁ =F ₁ σ characterizes the driving force. Similarly, the apparent spread σ(a) of the optical trap depends not only on the width σ _o of the focused beam, but also on the particle size:

${\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&ap;</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

Larger particles are affected by optical trapping over a greater range than smaller particles. The dependence of σ on a in this nature establishes the condition for exponential law sensitive separation. We continue to use formula (5) for illustration.

Thermal fluctuation analysis was used to characterize the depth of the optical trap for the data currently obtained for V _o /V ₁ = 1.3 and 0.73 for large and small spheres, respectively. The same analysis shows that the apparent widths of the wells are σ=0.94±0.07 μm and 0.74±0.07 μm. These results suggest a critical angle of θ = 14° ± 1° for large spheres and θ = 3° ± 2° for small spheres, consistent with the observation that large spheres are systematically trapped while small spheres escape. For barely trapped particles in an array of N traps, the total lateral deflection is (N-1)b sin θ. Therefore, by

Δ(a|b)=bsinθ (6)

The lateral deflection of each well was established to characterize the efficiency of the array. An interwell spacing b=2σ(a) over Δ=4/eV _o /F ₁ is chosen to optimize this efficiency. This result is useful for the design of a practical optical separation system, but it is not necessary to optimize its sensitivity to particle size.

Sensitivity can be expressed as

$\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&equiv;</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

and can be optimized by

$\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&PartialD;</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

so get

$\frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\sqrt{\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&chi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&chi;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

here

$\mathrm{<span\; class="notranslate">\; \&chi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; a</span>}}\frac{{\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

Equation (9) establishes the inter-trap spacing b at which the array of optical traps at angle Θ is most sensitive to distinguish between "large" particles that will be trapped and "small" particles that will escape.

As a practical example, these results can be used to optimize optical separation methods in viscous fluids. For particles whose size can be compared with the light wavelength or smaller, the depth of the potential well will be proportional to the particle volume V _o =Aa ³ , meanwhile, the viscous resistance is proportional to the radius V ₁ =Ba of the particle, so f(a) Proportional to ^a2 . Substituting the optimized interval b into the criterion based on fluid separation in equation (4), we get:

$\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; \&ap;</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; A</span>}{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; B</span>}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 3</span>}}{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 9</span>}}{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}}{{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 11</span>}<span\; class="notranslate">\; )</span>$

Equations (4) and (5) also show that the optical separation method depends only linearly on the depth of the potential well. Thus, variations in the actual depth of the actual array of optical vortex fluids only linearly degrade the resolution of the separation and can generally be compensated for by a generally stronger dependence on particle size.

In conclusion, the preceding examples have in fact illustrated the optical separation method using colloidal quartz sphere analog systems and also demonstrated that the technique is possible to achieve exponential sensitivity for size-based separations. The foregoing considerations illustrate that the geometry of the optical separation system can be chosen to optimize size-based separation, and exponential sensitivity is standard. Separation based on other features can be optimized using the same argument, although exponential sensitivity cannot generally be expected.

Equation (11) also provides insight into the possibility of applying optical separation methods to protein and nanocluster objects whose dimension a is measured in tens of nanometers. In particular, equation (11) shows that at a fixed angle θ, going from an object on the scale of 1 micrometer to an object on the scale of 10 nanometers will require orders of magnitude to increase the ratio A/B. This can in principle be achieved by increasing the intensity of the light, reducing its wavelength, and selecting a wavelength that resonantly enhances the interaction with the particles.

Implementing optical separation in this system involves creating an array of wells that span the incoming mixed fluid in such a way that the desired components of the particle are deflected across the demarcation line into the buffer fluid. On the one hand, successful manipulation requires the sample to have sufficiently low diffusivity or mobility that undesired components spontaneously cross the demarcation line at an acceptably low rate.

However, another aspect of the invention is directed to the opposite situation, where desired components are more diffuse or mobile than undesired components. Furthermore, this aspect also addresses situations where the desired component interacts more weakly than the other and thus is not selected by conventional optical separation methods. The greatest benefit of the present invention will be realized in a system that can implement both conditions, although not enough. Figure 1 depicts a microfluidic H-shaped intersection 100 with two liquid fluids. One fluid, the mixed input fluid 110, contains the heterogeneous sample to be separated. Another fluid, buffer fluid 120, consists of a background or buffer liquid. Objects in input fluid 110 encounter optical tweezers array 130 aligned at an angle θ to the fluid, which deflects selected components of the sample into buffer output fluid 140 for collection. The undeflected components of the sample remain in the original flow, output fluid 150, where they are collected.

As shown in FIG. 2, instead of creating an array of light traps that guide objects out of the mixed input fluid 110 and into the buffer fluid 120, the present invention can also use light as the object tries to cross the dividing line, either by diffusion or by active swimming The trap guides the object back into the incoming mixed fluid. As in the conventional method shown in Figure 1, the microfluidic H-shaped intersection 200 contains two flowing liquid streams, one of which 210 contains the heterogeneous sample to be separated, and the other 220 of which only Contains buffer solution. In the mixed input fluid 210, only those objects that attempt to cross the boundary between the two fluids encounter the optical trap array 230, which is arranged to guide objects back into the mixed input fluid 210. Objects that pass through the optical trap array 230 across the demarcation line are collected in the buffer output stream 240 . Those objects, either because they are less diffuse or because they are deflected by the optical trap array 230 , remain in the original input fluid and can be collected separately in the output fluid 250 . In this case, the less diffuse, or less mobile, objects are deflected back into the mixed input fluid, while the more mobile components will escape from the trap and be collected across the dividing line. Likewise, objects that are less strongly affected by the optical trap are more likely to be collected across the dividing line.

Whereas the optical separation method requires a sufficiently large number of optical traps to fill the entire mixed input fluid, the inverse process described above requires only enough optical traps to cover the area just around the boundary between the fluids. Therefore, the opposite optical separation method requires far fewer optical traps than conventional optical separation methods, and thus more efficiently uses the laser light needed to create the optical traps.

In this respect, optical separation methods have been well demonstrated to be superior to other sorting techniques, whereas optical separation methods conversely give the same advantages. These advantages include: continuous operation rather than batch operation; continuous optimization by adjusting laser power, laser wavelength, optical tweezers geometry, driving force, and exponential sensitivity to size. Optical separation methods, in contrast, extend these advantages to systems where conventional optical separation methods are either not applicable or impractical. Because with the conventional optical separation method, the opposite optical separation method can advantageously take advantage of the light polarization forming the trap, or the advantage of the beam mode structure forming the trap, according to the birefringence, optical rotation, elasticity of the object, and other factors such as size, Classify objects based on properties such as light scattering cross-section, light absorption rate, surface charge, and shape.

It is well known that microfluidic H-shaped crossings are useful for classifying objects according to their diffusivity. Adding an array of optical tweezers to the opposite optical separation method greatly enhances the selectivity of the process and gives a vast new physical basis for classifying objects.

Another aspect of the invention utilizes a thermal ratchet. Figure 6 (A-D) depicts the principle upon which the optical peristaltic operation is based and is used to explain the properties of the optical thermal ratchet. In FIG. 6A, a pattern of discrete light traps is drawn to position a single object. The pattern is represented schematically as two discrete potential energy wells, each of width σ, separated by a distance L. In fact, actual patterns may include many optical traps organized into conduits. The purpose of optical peristalsis and the optical thermal ratchet method disclosed herein is to transfer objects from one optical trap tube to another. The difference between the two approaches lies in how they accomplish this.

In optical creep, the initial pattern of wells is replaced by another pattern in which the tube is displaced by a distance comparable to σ (see Figure 6B). Because the new potential well overlaps the old one, the particles are deterministically transferred to the nearest conduit on the new pattern. Figure 6C repeats the process with yet another shifted well pattern. When projecting the original pattern, one cycle of optical peristalsis is completed (see Figure 6). The net effect of this cycle is to transport trapped particles from the conduit of one well in the first pattern to the next conduit also in the first pattern. In reality, there are many particles trapped in many optical traps; and in each cycle of optical peristalsis, all particles will be forwarded by a set of tubes. The direction of motion is unambiguously determined by the order of the sequence, and can be reversed by reversing the order.

Optical thermal ratcheting differs from optical peristalsis in that along the direction of motion the distance between the tubes is substantially greater than the width of the individual wells. Thus, particles trapped in the first pattern are free to diffuse when the second pattern is excited. Those particles that have diffused far enough to reach the nearest conduit in the second pattern are quickly localized. The localized components are then transported forward again (also by diffusion) when projecting the third pattern, and again when looping back to the first pattern. Unlike optical peristalsis, in which deterministic migration ensures that all objects trapped in the trap move forward in each cycle, the above-mentioned biased diffusion only migrates a certain component of the sample forward.

However, the aforementioned embodiments of the thermal ratchet lead to new opportunities. Particles that are too slow to catch up with the forward propagating waves may still have diffused far enough to catch up to the wells retrograde to their point of origin when illuminated by the third pattern of Figure 6C. The particles are transported backwards one-third of the distance between the tubes in each cycle. The population of the sequence of patterns through which the trap is formed, whether it moves forward or backward, is determined by the balance between the rate of particle diffusion and the rate of cycle of the sequence. Thus, changing the rate of circulation can change the direction of the average motion, a phenomenon also known as flow inversion.

Under the influence of the circulating optical tweezers pattern, the desired flux of particles can be calculated. The tweezers at position x _j can be simulated by a Gauss potential well.

${\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

The potential well has a depth V _o and a width σ. This potential well is obviously spatially symmetric. The pattern of wells establishes a state in a three-state cycle required for the ratchet to operate. As an illustrative example, consider that the wells in the pattern are separated by an equal distance L, so that the overall potential in state k is

${\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; N</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}\frac{\mathrm{<span\; class="notranslate">\; L</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

Here k=0, 1, or 2. Also, as shown in the illustrated example, the profile of the potential energy can be considered to cycle through these three states repeatedly at equal intervals T. This time can be compared with the time for particles with diffusivity D to diffuse through the system

$\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

quite. The result of the equilibrium between T and τ, determines the direction in which the sequence of potential energy states drives the particle through the system.

Under the combined influence of optical trap and random thermal force, the probability p(x, t)dx of finding a Brown particle at time t and position x in dx is determined by the main equation:

p(y,t+τ)=∫P(y,τ|x,0)p(x,t)dx (15)

Here, the propagation function for each state k is given by,

Pk(y,t|x,0)=e ^L (16)

For time τ<T, we have:

$\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\frac{{<span\; class="notranslate">\; \&PartialD;</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{<span\; class="notranslate">\; \&PartialD;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&beta;</span>}\frac{{<span\; class="notranslate">\; \&PartialD;</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 17</span>}<span\; class="notranslate">\; )</span>$

And here β ^-1 is the thermal energy ratio. The master equation for a complete three-state cycle is,

p(y, t+3T)＝∫dy ₃ P ₃ (y ₃ , T\y ₂ , 0)∫dy ₂ P ₂ (y ₂ , T\y ₁ , 0)∫dy ₁ P ₁ (y ₁ , T\x,0)p(x,t) (18)

For the symmetric optical tweezers potential, our consideration is that the master equation has a steady-state solution, so that:

p(x,t+3T)=p(x,t) (19)

The average velocity for this steady state is then given by:

$\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; \&Integral;</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\frac{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}\mathrm{<span\; class="notranslate">\; dx</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 20</span>}<span\; class="notranslate">\; )</span>$

Figure 7 plots the numerical solution to this system of equations for two representative values of βV _o = 10 and σ/L. For very small values of the cycle time T, the particles cannot keep up with the rapidly changing potential energy profile and thus diffuse randomly; the average velocity eventually equals zero in this limit. If the wells in successive patterns overlap (as shown by σ = 0.15L in Figure 7), the particles pass deterministically from well to well, resulting in a uniform positive drift velocity. This transfer achieves its maximum efficiency for moderate cycle times T and shows no improvement for longer dwell times. The result is that the drift speed drops when 1/T is in the long limit.

Wider spaced wells (shown in Figure 7 for σ = 0.10L) yield another behavior. Here, for sufficiently large values of T, the particle is able to keep up with the forward propagating wave. However, faster cycles lead to flow reversals characterized by negative v-values. This numerical result demonstrates the principle by which a fully symmetric thermal ratchet can be implemented with flow reversal using an array of optical tweezers.

As shown in Fig. 7, for the three-state cycle of Gauss well potential with βV _o =10, there is a transition from deterministic optical creep at σ = 0.15L to flow reversal thermal ratchet operation at σ = 0.10L.

So far, flow reversal due to changes in cycle time T has been described. For different populations in heterogeneous samples, their different diffusion coefficients produce different τ values, which can also cause flow reversal effects. These different populations can be caused to move in opposite directions simultaneously, simply by choosing T to drive one population forward and another backward. In this way, the optical thermal ratchet described is useful for the separation and purification of small liquid-borne objects.

A preferred optical approach to implementing a reversible thermal ratchet has advantages over other ratchet-based separation schemes. For example, a thermal ratchet based on an array of interdigitated electrodes has been applied to sort DNA fragments. But these thermal ratchets require complex microfabrication processes, whereas optical ratchets can be implemented cheaply and easily assembled into microfluidic devices for lab-on-a-chip applications. It was previously demonstrated that an optical ratchet based on a single time-shared scanning optical tweezer can induce flow reversal. Said approach relies on building up a spatially asymmetric potential energy profile in a time-averaged sense, so the system is based on a different principle than the above-mentioned process. In the preferred system described herein, each optical well in each pattern provides a spatially symmetric potential energy well; and the patterns themselves are also spatially symmetric. Unidirectional migration is driven by breaking spatiotemporal symmetry through sequences of at least three motifs per cycle.

One of the previously suggested examples of symmetric thermal ratcheting also involves a three-state sequence. However, the pathway relies on particles that are allowed to diffuse in only one state, and act as deterministic ratchets for the other two states, thereby biasing diffusion. The procedure described in this document involves diffusion and localization in all three states, thus giving more selective, and more rapid classification of inhomogeneous samples.

While preferred embodiments have been shown and described, it should be noted that changes and modifications may be made thereto by one of ordinary skill in the art without departing from the invention in its broader aspects. Various features of the invention are defined in the following claims.

Claims (22)

1. A device for classifying the population of tiny objects, comprising: first channel and second channel; The force source that drives the bulk passage of tiny objects; a plurality of conduits formed of laser beams located where the first and second channels converge to form a plurality of optical traps; The optical traps are organized into patterns, wherein the patterns are arranged such that the conduits of one pattern are separated by the conduits of the remaining patterns; and The optical traps are oriented at an angle relative to the driving force, and each pattern of optical traps is spaced apart from each other; Here, the population of minute objects is classified at least into required components and unnecessary components. 2. The apparatus according to claim 1, wherein the population of microscopic objects is dispersed in a liquid medium arranged in the first channel, wherein a buffer is also arranged in the second channel. 3. The apparatus according to claim 1, wherein the channel comprises an H-shaped cross. 4. Apparatus according to claim 1, wherein the optical trap is created by holographic optics. 5. The device according to claim 1, wherein the unwanted component is more diffuse or mobile than the desired component. 6. The device according to claim 1, wherein the desired component is more diffuse or mobile than the undesired component. 7. The apparatus according to claim 1, wherein the distance between the pipes along the direction of the driving force is greater than the width of the various wells. 8. Apparatus according to claim 7, wherein the light trap is adapted to move only a part of the population of microscopic objects forward when the patterns are excited sequentially. 9. Apparatus according to claim 7, wherein the patterns are adapted, when one pattern is excited, to cause a population of microscopic objects trapped in the other pattern to diffuse. 10. A method for continuously separating the totality of tiny particles into at least required components and unnecessary components, comprising the following steps: Provide an overall external force that drives tiny particles; Provide a plurality of patterns each including at least one pipe, each pipe includes at least one optical trap, and the distance between the pipes along the direction of the driving force is greater than the width of the optical trap; providing a laser beam so as to form a plurality of optical traps; and These optical traps are organized so that unwanted components can leave the traps while desired components remain. 11. The method according to claim 10, further comprising dispersing the population of microparticles in the liquid fluid fed into the first channel, wherein a buffer is also fed into the second channel. 12. The method of claim 10, wherein the channel comprises an H-shaped intersection. 13. The method according to claim 10, further comprising creating the optical trap using holographic optical tweezers technology. 14. A method according to claim 10, wherein only a certain fraction of the population of tiny particles moves forward through the optical trap when the patterns are excited sequentially. 15. A method according to claim 10, wherein when one pattern is excited, the population of tiny particles trapped in the other pattern is free to diffuse. 16. The method of claim 10, wherein some particles move forward through the array of optical traps and some particles move backward through the array of optical traps. 17. A method for continuously separating the totality of tiny particles, comprising the steps of: Provide an overall external force that drives tiny particles; focusing the laser beam to form many optical traps; and providing a plurality of patterns each comprising at least one conduit, each conduit comprising at least one type of optical trap; and At intervals, each pattern is excited, wherein, when one pattern is excited, particles trapped in the previously excited other pattern are free to diffuse; Here, the whole of fine particles is classified at least into necessary components and unnecessary components. 18. The method according to claim 17, further comprising creating the optical trap using holographic optical tweezers technology. 19. The method according to claim 17, wherein the distance between the pipes along the direction of the driving force is greater than the width of the various wells. 20. A method according to claim 17, wherein only a certain fraction of the population of tiny particles moves forward through the optical trap when the patterns are excited sequentially. 21. A method according to claim 17, wherein when one pattern is excited, the population of tiny particles trapped in the other pattern is free to diffuse. 22. The method of claim 17, wherein some particles move forward through the array of optical traps and some particles move backward through the array of optical traps.

Images