A 3D Integrated Circuit for Sensing Biological Nanoparticles

We have designed a dielectrophoresis-based lab-on-chip using MIT Lincoln Labs 0.18 μm 3D integrated circuit technology. Dielectrophoresis is the phenomena where electric fields are used to control the movement of particles in a fluid. The topmost tier of a three chip stack (figure 1) is used to create a microfluidic trench and dielectrophoresis electrode array. During the 3D fabrication process, the top chip tier is assembled upside down and the substrate material is removed, putting the polysilicon layer in close proximity to the outside surface. We used use this layer to create an array of 2,048 electrodes, each being 180 nm wide and 200 μm long with a gap spacing of 270 nm between electrodes. A microfluidic trench is created directly above the electrode array by designating a large region as an electrical contact pad and etching away the top-level metal. This technique allows for very dense electrode arrays and avoids the need for cumbersome post-processing steps to create on-chip microfluidic channels. The remaining two chip tiers are used for electronics. Each electrode is driven by analog circuitry on the middle chip tier and the bottom tier is used for logic to control the waveform on each electrode. Since individual nanoparticles are below the diffraction limit of conventional microscopes, detecting them on lab-on-chip devices is difficult. Additionally, in the case of analyzing living organisms, having to alter their biochemistry in order to detect them (e.g. fluorescence) is undesirable. Our approach is to use dielectrophoresis to arrange an ensemble of particles into periodic striped patterns that form a diffraction grating (figure 2) and measure its diffraction efficiency to sense the presence of particles. Particle-based gratings are created by programming the lab-on-chip to produce a pattern of alternating electric field maxima and minima. Particles in the solution will be attracted to the locations of the electric field peaks. Once the particles reach a steady-state, the resulting particle-grating will periodically modulate the index of refraction along the bottom surface of the microfluidic trench. The diffraction efficiency of this grating can be measured by using monochromatic incident light to observe the intensity of the reflected diffraction orders. A straightforward detection scheme that can be realized using this technique is a simple determination of whether or not particles are present in the sample under test. Since the periodicity of trapped particles (d) is programmable, the observation angle (θ) at which to expect diffracted orders for a given incident wavelength (λ) can be predetermined. Therefore, the discrete diffraction order (m), where mλ = d(sin θ + sin β), will only exist when particles are present in the sample. A more sophisticated analysis of the sample can be performed by creating a mapping between the magnitude of the diffracted orders and the physical dimensions of the particles (figure 3). This technique is especially well suited for biological particles because it does not require their natural optical properties to be artificially modified. Figure 1 Organization of 3D lab-on-chip Figure 3 Efficiency of reflected diffraction orders versus crosssectional area of trapped regions Figure 2 Nanoparticles arranged by dielectrophoresis to form a diffraction