Design simualtion and analysis of piezoresistive microcantilever for biosensing applications

In the past decade, several research works demonstrated the ability of Biological Microelectromechanical System (Bio-MEMS) biosensors to detect of biomolecules such as Deoxyribonucleic Acid (DNA), proteins, Bacteria and Antigens. But due to the low concentration of the analytes that need to be detected in the samples,a minuscule signal results in the output of the sensor. In response to this, a need arisen for an optimized biosensor capable of giving high output signal in response the detection of few analytes in the sample; the ultimate goal is being able to convert the attachment of a single biomolecule into a measurable quantity. For this purpose, MEMS microcantilevers based biosensors have emerged as a promising sensing solution because it is simple, cheap, highly sensitive and more importantly does not need analytes optical labeling (Label-free). Among the different microcantilever transducing techniques, piezoresistive based microcantilever biosensors seem to be a more attractive solution being cheap, high sensitive, miniature, works well in liquid environments and having integrated readout system. Even though there are many publications in literature that concentrated on increasing the piezoresistive microcantilevers sensitivity, they only considered in optimizing few design and process parameters thus the resultant sensitivity enhancements are not good enough for practical applications. After the analyzation of the work found in literature, it was found that the parameters/approaches that be can be optimized/used to enhance the sensitivity of Piezoresistive microcantilever-based sensors are: Cantilever dimensions, Cantilever Material, Cantilever Shape, Piezoresistor's material, Piezoresistor's doping level, Piezoresistor's Dimensions, Piezoresistor's position, Stress concentration Region's (SCR) shape and position. In this study, after a systematic analyzation of the effect of each design and process parameters on the sensitivity, a step-wise optimization approach was developed in which almost all these parameters were variated one at each step while fixing the others to get the maximum possible sensitivity at the end. Throughout this work, COMSOL Multiphysics 5.0, a commercial Finite Element Analysis (FEA) tool, was used to simulate the sensor performance. At each optimization step, the goal was to optimize the parameter in such a way that it maximizes and concentrates the stress in piezoresistors region for the same applied force thus get the higher sensitivity. In total, almost 46 different simulations were done to get the final optimized sensor. Starting with a rectangular cantilever, the piezoresistor material and doping level were optimized in two steps. When the piezoresistor material was varied (single crystal silicon and Poly-silicon), it was found that the ΔR⁄R sensitivity is higher in the case of single crystal silicon. xxiv But for this sensor design, polysilicon has been chosen as the piezoresistor material because it's sensitivity does not depend on the crystal orientation, the sensor fabrication is easier, cheaper and can be realized in ITUnano laboratory. Next, by changing the doping level in the range between 1×1015 cm−3 to 1×1020 cm−3 and calculating the ∆R/R sensitivity, the doping level that will be used throughout the following simulations was determined. It was found that, 1×1018 cm−3 doping level is high enough to reduce the thermal noise effect, at the same time it does not be affected the sensitivity that much. Thus this doping level was chosen and used throughout the following simulations. Afterward, the cantilever material is varied to find the material that gives maximum stress and deflection for the same applied force. It was found that SiO2 resulted into almost 2.5x higher deflection and 1.7x higher sensitivity when compared to single crystal silicon (the starting cantilever material) case thus SiO2 has been selected as the cantilever material for this biosensor and it is used in the following optimization steps. Next, various cantilever shapes (Rectangular, Pi-shape, T-shape, Trapezoid, SteppedTrapezoid, and Triangular) were introduced, and for each shape, the dimensions were varied bearing in mind the process and device limits. The results from all these simulations were compared to find the optimized shape which gives the maximum sensitivity. During the rectangular shape microcantilever optimization step, it was found that the cantilever thickness has the highest effect on the sensor sensitivity when compared to the change in cantilever length and width. In addition to that, after the different rectangular microcantilever dimensions were optimized (length, width and thickness), the sensitivity increased 18.3x folds. Also, adding two side holes to the rectangular cantilever structure (T-shape) increased the sensitivity by 1.6 factor. Overall, for the same applied force, the trapezoid-shaped microcantilever design gave higher sensitivity (more than 46x times greater than the starting sensor sensitivity) whereas the stepped-trapezoid shaped gave the highest maximum deflection. Afterward, Stress Concentration Region (SCR) was introduced in the optimized trapezoid structure in different locations and orientations seeking for further sensitivity enhancement. From the simulations, it was found that adding a 30µ×10µm SCR rectangular hole to the optimized trapezoid structure 15µm away from the clamped cantilever edge, resulted in almost 1.6x times sensitivity enhancement which gave the best sensitivity value compared to the other positions. Regarding the normalized change in resistance to the applied force the final sensor's sensitivity equals to -1.5×10-8 Ω/Ω ⁄pN; this means that for each 1pN (10-10 g) biomolecules attach to this biosensor; the piezoresistor resistivity will decrease by 1.5×10-8 Ω. When compared to the starting sensor, the final sensor design gave 73.5x times better ΔR⁄R sensitivity and it is more sensitive than the other sensor designs previously reported in the literature. The fabrication sequence for this sensor was prepared, but due to technical problems in some of the devices found in ITUnano laboratory, the sensor has not been fabricated.