Class 2 CRISPR-Cas systems are streamlined versions in which a single Cas protein (an effector protein, e.g., a type V Cas effector protein such as Cpf1) bound to RNA is responsible for binding to and cleavage of a targeted sequence. The programmable nature of these minimal systems has facilitated their use as a versatile technology that continues to revolutionize the field of genome manipulation.    Cas12 is an RNA-guided protein that binds and cuts any matching DNA sequence. Binding of the Cas12-CRISPR RNA (crRNA) complex to a matching single-stranded DNA (ssDNA) or double-stranded DNA (dsDNA) molecule activates the protein to non-specifically degrade any ssDNA in trans. Cas12a-dependent target binding can be coupled to a reporter molecule to provide a direct readout for DNA detection within a sample.  UC Berkeley researchers have developed compositions, systems, and kits having labeled single stranded reporter DNA molecules that provide a sensitive readout for detection of a target DNA. 

There are several emerging applications for Autonomous Underwater Vehicles (AUVs) where the agility and accurate control of location and/or orientation is critical. In the presence of random ocean currents and waves, conventional AUV systems need to use a combination of their thrusters to generate an appropriate force/torque and cancel the external disturbance to maintain the desired attitude or position. This is a relatively slow response since it requires accelerating and pushing water around the vehicle body. Thus, existing AUVs have disadvantages: (i) accurate and agile orientation and position control/stabilization is challenging; (ii) since thrusters are operational during reorientation maneuvers, a substantial amount of power is consumed to pump the bulk fluid, wasting the precious power storage of the vehicle and thus reducing its operational time; and (iii) drag forces and torques exerted on the thrusters significantly affect the efficiency of reorientation maneuvers.   UC Berkeley researchers have designed a new device for fast stabilization and control of an underwater robotic vehicle. In this architecture, the attitude maneuvers are performed using reaction torques that the body of the vehicle gains from a central inertial system.   

Rapid changes in the membrane potential of excitable cells (e.g., neurons and cardiomyocytes) play a central role in defining the cellular signaling and physiological profiles of these specialized cells. Typically, the membrane potential is monitored and measured via patch clamp electrophysiology, which involves the use of a micro-electrode attached to or near the cell of interests.  Unfortunately, the use of an electrode is highly invasive, limits records to the soma of a single cell and is extremely low throughput. Researchers at the University of California, Berkeley have designed and synthesized a voltage sensitive indicator that can provide excitation and emission profiles greater than 700 nm, and as such, represents an important method for visualizing membrane potential in living cells.

96 Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:"Calibri",sans-serif; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin;} The biological properties of trifluoromethyl compounds (e.g, CF3) have led to their ubiquity in pharmaceuticals, yet their chemical properties have made their preparation a substantial challenge, necessitating innovative chemical solutions.  For example, strong, non-interacting C-F bonds lend metabolic stability while simultaneously limiting the ability of chemical transformations to forge the relevant linkages and install the CF3 unit.  When these same synthetic considerations are extended toward the synthesis of trifluoromethylated positron emission tomography (PET) tracers, the situation becomes more complex.   UC Berkeley researchers discovered an unusual alternative mechanism, in which borane abstracts fluoride from the CF3 group in a gold complex. The activated CF2 fragment can then bond to a wide variety of other carbon substituents added to the same gold center. Return of the fluoride liberates a trifluoromethylated compound from the metal. This mechanism would be useful for the introduction of radioactive fluoride substituents for potential tracers to be used for positron emission tomography applications.

Researchers at Stanford and University of California, Berkeley, have developed an integrated MRI-Linac hybrid system that can increase the efficacy of image-guided radiotherapy (IGRT). This system allows more aggressive treatment strategies that employ dose escalation, tighter geometric margins and sharper dose gradients which can improve clinical outcomes. This radiotherapy treatment apparatus includes a treatment beam (charged by Linac, particle, proton, or electron beam), a magnetic field disposed parallel collinear to the treatment beam, and a target that is disposed along the treatment beam. MRI is ideal for IGRT, however, there is magnetic field and RF interference between the linear accelerator and MRI scanner. The configurations of this system overcome this issue.

The mechanical properties of cells derive from the structure and dynamics of their intracellular components, including the cytoskeleton, cell membrane, nucleus, and other organelles.  These, in turn, emerge from cell specific genetic, epigenetic, and biochemical programs, providing a link between cellular mechanics and the underlying molecular state.  Differences in mechanical properties reflect on cellular properties with clinical implications, including the metastatic potential, cell-cycle stage, and differentiation state of cells.  Yet, many mechanical aspects of various cells and sub-cell organelles remain unknown due to absence of appropriate analysis platforms. Atomic-force microscopy (AFM) and micropipette aspiration are the gold standards for performing mechanical measurements of cells, as they both provide controlled loading conditions and quantify such cellular properties as elastic modulus and cortical tension.  They are, however, burdened by slow throughput, capable of analyzing only just a few cells/hr.  Likewise, optical tweezers and microplate rheometry also suffer from low throughput.  Various microfluidic based platforms have been proposed for the high-throughput mechanical analysis of cells, including hydrodynamic stretching cytometry, suspended microchannel resonators (SMR), and real-time deformability cytometry (RT-DC).  Although each of these methods can analyze populations of cells in a relatively short time, they focus only on a single cellular property.  Consequently, these platforms, and the low-throughput traditional methods that under-sample, can neither identify cellular heterogeneity nor classify mechanical sub-phenotypes within a population. Investigators at UC Berkeley have developed a microfluidic platform, “mechano-node-pore sensing” (mechano-NPS), a rapid and multi-parametric cell screening platform, that simultaneously quantifies cell diameter, transit time through a contraction channel, transverse deformation under constant strain, and recovery time after deformation.  This platform efficiently reveals malignant-dependent mechanical phenotypes of cancer and normal epithelial cells, discriminates between sub-lineages of cells with accuracy comparable to flow cytometry, and determines the effects of chronological age and malignant progression on cell elasticity and recovery from deformation – based solely on a cell’s mechanical properties.

96 Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:Calibri; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin;} The CRISPR-Cas system is now understood to confer bacteria and archaea with acquired immunity against phage and viruses. CRISPR-Cas systems consist of Cas proteins, which are involved in acquisition, targeting and cleavage of foreign DNA or RNA, and a CRISPR array, which includes direct repeats flanking short spacer sequences that guide Cas proteins to their targets.  Class 2 CRISPR-Cas are streamlined versions in which a single Cas protein bound to RNA is responsible for binding to and cleavage of a targeted sequence. The programmable nature of these minimal systems has facilitated their use as a versatile technology that is revolutionizing the field of genome manipulation.  Current CRISPR Cas technologies are based on systems from cultured bacteria, leaving untapped the vast majority of organisms that have not been isolated.  There is a need in the art for additional Class 2 CRISPR/Cas systems (e.g., Cas protein plus guide RNA combinations).     UC Berkeley researchers discovered a new type of Cas protein, CasY.  CasY is short compared to previously identified CRISPR-Cas endonucleases, and thus use of this protein as an alternative provides the advantage that the nucleotide sequence encoding the protein is relatively short.  CasY utilizes a guide RNA to perform double stranded cleavage of DNA. The researchers introduced CRISPR-CasY into E. coli, finding that they could block genetic material introduced into the cell.  Further research results indicated that CRISPR-CasY operates in a manner analogous to CRISPR-Cas9, but utilizing an entirely distinct protein architecture containing different catalytic domains.   CasY is also expected to function under different conditions (e.g., temperature) given the environment of the organisms that CasY was expressed in.  Similar to CRISPR Cas9, CasY enzymes are expected to have a wide variety of applications in genome editing and nucleic acid manipulation.   

96 Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:Calibri; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin;} The CRISPR-Cas system is now understood to confer bacteria and archaea with acquired immunity against phage and viruses. CRISPR-Cas systems consist of Cas proteins, which are involved in acquisition, targeting and cleavage of foreign DNA or RNA, and a CRISPR array, which includes direct repeats flanking short spacer sequences that guide Cas proteins to their targets.  Class 2 CRISPR-Cas are streamlined versions in which a single Cas protein bound to RNA is responsible for binding to and cleavage of a targeted sequence. The programmable nature of these minimal systems has facilitated their use as a versatile technology that is revolutionizing the field of genome manipulation.  Current CRISPR Cas technologies are based on systems from cultured bacteria, leaving untapped the vast majority of organisms that have not been isolated.  There is a need in the art for additional Class 2 CRISPR/Cas systems (e.g., Cas protein plus guide RNA combinations).   UC Berkeley researchers discovered a new type of Cas protein, CasX, from groundwater samples. CasX is short compared to previously identified CRISPR-Cas endonucleases, and thus use of this protein as an alternative provides the advantage that the nucleotide sequence encoding the protein is relatively short.  CasX utilizes a tracrRNA and a guide RNA to perform double stranded cleavage of DNA. The researchers introduced CRISPR-CasX into E. coli, finding that they could block genetic material introduced into the cell.  Further research results indicated that CRISPR-CasX operates in a manner analogous to CRISPR-Cas9, but utilizing an entirely distinct protein architecture containing different catalytic domains.   CasX is also expected to function under different conditions (e.g., temperature) given the environment of the organisms that CasX was expressed in.  Similar to CRISPR Cas9, CasX enzymes are expected to have a wide variety of applications in genome editing and nucleic acid manipulation. 

Single-crystal x-ray diffraction is a powerful technique for the definitive identification of chemical structures.  Although most molecules and molecular complexes can be crystallized, often enthalpic and entropic factors introduce orientational disorder that prevent determination of a high-resolution structure.  Several strategies based on the inclusion of guests in a host framework that helps maintain molecular orientation have been used to overcome this challenge.  However, most of these methods rely primarily on weak interactions to induce crystalline order of the included molecules. Researchers at UC Berkeley have developed a strategy for crystallization of molecules within the pores of chiral metal-organic frameworks (MOFs) using coordinative bonding, which includes covalent and ionic bonds, and/or using chirality.  

96 Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:Calibri; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin;} The ∼10% optical contrast of graphene on specialized substrates like oxide-capped silicon substrates, together with the high-throughput and noninvasive features of optical microscopy, have greatly facilitated the use and research of graphene research for the past decade.  However, substantially lower contrast is obtained on transparent substrates. Visualization of nanoscale defects in graphene, e.g., voids, cracks, wrinkles, and multilayers, formed during either growth or subsequent transfer and fabrication steps, represents yet another level of challenge for most device substrates.     UC Berkeley researchers have developed a facile, label-free optical microscopy method to directly visualize graphene on transparent inorganic and polymer substrates at 30−40% image contrast per graphene layer.  Their noninvasive approach overcomes typical challenges associated with transparent substrates, including insulating and rough surfaces, enables unambiguous identification of local graphene layer numbers and reveals nanoscale structures and defects with outstanding contrast and throughput. We thus demonstrate in situ monitoring of nanoscale defects in graphene, including the generation of nano-cracks under uniaxial strain, at up to 4× video rate.  

Realizing the potential of massive sensor networks requires overcoming cost and power challenges. When sleep/wake strategies can adequately limit a network node's sensor and wireless power consumption, then the power limitation comes down to the real-time clock (RTC) that synchronizes sleep/wake cycles. With typical RTC battery consumption on the order of 1µW, a low-cost printed battery with perhaps 1J of energy would last about 11 days. However, if a clock could bleed only 10nW from this battery, then it would last 3 years. To attain such a clock, researchers at UC Berkeley developed a mechanical circuit that harnesses squegging to convert received RF energy (at -58dBm) into a local clock while consuming less than 17.5nW of local battery power. The Berkeley design dispenses with the conventional closed-loop positive feedback approach to realize an RCT (along with its associated power consumption) and removes the need for a sustaining amplifier altogether. 

Creating correct focus cues (blur and accommodation) has become a critical issue in the development of the next generation of 3D displays, particularly head-mounted displays. Without correct focus cues, current 3D displays create undue visual discomfort and reduce visual performance. Current attempts to solve the focus cues problem are limited in their practical use.  For example, volumetric displays are limited because the viewable scene is restricted to the size of the display volume.  Multi-plane displays require very accurate alignment between the display and the viewer’s eyes.  Light-field displays require demanding resolution requirements and computational workload.   Researchers at UC Berkeley have developed a system and rendering method to present correct focus cues with a conventional display. Taking the optics of the human eye into account, the researchers create on the user’s retina the correct depth-dependent optical effects that are experienced in the real world. The researchers have shown that their method, which currently incorporates chromatic aberration but can incorporate other effects, is better able to drive accommodation and create realistic depth appearance than conventional methods. The new method thereby enables perceptual realism rather than conventional photorealism, and this increases the immersion and realism of the visual experience. 

Piezoelectric Micromachined Ultrasonic Transducers (pMUTs) have attracted industry attention for their good acoustic matching, small geometry, low cost-by-batch fabrication, and compatibilities with CMOS and consumer electronics. While planar pMUTs have reasonable performance over bulk piezoelectric transducers, certain deficits remain in terms of coupling and acoustic pressure outputs, DC displacements, bandwidth, and power consumption. To address these deficiencies, researchers at the University of California, Berkeley, have developed a next generation of shaped pMUTs which are no longer fully defined by resonance frequency and can accommodate larger pressure outputs and bandwidths. This new pMUT apparatus can significantly boost overall performance while dramatically reducing power as compared to flat diaphragm state-of-the-art pMUTs.

Light is a wave, having both an amplitude and phase. Our eyes and cameras, however, only see real values (i.e. intensity), so cannot measure phase directly. Phase is important, especially in biological imaging, where cells are typically transparent (i.e. invisible) but yet impose phase delays. When we can measure the phase delays, we get back important shape and density maps.   Researchers at the University of California, Berkeley have developed a new method for recovering both phase and amplitude of an arbitrary sample in an optical microscope from a single image, using patterned partially coherent illumination. The hardware requirements are compatible with most modern microscopes via a simple condenser insert, or by replacing the entire illumination pathway with a programmable LED array, providing flexibility, portability, and affordability, while eliminating many of the trade-offs required by other methods. This enables quantitative imaging of phase from a single image, using partially coherent illumination, and in a way that is flexible and amenable to a variety of existing microscopy systems. 

The use of electric fields for signaling and manipulation is widespread, mediating systems spanning the action potentials of neuron and cardiac cells to battery technologies and lab-on-a-chip devices. Current FET- and dye-based techniques to detect electric field effects are systematically difficult to scale, costly, or perturbative. Researchers at the University of California Berkeley have developed an optical detection platform, based on the unique optoelectronic properties of two-dimensional materials that permits high-resolution imaging of electric fields, voltage, acidity, strain and bioelectric action potentials across a wide field-of-view.

The sensation of ocular discomfort commonly referred to as “dry eye” can be caused by various factors. The principal causative factors are (a) increased tear-evaporation rates attributable to meibomian gland dysfunction and insufficient/unbalanced tear-lipid films; (b) inadequate tear-aqueous production attributable to aging, medical procedures performed on the cornea (e.g., LASIK), or other general health conditions (e.g., autoimmune diseases); (c) environmental irritants (e.g., dust, smoke, wind, sun, or low humidity); and (d) eye strain attributable to extended viewing of computer monitors or other working environment-related factors. There are many different artificial-eye drops marketed and prescribed or recommended by medical practitioners to decrease dry-eye sensations. Unfortunately, all provide only short-term or no effects at all on tear-film stability and evaporation rates. Moreover, many artificial-tear formulations contain petrochemicals, (e.g., mineral oil) which have nothing in common with natural lipids comprising human tear-lipid films and might be potentially harmful to the eye.   Researchers at UC Berkeley have developed bicontinuous microemulsion formulations capable of delivering the components necessary to counteract compromised stability of tear-lipid layers and thus enhance the stability of entire tear films. These bicontinuous microemulsion components disperse spontaneously into a physical state that makes the microemulsion completely miscible with both human tear aqueous and human tear lipids. The components of these microemulsions are chemically identical or very close to natural tear lipids and tear aqueous and thus are completely biocompatible with human tear films. The lipids used in this formulation are biodegradable, and human tear enzymes will be able to metabolize these bicontinuous microemulsion lipids.  

Conventional cameras achieve color imaging by patterning organic dye color filters on top of photo detectors. However, due to the low absorption coefficients, organic dye filters cannot be made thinner than a few hundred nanometers, forbidding the realization of very small pixels. In addition, they are not durable under ultraviolet illumination or high temperature. Alternatively, optically thick plasmonic color filters have been realized, which can achieve pixel size down to a few microns. They are also superior to organic dyes regarding stability and design flexibility. However, the plasmonics color filters are still based on the conventional filtering scheme, which is intrinsically ineffective. Researchers at the University of California, Berkeley have developed a mechanism to achieve sub-micron pixel detection with very high photon efficiency. This novel mechanism is based on 3D semiconductor particles, which are more sturdy and easier to fabricate comparing to aforementioned techniques. At sub-micron pixel size, these resonant nano-structures outperform conventional color filters, which are limited by detrimental crosstalk between neighboring pixels.

Current super-resolution microscopy (SRM) methods have excellent spatial resolution, but no spectral information. Issues such as heavy color crosstalk, compromised image quality, and difficulties in aligning 3D coordinates of different color channels mean that high-quality multicolor 3D SRM remains a challenge. Another current imaging technique, single-molecule spectroscopy, is also limited in use because current methods are low throughput, have low spatial resolution, and cannot be used effectively for densely labeled biological samples.   UC Berkeley researchers have developed a 3-D super-resolution microscopy and single molecule spectroscopy system that addresses the issues inherent to both of these imaging techniques. By synchronously measuring the fluorescence spectra and positions of millions of single molecules within minutes, both spectrally resolved SRM and ultrahigh-throughput single-molecule spectroscopy are made possible.

The ability to see depth is a key visual function, as three-dimensional vision is used to guide body movements. Although many visual cues are used to infer spatial relationships, depth perception relies primarily on stereopsis, or the perception of depth based on differences in the images in the two eyes. More than 5% of the US population, however, is unable to see in three dimensions due to stereo-blindness and stereo-anomaly. Without depth perception, basic activities such as catching a ball or driving a car are not possible. Current therapeutic methods to address this issue include a set of eye-training exercises that aim to equalize the input from the eyes to the brain, which are collectively called orthoptics.   Researchers at UC Berkeley have developed an orthoptic method to train stereo depth perception. This method includes devices and systems for implementation, and it can be used in the home. 

Better understanding the brain's architecture and the behavior of neural networks requires non-invasive probes capable of monitoring brain activity at the scale of individual neurons.  Functional neuro-imaging methods have the advantage of being minimally invasive and can potentially resolve individual action potentials.  An ideal imaging method would be capable of quantifying many neurons simultaneously, have high spatial and temporal resolution, be non-invasive, and be accurate even in deep layers of brain tissue. There are a variety of current techniques available, many of which use mechanical scanning to reduce the effects of optical scattering and therefore have low temporal resolution. UC Berkeley researchers have developed a device capable of quantitative functional neuro-imaging in the thick brain tissue of live animals. By combining a detection method with algorithmic data processing, this device achieves single neuron resolution and fast sampling rates with high spatial and temporal resolution.  

Cathodoluminescence (CL) is used for nanoscale imaging by detecting the light generated in the sample by the application of an electron beam. Direct CL has also been used to image biological samples, but typically causes damage to the sample and can result in poor imaging quality.  Methods which incorporate inorganic cathodoluminescent nanoparticle labels into a biological sample result in less sample damage, but imaging with nanoparticle labels requires the electron beam to penetrate into the sample, which precludes repeated measurements or observations of dynamics. A UC Berkeley researcher has developed an optical imaging system and method for producing nanoscale images with high resolution, images of fragile samples without damaging the samples and that can be used for repeated imaging of a sample which allows observation of sample dynamics.  

Time-resolved phase contrast MRI (4D flow) can quantify cardiac function and flow. The technique may even permit complex anatomical assessment, thus comprising a comprehensive exam in a single scan. Unfortunately, artifacts from respiratory motion compromise this ability. Therefore, we developed a simple method to measure motion using readily available navigation information from the velocity-encoding gradients without any significant modification to conventional sequences.

Existing structured illumination microscopy can increase the resolution of an image by a factor of 2, however, it requires expensive optical or mechanical devices (e.g., spatial light modulators, digital micromirror devices, or piezo translation stages).  Also, the existing movable patterned mask generated by the light interference using SLM or DMD suffers from hysteresis and repeatability problems due to mechanical motion.  UC Berkeley researchers have developed a computational illumination hardware and software system that can achieve pattern-shift on the object plane without mechanically switching the patterns and therefore without movement of any component.  The lensless system can be easily implemented into existing microscopes without extra hardware can can be extremely fast and suitable for real-time imaging applications.     

Piezoelectric Micromachined Ultrasonic Transducers (pMUTs) have attracted industry attention for their good acoustic matching, large bandwidth, miniaturization, and low cost-by-batch fabrication. pMUTs have the advantages of low power consumption and large deflection for high-acoustic power applications. However, low electromechanical coupling has been a serious drawback for pMUT applications, in some cases foreclosing key opportunities. In response to this challenge, researchers at UC Berkeley have developed a bimorph pMUT with unique advantages which dramatically improve the device capabilities: the bimorph pMUT utilizing two active AlN layers in a CMOS-compatible process. This innovative design is the first bimorph pMUT with two active piezoelectric layers separated by a common electrode. The prototype bimorph pMUT has a resonant frequency of 198.8 kHz and central displacement of 407.4 nm/V. Under the differential drive scheme using the dual electrodes at low frequency, the measured central displacement is 13.0 nm/V, which is about 400% higher than that of a unimorph AlN pMUT. This revolutionary dual electrode bimorph pMUT presents a new class of design/fabrication for exciting pMUT applications, including range finders and gesture recognition devices.

Curved pMUTs, as developed at UC Berkeley, have been shown to have 2 orders of magnitude improvement over flat pMUTs as well as the capacity for post-processing tuning. However, it is desirable to improve production methods to make this innovation more commercially applicable.To meet this challenge, investigators at Berkeley have developed a self-curved diaphragm process using stress engineering to produce highly responsive curved pMUTs. This diaphragm pMUT can boost 6X better performance compared to the flat diaphragm state-of-the-art pMUT. CMOS foundry-based process flow has produced self-curved diaphragms by engineering residual stress in thin films to construct molds for fabrication. Benefits of the invention include achieving silicon curved molds by patterning thin layers of stressed silicon nitride and silicon oxide layers on top of a silicon plate of a predetermined thickness.