October 2012
Spotlight Summary by Francesco Morichetti
Direct fabrication of silicon photonic devices on a flexible platform and its application for strain sensing
Don’t be so rigid, even photonics can be softer...
Everybody working in photonics has the perception that the natural substrate for integrated devices is a bit of a rigid material, which is normally provided by a semiconductor wafer or by a crystal. This picture comes from well established fabrication methodologies, but it is not necessarily what has to be. Although being somewhat surprising, optical materials like semiconductors, that are brittle and fragile as a bulk, become soft and flexible when shrunk to nanometer scale. With this in mind, we can really imagine the concept of silicon waveguides, fabricated on a flexible substrate, which can be elastically deformed with neither cracks nor damages, as if they were made of plastic themselves.
Material properties tell us that there are no fundamental barriers to make silicon photonics on a plastic sheet, but how to do it in practice? Several strategies have been explored so far, in which photonic devices are patterned on a sacrificial solid substrate and then transferred to a flexible substrate. Yet, all these approaches suffer from severe limitations on the size and geometry of the manufactured devices, and require extremely smooth surfaces for chip bonding and extremely sophisticated alignment equipment.
The treasure map disclosing a viable route to fabricate silicon photonic devices directly on a flexible substrate seemed far away from being discovered, at least until the work by L. Fan and coworkers. They found a simple and reproducible way to integrate complex photonic structures on a plastic film without the need of transferring them from another rigid substrate. Simple to say, but to succeed in their goal they had to address and solve a number of critical problems.
Temperature, first of all, because the mechanical properties of conventional polymers dramatically suffer from high temperature process. To circumvent thermal budget issues, they deposited hydrogenated amorphous silicon (a-Si) by using plasma enhanced chemical vapor deposition (PECVD) at a temperature of less than 200° C (well below the 300°C - 400°C range of conventional processes) in order to avoid substrate fractures and deformation. A low temperature deposition process is responsible for a bit higher materials loss, resulting in microresonators with a Q factor (about 7000) slightly smaller than state-of-the–art a-Si resonators. But a route for process optimization is planned to improve the resonator quality in the future.
Second, surface quality of the polymer substrate. Minor imperfections on the surface profile (due to polymerization and cross-linking process) can be responsible for stitching errors in the electron beam-lithography (EBL) used for waveguide writing. Here, an accurate surface cleaning process was developed to remove particles and organics on the surface, smoothing the polymer surface roughness down to 1.5 nm.
Third, sample handling. By definition, a thin and flexible substrate is difficult to handle during processing. The solution here is to sandwich the polymer substrate with the deposited a-Si film between two Si substrates with a small window on the top to access the device region. This provides the required rigidity and flatness for all the subsequent process steps, that are resist deposition, EBL writing, reactive ion etching and polymethyl methacrylate (PMMA) upper cladding coverage.
The combination of these three simple ingredients is the recipe cooked by L. Fan and coworkers to build silicon photonic devices on a flexible substrate, enabling multiple fabrication steps and the manufacturing of complex photonic integrated circuits with no apparent size or shape limitations on the patterns that can be realized. Although EBL was used as a proof-of-concept of this new fabrication methodology, it can be replaced by more cost-effective lithography methods such as deep-UV stepper or nanoimprint lithography in order to achieve low-cost fabrication.
Speaking about flexibility in photonic integrated circuits, people have in mind the need for more complex circuit architectures and functionalities that can be implemented. Now we can also add a radically new concept of flexibility for photonics, which refers to the mechanical properties themselves of the integrated devices and of the substrate. In this view, the large scale of integration provided by silicon photonics combined with the flexibility of polymeric substrates is likely to open exiting perspectives in a variety of fields, especially in sensing applications.
Flexibility, with all its different connotations, sounds as a keyword for the development of future photonic integrated circuits. And it is also a multi-nuance concept that contributes to smoothing out the rigidity in our conventional way of thinking photonics.
You must log in to add comments.
Everybody working in photonics has the perception that the natural substrate for integrated devices is a bit of a rigid material, which is normally provided by a semiconductor wafer or by a crystal. This picture comes from well established fabrication methodologies, but it is not necessarily what has to be. Although being somewhat surprising, optical materials like semiconductors, that are brittle and fragile as a bulk, become soft and flexible when shrunk to nanometer scale. With this in mind, we can really imagine the concept of silicon waveguides, fabricated on a flexible substrate, which can be elastically deformed with neither cracks nor damages, as if they were made of plastic themselves.
Material properties tell us that there are no fundamental barriers to make silicon photonics on a plastic sheet, but how to do it in practice? Several strategies have been explored so far, in which photonic devices are patterned on a sacrificial solid substrate and then transferred to a flexible substrate. Yet, all these approaches suffer from severe limitations on the size and geometry of the manufactured devices, and require extremely smooth surfaces for chip bonding and extremely sophisticated alignment equipment.
The treasure map disclosing a viable route to fabricate silicon photonic devices directly on a flexible substrate seemed far away from being discovered, at least until the work by L. Fan and coworkers. They found a simple and reproducible way to integrate complex photonic structures on a plastic film without the need of transferring them from another rigid substrate. Simple to say, but to succeed in their goal they had to address and solve a number of critical problems.
Temperature, first of all, because the mechanical properties of conventional polymers dramatically suffer from high temperature process. To circumvent thermal budget issues, they deposited hydrogenated amorphous silicon (a-Si) by using plasma enhanced chemical vapor deposition (PECVD) at a temperature of less than 200° C (well below the 300°C - 400°C range of conventional processes) in order to avoid substrate fractures and deformation. A low temperature deposition process is responsible for a bit higher materials loss, resulting in microresonators with a Q factor (about 7000) slightly smaller than state-of-the–art a-Si resonators. But a route for process optimization is planned to improve the resonator quality in the future.
Second, surface quality of the polymer substrate. Minor imperfections on the surface profile (due to polymerization and cross-linking process) can be responsible for stitching errors in the electron beam-lithography (EBL) used for waveguide writing. Here, an accurate surface cleaning process was developed to remove particles and organics on the surface, smoothing the polymer surface roughness down to 1.5 nm.
Third, sample handling. By definition, a thin and flexible substrate is difficult to handle during processing. The solution here is to sandwich the polymer substrate with the deposited a-Si film between two Si substrates with a small window on the top to access the device region. This provides the required rigidity and flatness for all the subsequent process steps, that are resist deposition, EBL writing, reactive ion etching and polymethyl methacrylate (PMMA) upper cladding coverage.
The combination of these three simple ingredients is the recipe cooked by L. Fan and coworkers to build silicon photonic devices on a flexible substrate, enabling multiple fabrication steps and the manufacturing of complex photonic integrated circuits with no apparent size or shape limitations on the patterns that can be realized. Although EBL was used as a proof-of-concept of this new fabrication methodology, it can be replaced by more cost-effective lithography methods such as deep-UV stepper or nanoimprint lithography in order to achieve low-cost fabrication.
Speaking about flexibility in photonic integrated circuits, people have in mind the need for more complex circuit architectures and functionalities that can be implemented. Now we can also add a radically new concept of flexibility for photonics, which refers to the mechanical properties themselves of the integrated devices and of the substrate. In this view, the large scale of integration provided by silicon photonics combined with the flexibility of polymeric substrates is likely to open exiting perspectives in a variety of fields, especially in sensing applications.
Flexibility, with all its different connotations, sounds as a keyword for the development of future photonic integrated circuits. And it is also a multi-nuance concept that contributes to smoothing out the rigidity in our conventional way of thinking photonics.
Add Comment
You must log in to add comments.
Article Information
Direct fabrication of silicon photonic devices on a flexible platform and its application for strain sensing
Li Fan, Leo T. Varghese, Yi Xuan, Jian Wang, Ben Niu, and Minghao Qi
Opt. Express 20(18) 20564-20575 (2012) View: HTML | PDF