Patent Application
20070139150
Inductors have been among the most difficult circuit elements to form on an integrated circuit. Principle challenges include finding ways to form coil structure that are easy to replicate in a mass produced semiconductor processing environment.
One method of fabricating coils has been to use self assembled stress engineered thin films to form a coil structure. In the stress-engineered thin film process, a first portion of a metal strip is deposited over a substrate and a second portion of the metal strip is deposited over a release layer. The metal is deposited such that an internal stress exists in the metal. After the release layer is removed, usually via etching, the second portion of the metal strip curls to form a winding of the coil. A load layer deposited over the second portion of the metal strip controls the amount of curling.
Many different materials may be used as the load layer. In one implementation, a polymer acts as a load layer that controls the curling of the metal. When a polymer is used for loading, the polymer is heated to temperatures of around 180 to 290° C. The amount of heating controls the reflow rate of the polymer. The reflow rate of the polymer controls the amount of curling. Thus higher temperatures cause more polymer softening and allows the stress-engineered thin film to further curl out of the plane of the substrate.
Several problems arise from the above described technique. A first problem is that heating the polymer to such temperatures can cause the polymer to burn. Burned polymer can be hard to remove in subsequent processing steps.
A second problem with this coil fabrication technique is the close proximity of coil windings. Stress-engineered thin films are often formed from metals that are alloys and may not have the highest conductivity. In order to enhance coil conductivity, the coil may be plated with a higher conductivity metal. However, the close proximity and the unusual fields that exist around the coil base (the first portion of the metal) can cause the plating material to short adjacent windings.
Thus an improved method of loading the stress-engineered thin film springs and also of sealing the electrodes is needed.
An improved method for reflowing a polymer layer is described. The method involves placing the polymer laser in an enclosed environment and exposing the polymer to a solvent vapor such as an acetone vapor. The solvent vapor softens the polymer. The described method is particularly suitable for reflowing polymer used to load release portions of coil structures and to seal the metal edges prior to plating the coil structures.
An improved method of forming microcoil structures is described. The microcoil structures are particularly useful in forming integrated circuit (IC) solenoid type devices and/or inductors. The improved method uses a solvent vapor to soften and/or reflow a polymer avoiding many of the problems associated with using heat or liquid acetone for polymer reflow.
As used herein, a stress-engineered thin film is defined as a material that includes a built in stress differential. The stress differential is produced in the spring material by one of several techniques. According to one technique, different materials are deposited in layers, each having a desired stress characteristic, for example a tensile layer formed over a compressive layer. According to another technique a single layer is provide with an intrinsic stress differential by altering the fabrication parameters as the layer is deposited. The spring material is typically a metal or metal alloy (e.g., Mo, MoCr, W, Ni, NiZr, Cu), and is typically chosen for its ability to retain large amounts of internal stress. Microsprings are typically produced using known photolithography techniques to permit integration of the microsprings with other devices and interconnections formed on a common substrate. Indeed, such devices may be constructed on a substrate upon which electronic circuitry and/or elements have previously been formed.
In one example of fabricating a stress-engineered thin film, a metal or metal alloy is deposited in such a way that an internal stress gradient is built into the metal. In one example, stress-engineered thin film 108 is a nickel-zirconium alloy although many other materials may also be used. Typically, stress-engineered thin film 108 is deposited in a series of sublayers, 120, 124, 128 to create the internal stress gradient. Electroless or electroplating techniques may be used to deposit the stressed metal. For example, the built in stress gradient may be obtained by plating from two baths with different concentrations or by varying the current density during plating.
In another example technique, the sublayers are sputtered such that the atomic spacing is larger in the upper sublayers of the spring material resulting in a stress gradient. Different stress levels can be introduced into each sublayer of the deposited stress metal during sputter deposition in a variety of ways including adding a reactive gas to the plasma, varying the deposition angle or changing the pressure of the plasma gas. Varying the pressure of the plasma gas (typically argon) during deposition provides the simplest method of varying stress in deposited sublayers. As the pressure of the plasma gas increases, the film stress in the deposited layers becomes more tensile. Thus, by increasing pressure, a stress which is more compressive in the lower sublayers of the metal layer and that becomes increasingly tensile towards the upper sublayers of the metal layer is formed. A more detailed description of forming such stress gradients is provided in U.S. Pat. No. 5,613,861 entitled “Photolithographically Patterned Spring Contact” by Donald Smith et al. and hereby incorporated by reference in its entirety.
When the spring is ready to be released and to form a coil structure, release material 104 is etched away using an etchant. Etching the release material separates the release portion 116 of the stress-engineered thin film 108 from substrate 100. After release layer removal, only the loading provided by polymer layer 132 such as a photo-resist prevents release portion 116 from curling out of a plane parallel to a surface of substrate 100.
The rate and amount of polymer softening and reflow controls coil formation. In particular, after release layer removal, the reflow and softening of the loading polymer controls the release of the stress in stress-engineered thin film 108. Typically, a coil with 450 and 600 um diameter windings are particularly suitable for IC inductor formation.
An example of edge sealing is shown in
In order to reflow the polymer, the polymer is exposed to a solvent vapor. The solvent vapor initially softens the polymer. After extreme softening, the polymer reflows. The reflow causes polymer 908 to flow over the edge 912 of the thin film. Typically the polymer is formed within 5 micrometers of the thin film edge to facilitate reflow sealing. After reflow, the polymer contacts the substrate 900 and seals thin film edge 912 as shown in
After etching away the release layer, the polymer is softened and/or reflowed. Traditional polymer reflow uses heat or a liquid chemical applied directly to the polymer. However, it has been found that using a vapor induced reflow improves the removability of the polymer in later processing steps.
In
Increased acetone vapor concentration may be created by boiling the acetone. Acetone boils around 57 degrees centigrade. Other solvents may boil at higher temperatures, for example SVC 150 boils around 80 degrees centigrade and Futurex boils at around 110 degrees centigrade. Lowering the surrounding pressure allows the solvents to boil at lower temperatures. The acetone vapor partially dissolves and softens the polymer load on the sample allowing internal stresses in the stress-engineered thin film to release. When sealing is desired, significant softening may result in polymer reflow. The amount of acetone in the ambient controls how quickly the photoresist softens. The higher the acetone partial pressure, the faster the photoresist softens. A desired partial pressure can be established by using standard mass flow controllers to meter a controlled flow of carrier gas such as nitrogen through an acetone bubbler into the reflow chamber. In practice, we found that simply placing a 25 ml beaker of acetone in a sealed 0.25 ft3 reflow chamber creates an environment that is adequate for achieving controllable reflow. The beaker of acetone establishes a natural equilibrium partial pressure with the volume of air in the chamber. The volume of acetone needed can be easily scaled with other chamber sizes to achieve a desired partial pressure of solvent. When the amount of solvent is sufficient such that some liquid solvent remains in the chamber, the atmosphere in the reflow chamber is considered “saturated”.
To control the process of softening and/or reflow, the solvent vapor is regulated. As used herein regulated means controlled. Various methods of regulating the solvent vapor exist. In a first method, a determined quantity of liquid solvent is placed with a fixed volume of air in a sealed container as shown in
As used herein, polymer reflow is defined as a decrease in the polymer Young's modulus; i.e., becoming softer than its room temperature stiffness. When used as a load layer, an important property of the polymer (resist) is the dramatic lowering of polymer stiffness when heated. Sufficient softening to allow flow off of an edge is important only for sealing edges for plating. For loading control purposes, such a significant softening is not necessary. Gradual stiffness reductions result in higher yields compared to systems that quickly reduce the stiffness.
In the current example, in the release portion of the stress-engineered thin film, the polymer flow weakens the rigidity of the polymer allowing gradual release of the internal stress in the stress-engineered thin film and curling of the stress-engineered thin film out of the plane of the substrate. The movement of the metal further redistributes the polymer. In the non-release fixed regions, the polymer reflow seals the metal edges, preventing electrical shorting to adjacent metal edges during subsequent plating processes.
After stress-engineered thin film deposition, a polymer is deposited in block 412. The polymer may include two components, a first component deposited over part of the release portion of a coil winding loads the release portion. A second component deposited near winding edges in the ground plane helps to seal and electrically isolate the individual winding edges during future plating operations. Various methods may be used to deposit and pattern the polymer, including printing the polymer, using photolithographic techniques to pattern the polymer and other techniques known to those of skill in the art.
In block 416, an etchant removes the release layer. Release layer removal separates the release portion of the stress-engineered thin film from the underlying substrate. In block 420, the sample including the polymer is pre-annealed. The pre-anneal may be performed at any time after polymer deposition. Thus the pre-anneal may occur before or after release etching. The pre-anneal is typically done at a temperature between 175 and 195 degrees centigrade, and is done more typically between 180 and 190 degrees centigrade or as close to 185 degrees centigrade as possible. Typically, the pre-anneal temperature is maintained for a time period of between 0.5 and 3 minutes. It should be understood that the pre-anneal is optional, however it has been experimentally determined that pre-annealing prior to performing a vapor reflow of the polymer substantially improves pattern definition. It has also been experimentally determined that pre-anneal temperatures at 175 degrees and 195 degrees centigrade are substantially less effective then pre-anneals at 185 degrees centigrade. Accurate pattern definition is particularly helpful in preventing short circuits between adjacent coils during plating. Ideally, the coil should be electrically isolated when used at high frequencies which can be on the order of giga-hertz.
In block 424, the sample is placed inside a chamber and exposed to acetone vapors. Exposure may occur at room temperature (typically between 10 and 45 degrees centigrade) for a period of 1 to 3 minutes in a sealed acetone vapor environment. When extremely tight control is needed, pumps may be used to maintain a vacuum and valves may precisely control vapor flow and acetone partial pressures. However, placing acetone in a sealed chamber with the sample at room temperature for a sufficient time period generates sufficient acetone vapor concentrations to soften the polymer and allow curling of the release portion of the stress-engineered thin film.
After polymer reflow, the coil is plated in block 428. Plating forms a thin layer of high conductivity metal, such as copper, onto the exposed portions of the coil. The plating improves coil conductivity. When the coil serves as an inductor, reducing the coil resistance improves the Q factor. The plated layer is typically less than 20 micrometers thick. However, in an inductor the skin effect confines current to the outer surface of the coil. Due to the skin effect, the outer layer conductivity dominates the high frequency resistance of the inductor loop. Plating also improves the connection where the coil windings close on themselves during assembly.
The plating process also attaches high conductivity metal to exposed portions of the coil in the ground plane. Since there is minimal spacing, often less than 15 micrometers, separating adjacent windings, slight additions of excess conductor plating material can create an electrical short between adjacent windings. Thus plating should be controlled to prevent short circuits. One method of preventing shorts uses reflowed polymer to seal coil edges in the ground plane. Sealing the winding metal edges prevents plating material from laterally growing from the edge of a first coil winding to contact an adjacent coil winding and creating an electrical short.
After plating, remaining polymer may be removed as shown in block 432. Because the polymer has not been heated beyond the temperatures used in the pre-anneal, the polymer typically has not been burned. The lack of heating beyond the pre-anneal avoids polymer burning and facilitates polymer removal. Thus a standard polymer remover such as liquid acetone can be used to chemically remove the polymer.
Although two example coils have been shown, numerous other coil structures may be formed. Examples of alternative coil structures are described in U.S. Pat. No. 6,621,141 entitled Out-of-Plane Microcoil With Ground Plane Structure by Van Schuylenbergh et al. which is hereby incorporated by reference in its entirety.
The preceding description includes a number of details such as dimensions, temperatures, types of material and example coil implementations. Such details are intended to facilitate understanding of the invention and provide examples. Such details should not be used, and are not intended, to limit the invention. Instead, the invention should only be limited by the claims, as originally presented and as they may be amended, encompass variations, alternatives, modifications, improvements, equivalents, and substantial equivalents of the embodiments and teachings disclosed herein, including those that are presently unforeseen or unappreciated, and that, for example, may arise from applicants/patentees and others.
