Patent Application
20170137959
Manufacture of semiconductor integrated circuits and other micro-scale devices typically requires formation of multiple metal layers on a wafer or other substrate. By electroplating metals layers in combination with other steps, patterned metal layers forming the micro-scale devices are created.
Electroplating is performed in an electroplating processor with the device side of the wafer in a bath of liquid electrolyte in a vessel, and with electrical contacts on a contact ring touching a conductive seed layer on the wafer surface. Electrical current is passed through the electrolyte and the conductive layer. Metal ions in the electrolyte plate out onto the wafer, creating a metal layer on the wafer.
Electroplating processors typically have consumable anodes, which are beneficial for bath stability and cost of ownership. For example, it is common to use copper consumable anodes when plating copper. The copper ions moving out of the plating bath to form the plated copper layer on the wafer are replenished by copper ions coming off of the anodes, thus maintaining the copper ion concentration in the plating bath. This is a cost effective way to maintain the concentration of metal ions in the bath compared to replacing the electrolyte bath. However, using consumable anodes requires a relatively complex and costly design to allow the consumable anodes to be periodically replaced. If the anodes are replaced through the top of the chamber, then the electric-field shaping hardware is disturbed requiring re-checking the performance of the chamber. If the anodes are replaced from the bottom of the chamber, then extra complication is added to the chamber body to easily remove the lower section of the chamber and add reliable seals.
Even more complexity is added when consumable anodes are combined with a membrane (for example a cation membrane) to avoid degrading the electrolyte, or oxidizing the consumable anodes during idle state operation, and for other reasons. Cationic membranes allow some metal ions to pass, which lowers the efficiency of the replenishment system and may require an extra compartment and electrolyte to offset loss of metal ions through the cationic membrane.
Electroplating processors using inert anodes have been proposed as an alternative to using a consumable anode. An inert anode processor may reduce complexity, cost, and maintenance. However, use of inert anodes has led to other disadvantages, especially related to maintaining the metal ion concentration in a cost effective manner compared to consumable anodes, and the generation of gas at the inert anode which can cause defects on the wafer. Accordingly, engineering challenges remain to providing an inert anode electroplating processor.
In one aspect, an electroplating processor has a vessel having a first or upper processor compartment and a second or lower processor compartment with a processor anionic membrane between them. Catholyte (a first electrolyte liquid) is provided in the upper compartment above the processor anionic membrane. Anolyte (a second electrolyte liquid) is provided in the lower compartment below the processor anionic membrane and in contact with the processor anionic membrane. At least one inert anode is located in the second compartment in contact with the anolyte. A head holds a wafer in contact with the catholyte. The wafer is connected to a cathode, and the inert anode is connected to an anode, of a power supply.
A replenisher is connected to the vessel via catholyte return and supply lines and anolyte return and supply lines, to circulate catholyte and anolyte through first and second replenisher compartments in the replenisher separated by an anionic membrane. The replenisher adds metal ions into the catholyte by moving ions from a bulk metal source, such as copper pellets, into the catholyte in the first replenisher compartment. Simultaneously, anions, such as sulfate ions in the case of plating copper, move from the anolyte in the second replenisher compartment, through the anionic membrane, and into the catholyte in the first replenisher compartment. Ion concentrations in the catholyte and in the anolyte in the processor remain balanced.
In
The electroplating processor 20 may alternatively have various other types of head 22. For example the head 22 may operate with a wafer 50 held in a chuck rather than handling the wafer 50 directly, or the rotor and motor may be omitted with the wafer held stationery during electroplating. In some applications, a seal on the contact ring presses against the edge of the wafer 50 to seal the contact fingers 35 away from the catholyte during processing.
During processing, the head 22 is positioned over an electroplating vessel 38 of the electroplating processor 20. The vessel 38 is divided by an processor anionic membrane 54 into a first or upper processor compartment 36 above a second or lower processor compartment 52. A di-electric material membrane support 56 may be provided below, or above and below, the processor anionic membrane 54 to better hold the processor anionic membrane 54 in place.
The first processor compartment 36 is filled with a first electrolyte referred to as catholyte, with the catholyte in contact with the top surface of the processor anionic membrane 54. The second processor compartment 52 is filled with a second electrolyte referred to as anolyte, which is in contact with the bottom surface of the processor anionic membrane 54. One or more inert anodes 40 are provided in the vessel 38 in the lower compartment 52. A di-electric material field shaping element 44 is provided in the upper compartment 36 to shape the electric field in the catholyte during processing. A current thief electrode 46 near the top of the upper compartment 36 is connected to a second cathode current source which is selected to influence the electric field around the perimeter of the wafer 50.
Referring now to
The catholyte in the first processor compartment 36 circulates through the first replenisher compartment 62 via supply and return lines 80 and 82. The anolyte in the second processor compartment 52 circulates through the second replenisher compartment 66 via supply and return lines 84 and 86. The supply and return lines may connect to one or more intermediate pumps, filters, tanks or heaters. Tanks 92 may be provided to hold replenished anolyte and catholyte, with multiple electroplating processors 20 supplied from the tanks 92 rather than directly from the replenisher 60.
A source of bulk metal 68, such as copper pellets, is provided in the first replenisher compartment 62. The bulk metal 68 may be contained within a di-electric material holder 74 having perforated walls or made as an open matrix or screen, so that the bulk metal 68 is held in place while also exposed to the catholyte in the first replenisher compartment 62. The holder 74 generally holds the bulk metal 68 in a relatively thin layer, to increase the surface area of the bulk metal exposed to the catholyte. The holder 74 may be attached to a vertical side wall of the first replenisher compartment 62, opposite from the replenisher anionic membrane 64.
An inert cathode 70 is provided in the second replenisher compartment 66. Typically the inert cathode 70 is a metal plate or wire mesh, for example a platinum clad wire mesh or plate. The inert cathode may be attached to a vertical side wall of the second replenisher compartment 66, opposite from the replenisher anionic membrane 64. The bulk metal 68 is electrically connected to an anode current source of a power supply 72. The inert cathode 70 is electrically connected to a cathode current source of the power supply 72.
Multiple electroplating processors 20 may be provided in columns within an electroplating system, with one or more robots moving wafers in the system. A single replenisher 60 may be used to replenish the catholyte in multiple electroplating processors 20. The power supply 72 connected to the replenisher 60 is separate from, or separately controllable from, the power supply connected to the processors 20.
In use for electroplating copper, for example, the catholyte includes copper sulfate and water, and the bulk metal 68 is copper pellets. The head 22 is moved to place a wafer 50, or the device side of the wafer 50, into contact with the catholyte in the upper compartment 36 of the vessel 38. Electric current flows from the inert anode 40 to the wafer 50 causing copper ions in the catholyte to plate out onto the wafer 50. Water at the inert anode is converted into oxygen gas and hydrogen ions.
Sulfate ions move through the processor anionic membrane 54 from the catholyte in the first processor compartment 36 into the anolyte in the second processor compartment 52. To maintain the concentration of copper ions in the catholyte, the catholyte is circulated through the first replenisher compartment 62. To avoid a buildup of sulfate ions in the anolyte, the anolyte is circulated through the second replenisher compartment 66. Within the replenisher 60, electric current flows from the bulk metal through the catholyte, the replenisher anionic membrane 64 and the anolyte to the inert cathode, via power supply 72. Copper ions from the copper pellets, and sulfate ions from the anolyte, are replaced into the catholyte. As a result, the copper and sulfate ions in catholyte and in the anolyte remain balanced during processing.
As the inert cathode 70 is vertical, gas bubbles generated at the inert cathode 70 tends to rise to the top of the second replenisher compartment 66 and are removed. If necessary, the replenisher 60 may be temporarily disconnected from the processors 20, or turned off, e.g., for maintenance, while the processors continue to operate, as the metal ion and anion concentrations change gradually.
With a single replenisher 60 connected to e.g., 10 processors, the power requirements of the replenisher 60 may be significant. The replenisher 60 may be designed to minimize the spacing between the bulk metal 68 and the inert cathode 70, to reduce the voltage drop between them, which in turn reduces the power consumption of the replenisher 60. For example, with processors 20 for 300 mm diameter wafers, the processor anionic membrane 54 has a diameter nominally larger than 300 mm. The replenisher anionic membrane 64 may have a surface area 100% to 300% larger than the surface area of the processor anionic membrane 54. The dimension DD between the bulk metal 68 and the inert cathode 70 may be e.g., 10 to 25 cm, with the bulk metal 68 and/or the inert cathode 70 having a height of 150% to 300% of DD.
In an alternative design shown in
In contrast to other replenishment techniques, the present system and method uses only a single membrane in the processor and in the replenisher, a single catholyte, and a single anolyte, with no additional intermediate electrolytes or compartments needed. Hence, the replenisher requires only two compartments. As the anionic membranes prevent metal ions from passing, the system maintains a high level of efficiency. Although explained above in an example for electroplating copper, the present system and method may also be used to electroplate other metals as well.
Thus, novel systems and methods have been shown and described. Various changes and substitutions may of course be made without departing from the spirit and scope of the invention. The invention, therefore, should not be limited except by the following claims and their equivalents.
