US DOE publishes solid state lighting manufacturing roadmap

USA: Last week, the US Department of Energy (DOE) published the Solid-State Lighting Manufacturing R&D Roadmap, the outcome of two workshops and considerable dialog between manufacturers, suppliers, researchers and other organizations involved in expanding solid state lighting (SSL). The purpose of the roadmap was to guide the US DOE R&D program, help act as a guide for equipment and material suppliers, and reduce risk and manufacturing cost.

The document can be downloaded from the DOE.

SSL manufacturing roadmap priorities
The manufacturing roadmap focuses on identifying priority needs for achieving cost-effective, high quality manufacturing capabilities for SSL components ranging packaged devices, replacement lamps, and complete luminaires, in both OLED and LED technologies. In high-brightness LED packages, the report focuses on four key manufacturing needs:

* The need to advance the epitaxy process and equipment to improve wavelength uniformity and reproducibility, reduce variations in chip output power, increase throughput, and improve yields.

* The need to address substrate-related issues, ranging from warping and defects in present materials to cost and availability of potentially better approaches such as native substrates.

* The lack of suitable manufacturing equipment. Increased automation should be introduced into the wafer processing, die packaging and testing activities. A lower cost-of-ownership for all equipment is required.

* The inadequacy of process controls. There is a need for improved inspection and testing equipment in a number of areas including epitaxial growth, wafer processing and die packaging throughout the process. Active feedback control for critical process steps such as epitaxial growth is also required.

Epitaxial growth is the key enabling technology for HB-LEDs. Several critical issues regarding epitaxial growth equipment and processes were identified, including wavelength uniformity, low throughout, lack of in-situ monitoring, and managing wafer bow.

All GaN-based HB-LED epiwafers are currently manufactured using MOCVD, the only technology currently capable of growing all layers of the device structure. The primary drawback of MOCVD is the relatively slow growth rate, resulting in long cycle times (typically 5-10 hours). Actions to increase the growth rate, reduce the overall cycle time, or expand the reactor capacity are required to raise throughput.

Hydride Vapor Phase Epitaxy (HVPE) is a potential alternative growth method which has the advantage of significantly higher growth rates and offers the prospect of much higher throughputs.

The substrate roadmap supports two paths: improved substrates for heteroepitaxial growth (sapphire and SiC), and improved substrates for homoepitaxial growth (GaN). In both cases, the report concludes that “improvements in substrate quality (surface finish, defect density, flatness, etc.) and product consistency are required in order to meet the demands of high volume manufacturing. For GaN substrates cost must also be dramatically reduced in order to become a viable option for LED manufacturing.”

Better cost of ownership model
During the DOE workshops, there was a general agreement that equipment improvement was required to advance process control, manufacturing throughput and yield. The need for an accurate cost model was the main take-away from these discussions with many participants saying that such a model would have much wider application in identifying areas in the packaged LED process which had the largest impact on ultimate device costs.

Such a model would allow the community to identify equipment and processes lying on the critical path and offer a more quantitative assessment of the beneficial cost impact of addressing each issue.

The total cost of ownership (COO) is the widely used in the semiconductor industry (see SEMI standard E35 ‘Cost of Ownership for Semiconductor Manufacturing Metrics’) and its applicability to LEDs may merit further discussions between manufacturers and suppliers. COO can be defined as the full cost of embedding, operating and decommissioning a system needed to accommodate a required volume.

It is the total cost of producing a good part from a piece of equipment, which is obtained by dividing the full cost of the equipment and its operation by the total number of good parts produced over the commissioned lifetime of the equipment. In addition to new equipment purchases and manufacturing process changes, COO can help with decisions about materials use, equipment operations and process improvements. It can help identify any bottlenecks in the process, and it can foster communication and understanding throughout the supply chain.

The DOE has announced intentions to continue the manufacturing workshops in 2010. It is expected that many of these issues, including the development of a better cost model, will be discussed in greater detail.

Source: SEMI, USA