Technology News

A Node To Semiconductor Manufacturing

In the past 7 years of semiconductor manufacturing the industry went through a significant technology node. An increase in size of the silicon wafer, to 300 mm, was adopted to increment productivity. Concurrently the logistics of wafer lot movements between process tools was completely automated, for the sake of productivity, and due to size and weight of the wafer lot carrier (FOUP). Here Middlesex Industries discusses a static node solution for problems incurred during wafer lots transport.

With respect to logistics, manufacturers of semiconductors adopted a policy to cooperate in standardising the logistics hardware, and chose a technology based on individual vehicles transporting wafer lots between process steps.

Such decision to standardise had its benefits as well as its drawbacks. The drawback was that such a single technology could not satisfy all the requirements for modern transport needs in an efficient manner. This forced the suppliers of the vehicle technology to adopt costly solutions, within that technology, in an attempt to comply with performance needs. This trend today continues. And, has significant effects on over all factory productivity (Asset utilisation and factory cycle time).

Incurred costs can be significant because factory assets (clean facilities and process equipment in the billions of dollars) are very expensive. An example of the limitations of vehicle technology is in their capacity to deliver. Inherent with that system design is the limitation on vehicle utilisation. Each vehicle can carry one wafer lot, and each vehicle is scheduled for sequential pick up and delivery. In this process, it cannot be avoided that vehicles sometimes run empty, thus limiting the systems over all utilisation or capacity to deliver.

Cause and effects
The adverse effect of such a limitation cascades down to the utilisation of the semiconductor process equipment, as well as to the time (cycle time) it takes for the semiconductor product to transit through all factory processes to finish. To counter act these problems with the same technology calls for more vehicles, and ever increasing speed for those vehicles. Both solutions are extremely expensive.

In addition to being expensive, the solution also has its practical limit. The more vehicles are employed on a track the lower their speed must be, due to the essential mode of stop and go operations. The current invention side steps the limitations of the discrete vehicles by introducing a different transport technology (conveyors) into vehicle logistics, which technology does not have such capacity limitations. Such a combination of the two transport technologies extends the performance of the discrete vehicle solution to meet the throughput requirement of the modern factory. Analysis of transport requirements in the clean semiconductor environment shows a spatially and sequentially chaotic distribution of work in process lot destinations. The process is recursive, due to process equipment costs. It cannot be arranged as a linear, sequential production line. To satisfy the lot transport requirements in this environment is difficult with discrete vehicles travelling from source to destination, and thus the use of vehicles implies its corollary, the employment of static transfer and storage devices. Such devices allow the break up of the transport system into several closedloop lines, where the static transfer devices serve as nodes of cross over paths between the loops.

The ups & downs
This static node solution works with limitations. It reduces the necessity of non-stop transports from all sources to all destinations, but adds the cost of additional summed dwell times of vehicles, and wafer lots, and so cuts into over all transport capacity. The cost of the solution is also very high. Transfer stockers with their robotic handling devices, and additional vehicle requirements are costly technology. The current invention improves on the above with dynamic transfer nodes. These nodes are capable of relocating the wafer lots from downstream to upstream, and vice versa, thus increasing the amount of time a vehicle is utilised, and thus the amount of separate move tasks a vehicle can perform. The net result is a higher throughput for the same discrete vehicle logistics. Additional benefits are: a reduction of wafer lot storage requirement, which also reduces cycle time of production; the new possibility of inserting dynamic transfer points within a vehicle transit loop, which increases the utilisation of each vehicle within that loop; the reduction in the number of hardened transit nodes, or the reduction of the number of loops required, which increases system configuration flexibility; the introduction of the capability of one vehicle delivering more then one wafer load in a circuit of its transit loop, which is theoretically equivalent of the extension of its load carrying capacity beyond the just one load at a time. And all of the above is achieved with a very significant reduction in equipment cost. The logic used in the above may be compared to that of airline hub, where an airliner docks at one terminal to disembark the passengers, who then are moved to an other gate or terminal to pick up the continuation of their flight, instead of waiting at the same arrival gate for an other plane. And in other instances to a bus route, where the bus makes a stop by virtue of which it is able to carry more then one passenger to a single destination, and thus the equipment is utilised more on each transit of the route. Embodiment of the above ideas can be made in several instances when added to an existing monorail vehicle system design.