UPW from wastewater

Published: 25-Jun-2002

Ton A.P.G. Roijackers, Jos G.M.A. Bouwman and Andreas Giesen of DHV Greenfab, The Netherlands asks if it is fearful or feasible?


Many regions all over the world suffer from shortage of potable water as a result of drought, pollution of groundwater, rivers and lakes or by dissipation. There is a worldwide need and requirement for a more sustainable use of fresh water, resulting in a trend towards a more integrated water resource management. Like all industries, the continuity of the microelectronics industry depends strongly on the availability of sufficient raw water. However, a major drawback for microelectronics industry is that raw water of a superb quality is required to be able to meet the stringent specifications for ultra pure water. Furthermore, the improved production technologies (300mm, reduced line width) demand for increasing strict requirements of both water availability and quality. In addition to this, there is an obvious clear trend in the European Union to draw up more significant stringent requirements to the discharged wastewater and the maximum allowable raw water intake. Above-mentioned trends result in the need for a more sustainable use of water and, in the long term, make it essential to reuse and recycle water. Moreover, although maybe at first glance apparently less important, water recycling also generates the following very important benefits:

  • Contribution to the 'clean and green' image;
  • Improved feed water quality (!);
  • Less dependence on local water resources and supply.
  • To give water recycling in European microelectronics a strong impulse, in 1999 and 2000 a research and innovation project focusing on zero-water discharge was executed. The MEDEA T650-project was jointly undertaken by a group of fabs, consultants, equipment suppliers and research institutes. The project was funded by the Dutch, Belgium, French, Swiss and German governments, while Dutch consultancy and engineering group, DHV 1), with specialists in water treatment and cleanroom design (facilities and utilities) was responsible for the project management. The major technical outcome of the project, communicated earlier in the magazine European Semiconductor, was as follows: Water recycling is feasible, up to a technological very high level of at least 95%, by using a smart wastewater segregation design in combination with a dedicated wastewater treatment set-up. In this paper DHV Greenfab describe the economics of combined treatment and reuse of wastewater from the UPW production (Fig. 1). The balance between the water recycling options and the economical feasibility is presented. We show the benefits generated by reduced water intake and discharge costs, will outbalance the investments and operational costs for additional treatments steps and recycled water quality monitoring and assurance.

    1 UPW
    1.1 General Ultrapure water (UPW) is critical for the semiconductor production process. Although, the International Technology Roadmap for Semiconductors has predicted a decrease of UPW usage per wafer, lately, semiconductor production facilities have rapidly expanded their wafer production capacities, and subsequently dramatically increased their water consumption while this trend is even further increased by the demand for larger wafer sizes. Therefore, "Zero-Discharge" options are becoming more and more important. Looking at Fig. 1 where a water system set-up according to the 'Zero Discharge' concept is presented, three main parts of the water system can be distinguished namely:

  • UPW production;
  • Semiconductor production process, where the UPW is used;
  • Wastewater treatment plant.
  • In case a true zero-discharge situation is achieved, no wastewater will be generated at all and the water production and usage has to be optimised. In this article, only the UPW production area is discussed. In "Do economy and water reuse fit together?", Cleanroom Technology, March 2002", Bouwman and Van Gool of DHV Greenfab discuss the economical evaluation of waste water segregation and reuse in the semiconductor industry. In several parts of a 'conventional' UPW-installation, precious water is lost due to the selected type of techniques and the necessary cleaning and regeneration procedures. This results in water lost of more then 150,000 m3/year. The parts of the installation in which the most water is lost are:
  • Multi Media Filters > 2000 m³/year
  • Softening > 20,000 m³/year
  • RO and EDI > 130,000 m³/year
  • The trend in UPW production set-ups is growing towards a reduction of wastewater discharge and an increase of the efficiency through reuse of used water. Different methods can be used to prevent the loss of this water. The determining factor for the selection of these methods is, next to the economical effects, the direct effect on the UPW quality.

    Economical aspects Fig. 2 reflects one example out of (many) possible UPW production alternatives resulting in very low water losses. The basic starting point is a conventional UPW installation. In this example, additional tools and equipment are used to reduce the loss of water to a minimum. The total efficiency on water savings can go up to 95% or even more, without any significant loss of UPW quality. The presented system comprises three internal recycle loops, which are added to the conventional UPW production line. The conventional installation (without water reuse) and the alternative one with various internal recycle loops (with water reuse) are compared and discussed on technical, technological and economical aspects.

    Recycle loop 1 The first recycle loop is for recycling the backwash water form the Multi Media filters. This recycle loop (see Fig. 3) is a discontinuous flow and contains a tank (40m³) for storage of the backwash water from the multi media filter and a centrifuge unit. During uptime of the MM-F the storage tank is processed over the centrifuge unit. The greater particles are removed from the water and the cleaned water is led back into the raw water tank.

    Recycle loop 2 As can be seen from Fig. 4, recycle loop 2 comprehends a buffer tank, a additional Reverse Osmosis unit and an Evaporation unit. Waste waters from the Softening unit, EDI and Mixed bed Ion Exchange are led over an additional Reverse osmosis unit. The wastewater from the RO is treated in an evaporation unit. The RO1 concentrate water is reused within the EDI unit.

    Recycle loop 3 Recycle loop 3 is without extra apparatus. The waste from the ultra filtration unit is led back to before the reverse osmosis unit. The economical consequences of the mentioned extraordinary optimisation of a UPW installation can be deduced from the diagram in Fig. 5. For a modern fab with quality demands in conformance with 150 nm technology and a production capacity of 520,000m³ UPW/year, the diagram presents the UPW cost price as a function of the total raw water cost price (water intake plus additional effluent costs). The applied economical basics are summarised in Table 1. In order to justify a comparison between both options, all costs are included in the resulting UPW production price, such as: fixed costs (based on an annuity of 24.4 %); variable costs (water loss, electricity, water heating, chemicals and replacements); raw water cost price and water discharge costs. The payback time of the additional equipment for internal recycle loops, as a function of the raw water cost price (intake + discharge) was calculated. Already at a raw water cost price of €0.50 the reuse option is in favour. As can be concluded from Fig. 5, with increasing raw water cost prices the reuse option becomes more favorable as it will lead to a further increase of annual savings (the difference in UPW price multiplied with the UPW production) and consequently a shorter pay back time for the 'extended' UPW installation (see Fig. 6) The extra costs for the required additional equipment (hardware costs) for adaptation into an installation with internal UPW water recycle loops, are calculated at approximately €350,000 (see Table 2). Fig. 6 shows that for the presented UPW optimisation, an attractive payback time of 2.5 year is already achieved at a raw water cost price of €1.50. It is obvious that at higher water prices recycling will become more and more attractive and the payback period will amount to less than 30 months. But one other great advantage of water recycling is increasing the UPW production amount without increasing the raw water intact. For example: an UPW installation for the production of approx 410,000 m³/year DI water can be expanded into an installation for the production of 520,000 m³/year DI water without even increasing the amount of raw water intake. As can be seen from Fig. 7 a saving in raw water intake of about 25 % is possible.

    Quality control But even more important than the water amount used in the semiconductor industry are the UPW water quality demands. In order to be able to predict the UPW quality for the installations, the quality parameters have been calculated, based upon basic performance of each of the unit operations. For both options, the conventional UPW installation and the 'improved' UPW installation the UPW quality are compared. In Table 3 and Fig. 7 the calculated UPW quality parameters for both options are presented. Fig. 7, presents a quality based comparison in which the values of the conventional installation are used as reference and set on 100 %. For the re-use options parameters such as ions, particles, silica, bacteria and hardness appear to be at least 15 % lower than in the conventional option. On the other side, TSS and TOC are about 15% higher for the Re-use option, but are absolutely still within acceptable ranges. The UPW quality from the calculations are perhaps not quite in accordance with the real quality, but the calculations can certainly be used to predict a tendency.

    1.2 Conclusions In this article only one of many possible options are shortly discussed. DHV Greenfab has developed a method to carry out basic calculations to predict the possibility (both economical and technical) of reuse within an UPW installation. This exercise can in principle be carried out for every possible UPW production system. Depending on the extensiveness of the installation and the raw water cost price, a balance between economical and technical demands can be found, most likely in favor of an extended or modified installation in which water is recycled and saved. The payback period for reusing the water in the UPW installation amounts to 30 month (at a raw water price of €1.50). Overall it can be stated that important savings on water usage and hence savings on water costs can very well be established within the "standard" UPW production installation, without loss of UPW quality and consequently, without affecting the high semiconductor wafer quality.

    You may also like