1、2010 ASHRAE 293ABSTRACTIn a traditional ventilation arrangement, a wall returnsystem containing a ceiling air supply and wall exhaust is usedin non-unidirectional airflow type cleanrooms. In industrialcleanrooms, such as semiconductor cleanrooms, however,larger process tools result in smaller passag
2、e of airflowstowards the wall-exhaust vents due to larger footprints andvolumes of process tools. Additionally, only sub-micron parti-cles are a concern for semiconductor cleanrooms (Hu and Wu,2003) and these particles are with very small magnitudes ofterminal velocities when moving in cleanroom air
3、 (Hinds.1999). The conventional wall return system used in existinghigh cooling load cleanrooms aims to create a uniform airtemperature distribution and a diluted particle concentrationenvironment in which the downward cold supply air from ceil-ing filters encounters upward air currents due to exhau
4、st heatas well as particle sources from process tools. To address theproblem, this study proposes a unique local air exhaust systemthat can significantly improve the air cleanliness level andremove heat loads efficiently. The main component is a Fan DryCoil Unit (FDCU) installed on the ceiling above
5、 the processtool in order to extract heat and particles released from thetool. The new proposed system is therefore referred as theFDCU system. An experimental study was conducted in a full-scale cleanroom with real process tools to compare the perfor-mance of both systems. Results show that more th
6、an 50% ofparticles were eliminated by the proposed FDCU system ascompared to that of the conventional wall return system. Theproposed FDCU system is also useful for solving problemssuch as a high negative static pressure in the supply chamberand high flow resistance in the flow path of the conventio
7、nalwall return system. The measured pressure in the supply cham-ber was 2 Pa (2.9 104psi) for the FDCU system, while thatof the wall return system was a negative value which variedproportionally to the air change rates of the cleanroom.INTRODUCTIONIn the traditional arrangement of the airflow path i
8、n acleanroom, shown in Figure 1, supply air (SA) is introducedfrom the ceiling and return air (RA) is located on the wall ata low level (IEST, 1995 and Whyte 2001). This system ishereafter referred to as a wall return system cleanroom. In thissystem, the direction of airflow is maintained consistent
9、 withthe movement of gravitationally settling particles (or bio-aerosols in a bio-cleanroom such as a pharmaceutical clean-room or hospital operating room). In this type of cleanroom,the return air shaft (RAS) occupies space in the cleanroomand increases the power input of fan-filter units (FFU) due
10、 toa long flow path to/from the FFU. Moreover, the return air-flow is significantly influenced by production tools anddevices, as well as the movements of operators. In summary,existing problems in the wall return system are as follows:(1) the external static pressure of FFU must be high (to over-co
11、me pressure resistance of the return air grill (RAG), RAS,and dry cooling coils (DCC), (2) the downward cold supplyair from the FFU encounters upward air currents due to theheat-load from process tools, and (3) fixed positions of theRAS and DCC. Installation of supply inlets and exhaust out-lets on
12、the ceiling conflicts with the traditional concept ofcleanroom airflow path arrangement. Murakami et al. (1992)indicated that changes in arrangement or in the number of ex-haust openings do not have a significant effect on the entireflow field; however, such changes often have a large influenceon th
13、e particle diffusion field (for 0.3 m particles) since theAn Innovative Ventilation System for Cleanrooms with High Cooling LoadsTi Lin Shih-Cheng Hu, PhDMember ASHRAEAndy Chang Cheng-Yan LinTi Lin and Andy Chang are PhD students, Shih-Cheng Hu is professor and Cheng-Yan Lin is MS student, in the De
14、partment of Energy andRefrigerating Air-Conditioning Engineering, National Taipei University of Technology, Taipei Taiwan.OR-10-031 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For per
15、sonal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 294 ASHRAE Transactionsparticle transportation path is changed by the position of theexhaust outlets. Murakami et al. (1989) also indicate
16、d that alocally balanced supply-exhaust airflow system (the supplyand exhaust airflow rates balanced locally within a flow unit)exhibits better exhaust particle performance than that of awall return type cleanroom. For industrial cleanrooms with ahigh cooling load, such as semiconductor cleanrooms f
17、orsemiconductor testing and for thin filming (with furnaceequipment), the cleanliness level must be maintained andtemperature controlled within a strict interval, and currently,the size of the particles of concern fall into the sub-micronregion due to advances in semiconductor manufacturing tech-nol
18、ogy. To solve the problems mentioned above, this studyproposes a unique local air distribution scheme that can main-tain the cleanliness level within the requirements while re-moving cooling load efficiently. The new proposed system(see Figure 2) is referred as a Fan Dry Coil Unit system, inwhich th
19、e supply inlet and exhaust outlet are installed in theceiling and heat is removed by a dry coil located just above theprocess tools. This study aims to evaluate the performance ofthe proposed system and to compare the system with the tra-ditional wall return system by experimental measurements in(a)
20、(b)Figure 1 (a) An existing wall return system and (b) the proposed FDCU system. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distributi
21、on, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. ASHRAE Transactions 295a full-scale cleanroom with a typical cooling load using asemiconductor testing machine.EXPERIMENTAL SET-UPA full-scale cleanroom was set up with dimensions of4.8 m 6
22、.3 m 2.8 m (15.7 ft 20.6 ft 9.2 ft) with two setsof dummy tools with dimensions of 1.6 m 1.2 m 2.4 m(5.25 ft 4.0 ft 7.9 ft) and one set of process tool (withdimension of 0.6 m 0.9 m 1 m (2.0 ft 3.0 ft 3.3 ft). Thisprocess tool is an semiconductor testing machine. Figure 3shows the domains studied, i
23、n which the FDCU system and thewall return system can be arranged in the same space. In thefigure, the dark gray areas show the outlets of the cleanroom,and the black area on the process tool represents the toolexhaust, which releases heat/particles to the cleanroom. Thecooling load is about 3.7 kW
24、(12,624 Btu/h), about 120 W/m2(38.26 Btu/hft2) floor area. The exhaust area, temperature, andvelocity were 0.15 m 0.15 m (0.5 ft 0.5 ft), 37C (98.6F)and 1500 m3/h (882.5 cfm), respectively. A 0.8 m 0.15 m (2.6ft 0.5 ft) suction inlet was located at the front and back sidesof the process tool. In the
25、 FDCU system, three FDCU unitswere installed with suction flow rate of 3240 m3/h (1905 cfm).Additionally, an ultrasonic vibrator located inside the tool wasused to generate particles using polystyrene latex sphere (PSL)as the particle source (injection velocity can be neglected) withmore than 85% of
26、 0.1 m diameter particles. Eventually, theconcentration produced at the exhaust of process tool wasapproximately 900,000 particles/ft3. A He-Ne laser particlecounter (with 1 cfm sampling flow rate) connected a manifoldwith 14 sampling point with 1 cfm sampling flow rate wereused to measure the parti
27、cle concentration in the measuringpoints. The locations of the 14 sampling points were based onthe airflow simulation results, which exhibit recirculationzones in the cleanroom. The wall return system has six outlets,represented by exhaust grills. Both schemes have 12 FFUs,which supply filtered air
28、to the cleanroom. The cleanroom wasmaintained with a positive static pressure of 12 Pa (17.5 104psi). Particle concentration monitoring was carried out atfourteen points with a multi-channel Laser Particle Counter.Figure 4 shows the locations of each point where a probe waslocate at two elevations,
29、i.e., 1.2 m (for points 1, 3, 5, 7, 9, 11,and 13) and 1.5 m (for points 2, 4, 6, 8, 10, 12, and 14).RESULTS AND DISCUSSIONFigure 5 shows the variation of ACH and FFU powerconsumption of the two systems. As shown in Figure 5. TheACH values of FDCU system exhibits slightly higher valuethan that of the
30、 wall return system for the same powerconsumption of the FFU. This is mainly due to a lower airflowresistance in the FDCU system.Figure 6 shows the variation of ACH over the normalizedconcentration ratio of the 14 measurement points. Points 1 and2 were behind the dummy tools and they were the furthe
31、staway from the process tool, which was a particle source.Therefore the normalized concentrations levels were thelowest at points 2 and 3. The normalized concentration inpoints 3 and 4 were high because they were between thedummy tools, where the width was only 1.1 m (3.63 ft).Recirculation zones we
32、re expected to be in these places. Points5 to 14 were located in the recirculation zones around theprocess tool. Figure 7 shows that the normalized concentrationof FDCU system was much lower than that of wall returnsystem in the same air change rate cases. In general, the aver-aged normalized concen
33、tration of the FDCU system was about25 70% lower than that of wall return system for all ACHFigure 2 (a) The lay-out of the cleanroom studied and (b)sampling probe locations of the particle counter.(a)(b) 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashr
34、ae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 296 ASHRAE Transactionscases. For the wall return system, the roo
35、m average normal-ized concentration were between 0.39 0.49. For the FDCUsystem, the room average particle concentration fell to 0.29 0.20. Therefore, the FDCU was essential for the particleremoval.For the wall return system, static pressure value in thesupply chamber was a negative and increased wit
36、h theincreases in ACH cases, while that of the FDCU system wasa positive, i.e., 2 Pa (2.9 104psi) (see Figure 8). It can benoted that a high negative static pressure in the supply cham-ber will infiltrate outdoor air and moisture, resulting an unsta-ble relative humidity control and increasing cooli
37、ng load.CONCLUSIONThis experimental work clarifies the effects of the clean-room exhaust outlet on sub-micron particle removal effi-ciency. For sub-micron particles, more particles wereexhausted by the proposed FDCU system than that by aconventional wall return system. The proposed FDCUsystem is als
38、o beneficial for solving problems such as highnegative static pressure in the supply chamber and high flowresistance in the flow path of the conventional wall returnsystem.REFERENCESIEST-PR-CC-12.1. 1995. Considerations in cleanroom de-sign, Institute of Environment Science and Technology.Cleanroom
39、Technology: Fundamentals of Design, Testingand Operation. 2001. W. Whyte. John Wiely and Sons.Hu, S.C. and T.M. Wu. 2003. Experimental studies of air-flow and particle Characteristics of a 300 mm POUP/LPU minienvironment system. IEEE Transactions onSemiconductor Manufacturing, 16(4):660-667.Hinds, W
40、.C. 1999. Aerosol Technology. 2nd ed. New York,John Wiley & Sons, Inc. 296.Murakami, S., S. Kato, and S. Nagano. 1992. Numericalstudy on diffusion in a room with a locally balancedsupply-exhaust airflow rate system. ASHRAE Transac-tions, pp. 218-238.Murakami, S., S. Kato, and Y. Suyama. 1989. Numeri
41、calstudy of diffusion field as affected by arrangement ofsupply and exhaust openings in conventional flow-typeclean room. ASHRAE Transactions 95(2):113-127.Figure 3 ACH and FFU power consumption of the twosystems.Figure 4 Normalized particle concentration for different ACH cases, ranging from 76 1/h
42、 to 136 1/h. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted
43、 without ASHRAEs prior written permission. ASHRAE Transactions 297Figure 4A (a) Room average concentration for different air change rate and (b) outdoor air infiltration in supply air plenum.(a) (b) 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.
