滑動式隔震平台應用於振動敏感性設備之地震模擬試驗

科技廠許多振動敏感性製程設備在中、小地震中極易受損，尤其是垂直爐管，這是震害損失主要來源之一。填滿晶圓片的石英晶舟連同外罩之石英管係以機械手臂置入機台定點後退出，因此無法鎖固於機台上，僅以自由站立的形式置於其中。由於晶舟與石英管屬細長型結構，高寬比約6:1，極易產生晃動。此外，因晶圓與石英間的摩擦係數很小，40 gal以上的樓板加速度峰值即可能造成滑片導致晶圓受損；當樓板加速度峰值達120 gal以上時，晶舟就可能與石英管碰撞，甚至傾倒，帶來巨大損失。有鑑於此，中、小地震下，垂直爐管隔震平台必須在樓板加速度達40 gal之前就能啟動，避免晶圓受損；在大地震中，隔震平台的加速度必須控制在100 gal以內，避免晶舟及石英管傾倒。此外，受限於特殊氣體管線的容許伸長量，在設計地震強度下，隔震平台的最大位移必須控制在18 cm之內，以確保安全；殘餘位移在3 cm以內，機台毋須重新定位。前述加速度控制目標，採滑動或滾軸式隔震支承均能達到；若要兼顧強震作用下的加速度與位移控制，則以滑動式支承較為可靠。本研究提出以摩擦單擺支承(Friction Pendulum Bearing, FPB)為隔震系統，其滑動介面為曲面，可於任意方向移動，並以滑動介子維持上部結構與水平面永遠正交。欲達到垂直爐管的防震設計目標，摩擦係數必須控制在3%以下，這須仰賴高精密的加工技術與適當的界面材料，所用的滑動材料必須滿足無塵室所要求的不揮發、耐腐蝕、耐磨耗及免維護等條件。針對垂直爐管的隔震應用，交通大學土木工程系與劦承精密股份有限公司合作開發了以摩擦單擺支承為基礎的隔震平台，並進行了一系列的振動台試驗，包括以正弦波來檢測隔震平台的啟動門檻，以Kobe震波來探討隔震平台在不同震度條件下的性能表現。測試結果顯示，隔震平台在尖峰樓板加速度(PFA)達21gal時即能啟動，界面摩擦係數μ≦0.03，當尖峰樓板加速度達747gal時，平台最大加速度僅 70 gal，無任何的晶圓滑落，最大位移僅為10.77cm，滿足性能設計目標。此外，在所有的測試中，隔震支承的最大殘餘位移(residual displacement)不逾0.42cm，遠低於容許值，證明各項指標都能符合規格要求，可實際應用於科技廠。

關鍵字

摩擦單擺支承；隔震平台；科技廠房；晶圓；垂直爐管

並列摘要

Many vibrant-sensitive process tools, in particular the vertical furnaces, in the high-tech fabs are seismically vulnerable to moderate earthquakes. This is one of the major seismic losses in the past earthquakes. Quartz boats filled with wafers along with the enclosing quartz tubes are handled with automatic robots. They cannot be fixed in place on the platform of the furnace and, as a result, can only be left free standing. Quartz boat and quartz tube are slender with an aspect ratio of 6 approximately, which makes them easy to be rocked. In addition, the friction between the wafer and quartz is so small that sliding and damage of the wafers could occur as the peak floor acceleration (PFA) reaches 40gal during earthquakes. As the PFA goes beyond 120gal, the quartz boat could collide with the inner tube and overturned even worse, leading to significant losses. Under this circumstance, the isolation platform for vertical furnaces needs to be activated before the PFA reaches 40gal to prevent damage of the wafers under moderate earthquakes. Whereas the peak acceleration of the isolation platform has to be kept within 100gal in strong earthquakes to avoid overturning of the quartz boat and tube. Moreover, restrained by the allowable extension distance of special gas lines, the maximum displacements of the isolation platform cannot go beyond 18cm under the design earthquake for safety reasons. The residual displacement has also to be kept within 3cm after earthquakes to avoid reposition of the vertical furnaces otherwise. The aforementioned control target on the acceleration can be achieved with both sliding or rolling-type isolation systems. However, if both the acceleration and displacement design criteria are to be met under strong earthquakes, the sliding-type system is more reliable. In this study, the friction pendulum bearing (FPB) is proposed for the seismic isolation platform. The sliding surface of the FPB is concave that allows the system to move in any direction. With an articulated slider in between, the isolation system can always keep the superstructure in an up-right position. To achieve the seismic design goals for vertical furnace, the frictional coefficient of the FPB must be controlled well within 3%. This would have to rely on both the high-precision processing technology and appropriate interface materials that must be non-volatile, corrosion-resistant, durable and maintenance-free as required by cleaning-room standards. In collaboration with Well-Link Co., the Dept. of Civil Engineering of NCTU has developed an FPB-based seismic isolation platform for the vertical furnaces. A series of shaking table tests has been conducted to verify the performance of the isolation platform, which includes excitations with sinusoidal waves to identify the triggering threshold (in PFA) of the isolation platforms, and Kobe earthquake in various intensities to explore the seismic performance of the isolation platform. The test results indicate that the isolation platform can be activated as the peak floor acceleration reaches as low as 21gal, implying a frictional coefficient of less than 3% between the sliding interfaces of the FPB. The maximum acceleration of the platform measured to be only 70gal without sliding of any wafer as the PFA of the input disturbance reaches 747gal, while the maximum sliding displacement of the platform is only 10.77cm. Moreover, in all tests, the maximum residual displacements recorded are no more than 0.42cm which is far less than the allowable range. Seismic performance of the proposed isolation system meets all the design criteria and the practical use of the system in high-tech factories is confirmed.