以暫態平面熱源快速量測材料熱性質之探討

Abstract

一般常用之熱傳導係數量測機台，需花長時間等待系統溫度達到穩態，再由已知間距量測點所測得試片之溫度梯度，以傅立葉熱傳導方程求出材料熱傳導係數，因此量測材料熱傳導值所需耗費時間較長，一般約需數十分鐘至數小時。且此穩態量測法其待測材料熱傳導係數與適合量測之試片厚度有相對應關係，對低熱傳導材料而言，此方法適於較薄試片量測，在厚試片量測上因等待系統達到穩態時間過長，存在嚴重熱損耗問題；對高熱傳導材料來說，此方法僅適於較厚試片量測，在薄試片量測上因各溫度測量點間之溫度差異小，於熱傳導值的換算易有較大誤差。 本論文研究探討以暫態熱傳導之特性基礎，自製一兼具加熱與溫度感測功能之平面雙螺旋熱源量測元件，將元件緊密夾於待測試片中，量測此暫態平面熱源其溫度隨時間之響應以推得待測材料之熱傳導係數及熱擴散係數。此暫態平面熱源量測技術著重在加熱時熱源與待測試片其溫度隨時間之變化關係，不需等待系統達到熱穩態，量測時間遠低於傳統穩態量測技術。此外，吾人亦探討此暫態平面熱源量測技術於大氣及抽真空環境下之量測差異，發現在抽真空環境下可有效隔絕外界環境對流對待測試片之影響，大幅提升量測準確度。 本暫態平面熱源量測系統在5.75秒內量測出紅銅(JIS C1100)之熱傳導係數370.6 W/m•k，與標稱平均值差異<5%；在3.5秒內量測出氮化鋁(AIN)材料之熱傳導係數188.5 W/m•k，與中科院量測值誤差<0.78%；在5秒內量測出矽晶圓之熱傳導係數146.2 W/m•k，與標稱平均值差異<7.47%；在4.6秒內量測出黃銅(JIS C2680)之熱傳導係數123.9 W/m•k，與標稱平均值差異<4.6%；在80秒內量測出不鏽鋼(Stainless steel 304)之熱傳導係數20.69 W/m•k，與標稱平均值差異<3.45%；在180秒內量測出壓克力板之熱傳導係數0.202 W/m•k，與標稱平均值差異<3.35%。本論文研究以低成本成功建立一快速暫態熱傳導係數量測系統，於大熱傳導係數範圍(0.202 ~ 370.6 W/m•k)量測中整體誤差低於7.47%、量測時間短於180秒，在學術研究上對未知熱傳導值材料來說提供一準確且快速之量測利器。

A thermal conductivity meter that we have used commonly requires a long time to allow system temperature to reach steady state. With the Fourier Heat Conduction Equation, we evaluate thermal conductivity by using the temperature gradient that we measured from the known measurement points of spacing. Therefore, it takes a longer period of time to measure the amount of the thermal conductivity which generally requires 10 minutes to couples of hours. By this, there is a correspondent relationship between the thermal conductivity of steady-state measurement method and the thickness of test specimens. For the lower thermal conductivity material, this solution is suitable for a thinner test specimen. It usually causes serious concerns of heat loss when a thicker test specimen spends too much time waiting for the system to reach the steady state. As for the higher thermal conductivity material, this method is better merely used on a thicker test specimen. In contrary, it would tend to have larger error in the value of thermal conductivity because the difference of temperature measurement points we gathered from a thinner specimen was small. In this essay, we study the characteristic foundation of transient heat conduction based on a self-made raw model composed with flat double-spiral heat source measurement which qualify both heating and temperature sensing functions. We make the components tightly clipped in between the test specimens which we use for testing later. After that, we evaluate the thermal conductivity and thermal diffusivity of the test specimens by measuring the temperature of the transient plane heat source which is reactivating with the passing of time. This measuring techniques of Transient Plane Source is concentrated on the changes between the heat source when heating and its own temperature changing over time. It takes less period of time than traditional steady-state measurement techniques and we don't have to wait for a system to keep heat stable. In addition, I examine the measurement difference of the measuring techniques of Transient Plane Source in the atmospheric and vacuum environment. My finding in this research discovers that the vacuum environment can efficiently isolate from the impact to the test specimens which is potentially influenced by external environment convection. With this discovery, we can significantly increase the measurement accuracy. This Transient Plane Source Measurement System I have worked on has been helping me to obtain the following data. My research measured from the Brass JIS C1100 within 5.75 seconds appeared that the value of thermal conductivity was 370.6 W/m•k and the normal average difference was less than 5%. As for the Aluminum Nitride, AIN, it took me 3.5 seconds to gather the value of thermal conductivity at 188.5 W/m•k which was less 0.78% different from the statistics recorded at the Chung-shan Institute of Science and Technology in Taiwan. The value of thermal conductivity I collected when tested on Silicon Wafer in 5 seconds was 146.2 W/m•k and the normal average difference was less than 7.47%. Spent 4.6 seconds to work on the Brass JIS C2680, I obtained the value of thermal conductivity at 123.9 W/m•k which was less 4.6% different from the nominal average. As for the value of thermal conductivity measured from Stainless Steel within 80 seconds, it was 20.69 W/m•k and the normal average difference was less 3.45%. When tested on a acrylic sheet within 180 seconds, I obtained the values of thermal conductivity at 0.202 W/m•k which was less 3.35% different from the nominal average. In this thesis, we have successfully established a efficient measuring system of transient heat conduction coefficient by taking lower cost efficiency. In the measurement of large thermal conductivity range from 0.202 to 370.6 W/m•k), the overall error was lower under 7.47% and the time of measurement was within 180 seconds. To the unknown material of thermal conductivity, this device provides the academic research a efficient, precise and time-saving measurement application.