建立磁光科爾量測系統與製作鐵、鍺超薄膜之磁特性研究
Abstract
由廣義的Fresnel方程得知非鐵磁覆蓋層會影響鐵磁薄膜的磁光科爾效應(Magneto Optical Kerr Effect, MOKE)，其增強或減弱取決於覆蓋層的折射率；鍺(Ge)是半導體材料，但它的奈米顆粒結構被發現在室溫下具有鐵磁性，雖然產生鐵磁性Ge的機制仍不清礎，且其實驗的重複性也不佳，然而Ge/Fe膜層間的磁性現象會是一個有趣的課題，本文主要以縱向磁光科爾效應(LMOKE)來探討Ge/Fe與Fe/Ge的超薄膜分別在SiO2/Si(100) 行為是否如廣義Fresnel方程預測及從實驗數據反推Fe與Ge超薄膜在基板上的析射率。為達成研究目標，必先建立一套高靈敏的科爾磁光量測系統和一個超高真空熱蒸鍍系統來製作樣品。
科爾磁光量測系統的基本組件是一支波長為632.8nm的10mW氦氖雷射、二個Glan Thompson線偏振晶體（高消光比10-6)，一個置於雷射後來增強光的偏振及另一個置於光感測器前作檢偏用，光感測器則是一個矽基光二極體，為測量微弱的磁光訊號，光二極體與檢偏器間置入一個波長為632.8nm的帶通處波器來阻擋外界光害干擾，整個量測及數據擷取由電腦控制，此外訊號的訊噪比也可籍平均法來提高，在大氣下量測2.3nm鐵膜得到的磁光訊號，其訊噪比高於40；低2.3nm的鐵膜沒法量出磁光訊號，但這並不表示這個磁光儀的極限，而是小於2.3nm的鐵膜在空氣中氧化後不再是鐵磁性，又或其居禮溫度低於室溫。
熱蒸鍍系統的真空度高達 torr，並配有樣品交換腔及六個由烏線和氧化鋁坩堝組成的穩定蒸鍍源，可製作六種不同成份組成的薄膜，對於Fe及Ge，在 torr下製備的鍍率可控製在0.01~0.06Å/s，這樣的高真空和低鍍率是製作超薄膜的有利根基。
樣品有三系列：Fe(Xnm)/SiO2/Si(100)、Fe(Xnm)/Ge 30nm/SiO2 /Si(100)和Ge 4nm/Fe(Xnm)/SiO2/Si(100)，鐵膜厚度從1.5至15nm不等，其科爾旋轉角隨著鐵磁層厚度增加而變大，從理論計算印證實驗結果並推得Fe、Ge薄膜折射率分別為2.67+1.36i和5.25，且Fe、Ge相互覆蓋也無法增強磁光訊號現象。不同角度量測LMOKE得到的磁滯曲線顯示其磁性為各向同性，可知樣品的晶格結構為多晶體(polycrystal)磁滯曲線也透露各系列磁性膜有不同的磁翻轉過程， 覆蓋Ge能填補薄膜缺陷減少針扎基點(pinning center)，使磁壁位移容易，此現象將反應在磁滯曲線方正度及矯頑力上。
The magneto optical Kerr effect (MOKE) of magnetic multilayer can be described by generalized Fresnel equation that the refractive index of the non-magnetic capping has significant influence on MOKE signal. Besides, Ge belongs to the group of semiconductor, whose nano structure is found ferromagnetic even at room temperature. It is interesting to study ultra thin films of either Fe on Ge, or Ge on Fe by means of MOKE, and examine if the generalized Fresnel equation is still valid in ultra thin regime. In order to carry out this study, a MOKE measurement system and an ultra high vacuum thermal evaporator were built. MOKE system consists a 10 mW He-Ne laser, two Glan Thompson polarizers with extinction ratio of 10-6 , a photo-detector, and a electro-magnet. A 632.8nm band pass filter was insert between the photo-detector and analyzer, such that light pollution from environment can be eliminated. The MOKE measurement is conducted by computer, and data acquisition with signal averaging can improve the signal to noise (SNR) further. The SNR obtained from a 2.3nm Fe measured at atmosphere was better than 40. MOKE was not observed for Fe thickness below 2.3nm. However, this does not mean our MOKE system is not sensitive enough. This could be due to the oxidation of the Fe layer, or the Curie temperature of this thin Fe layer is below room temperature. The ultra high vacuum thermal evaporator with base pressure of better than 5*10-10 torr is equipped with six crucibles and a sample exchanged load lock. The deposition rates for Ge and Fe were 0.01 ~0.06 Å / s at torr. Fe (Xnm) / SiO2/Si (100), Fe (Xnm) / Ge 30nm/SiO2 / Si (100) and Ge 4nm/Fe (Xnm) / SiO2/Si (100) were made, in which X varies from 1.5 to 15nm. The MOKE signal is found to be proportional to the thickness of Fe. It is found that the refractive index of Ge , Fe are 2.67 +1.36 i, and 5.25, respectively, and Ge reduces the MOKE signal as predicted by generalized Fresnel equation. All samples are isotropic magnetically as revealed by angular resolved MOKE measurements. This could be due to the very random polycrystalline structure of the films. The hysteresis obtained from MOKE shows different magnetization reversal for different Fe thickness, and the capping Ge reduce the pinning center, and hence having smaller coercivity.
The magneto optical Kerr effect (MOKE) of magnetic multilayer can be described by generalized Fresnel equation that the refractive index of the non-magnetic capping has significant influence on MOKE signal. Besides, Ge belongs to the group of semiconductor, whose nano structure is found ferromagnetic even at room temperature. It is interesting to study ultra thin films of either Fe on Ge, or Ge on Fe by means of MOKE, and examine if the generalized Fresnel equation is still valid in ultra thin regime. In order to carry out this study, a MOKE measurement system and an ultra high vacuum thermal evaporator were built. MOKE system consists a 10 mW He-Ne laser, two Glan Thompson polarizers with extinction ratio of 10-6 , a photo-detector, and a electro-magnet. A 632.8nm band pass filter was insert between the photo-detector and analyzer, such that light pollution from environment can be eliminated. The MOKE measurement is conducted by computer, and data acquisition with signal averaging can improve the signal to noise (SNR) further. The SNR obtained from a 2.3nm Fe measured at atmosphere was better than 40. MOKE was not observed for Fe thickness below 2.3nm. However, this does not mean our MOKE system is not sensitive enough. This could be due to the oxidation of the Fe layer, or the Curie temperature of this thin Fe layer is below room temperature. The ultra high vacuum thermal evaporator with base pressure of better than 5*10-10 torr is equipped with six crucibles and a sample exchanged load lock. The deposition rates for Ge and Fe were 0.01 ~0.06 Å / s at torr. Fe (Xnm) / SiO2/Si (100), Fe (Xnm) / Ge 30nm/SiO2 / Si (100) and Ge 4nm/Fe (Xnm) / SiO2/Si (100) were made, in which X varies from 1.5 to 15nm. The MOKE signal is found to be proportional to the thickness of Fe. It is found that the refractive index of Ge , Fe are 2.67 +1.36 i, and 5.25, respectively, and Ge reduces the MOKE signal as predicted by generalized Fresnel equation. All samples are isotropic magnetically as revealed by angular resolved MOKE measurements. This could be due to the very random polycrystalline structure of the films. The hysteresis obtained from MOKE shows different magnetization reversal for different Fe thickness, and the capping Ge reduce the pinning center, and hence having smaller coercivity.
Description
Keywords
磁光科爾效應, Fresnel方程, Magneto Optical Kerr Effect, Fresnel equation