Search

Article

x

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

Fabrication and thermoelectric properties of Si micro/nanobelts

Wei Jiang-Tao Yang Liang-Liang Wei Lei Qin Yuan-Hao Song Pei-Shuai Zhang Ming-Liang Yang Fu-Hua Wang Xiao-Dong

Fabrication and thermoelectric properties of Si micro/nanobelts

Wei Jiang-Tao, Yang Liang-Liang, Wei Lei, Qin Yuan-Hao, Song Pei-Shuai, Zhang Ming-Liang, Yang Fu-Hua, Wang Xiao-Dong
PDF
HTML
Get Citation
  • Currently, low-dimensional materials are a hot spot in the field of thermoelectric research, because the thermoelectric properties will be significantly improved after the low-dimensionalization of bulk materials. In a bulk material, its thermoelectric figure of merit ZT value cannot be increased by changing a single parameter, because the parameters of the material are interrelated to each other, which is not conducive to the research of internal factors and thus limiting the efficiency of thermoelectric material, but thermoelectric material on a micro-nano scale is more flexible to adjust its thermoelectric figure of merit ZT value. There are many different kinds of methods of implementing the low-dimensionalization of bulk materials. In this paper, size-controllable Si micro/nanobelts are prepared based on semiconductor micromachining and focused ion beam (FIB) technology, and the thermoelectric properties of Si micro/nanobelts of different sizes are comprehensively studied by the micro-suspension structure method.In this experiment, we find that the conductivity of doped Si micro/nanobelt is significantly better than that of bulk Si material, that as the width of the Si micro/nanobelt decreases, the thermal conductivity of the material decreases significantly, from 148 W/(m·K) of bulk silicon to 17.75 W/(m·K) of 800 nm wide Si micro-nanobelt, that the Seebeck coefficient of the material is lower than that of the corresponding bulkmaterials. The decrease of thermal conductivity is mainly due to the boundary effect caused by the size reduction, which leads the phonon boundary scattering to increase, and thus significantly inhibiting the behavior of phonon transmission in the Si material, thereby further affecting the transmission and conversion of thermal energy in the material. At 373 K, the maximum ZT value of the 800 nm wide Si micro/nanobelt reaches ~0.056, which is about 6 times larger than that of bulk silicon. And as the width of the Si micronanobelt is further reduced, the thermoelectric figure of merit ZT value will be further improved, making Si material an effective thermoelectric material. The FIB processing technology provides a new preparation scheme for improving the thermoelectric performances of Si materials in the future, and this manufacturing technology can also be applied to the low-dimensional preparation of other materials.
      Corresponding author: Wang Xiao-Dong, xdwang@semi.ac.cn
    [1]

    Zhou Y, Guo Z, He J 2020 Appl. Phys. Lett. 116 043904

    [2]

    袁国才, 陈曦, 黄雨阳, 毛俊西, 禹劲秋, 雷晓波, 张勤勇 2019 物理学报 68 117201

    Yuan G C, Chen X, Huang Y Y, Mao J X, Yu J Q, Lei X B, Zhang Q Y 2019 Acta Phys. Sin. 68 117201

    [3]

    邹平, 吕丹, 徐桂英 2020 物理学报 69 057201

    Zou P, Lv D, Xu G Y 2020 Acta Phys. Sin. 69 057201

    [4]

    Snyder G J, Toberer E S 2008 Nat. Mater. 7 105

    [5]

    Vining C B 2009 Nat. Mater. 8 83

    [6]

    Sales B C, Mandrus D, Williams R K 1996 Science 272 1325

    [7]

    Kim H S, Liu W, Chen G, Chu C W, Ren Z 2015 Proc. Natl. Acad. Sci. U.S.A. 112 8205

    [8]

    Goldsmid H J, Douglas R W 1954 Br. J. Appl. Phys. 5 386

    [9]

    Zhao H, Sun X, Zhu Z, Zhong W, Song D, Lu W, Tao L 2020 J. Semicond. 41 081001

    [10]

    Cai X, Han X, Zhao C, Niu C, Jia Y 2020 J. Semicond. 41 081002

    [11]

    Castenmiller C, Zandvliet H J W 2020 J. Semicond. 41 082003

    [12]

    Boukai A I, Bunimovich Y, Tahir-Kheli J, Yu J K, Goddard W A III, Heath J R 2008 Nature 451 168

    [13]

    Hochbaum A I, Chen R, Delgado R D, Liang W, Garnett E C, Najarian M, Majumdar A, Yang P 2008 Nature 451 163

    [14]

    Zhang Y, Su Q, Zhu J, Koirala S, Koester S J, Wang X 2020 Appl. Phys. Lett. 116 202101

    [15]

    Pettes M T, Jo I, Yao Z, Shi L 2011 Nano Lett. 11 1195

    [16]

    Liu H, Yang C, Wei B, Jin L, Alatas A, Said A, Tongay S, Yang F, Javey A, Hong J, Wu J 2020 Adv. Sci. 7 1902071

    [17]

    Shrestha R, Luan Y, Shin S, Zhang T, Luo X, Lundh J S, Gong W, Bockstaller M R, Choi S, Luo T, Chen R, Hippalgaonkar K, Shen S 2019 Sci. Adv. 5 eaax3777

    [18]

    Choe H S, Prabhakar R, Wehmeyer G, Allen F I, Lee W, Jin L, Li Y, Yang P, Qiu C W, Dames C, Scott M, Minor A, Bahk J H, Wu J 2019 Nano Lett. 19 3830

    [19]

    Choe H S, Li J, Zheng W, Lee J, Suh J, Allen F I, Liu H, Choi H J, Walukiewicz W, Zheng H, Wu J 2019 Appl. Phys. Lett. 114 152101

    [20]

    Park J, Bae K, Kim T R, Perez C, Sood A, Asheghi M, Goodson K E, Park W 2021 Adv. Sci. 8 2002876

    [21]

    Zhao Y, Zheng M, Wu J, Huang B, Thong J T L 2020 Nanotechnol. 31 225702

    [22]

    Madarasz F L, Lang J E, Szmulowicz F 1981 J. Electrochem. Soc. 128 2692

    [23]

    Wada H, Kamijoh T 1996 Jpn. J. Appl. Phys. 35 L648

    [24]

    Asheghi M, Kurabayashi K, Kasnavi R, Goodson K E 2002 J. Appl. Phys. 91 5079

    [25]

    Li D, Wu Y, Kim P, Shi L, Yang P, Majumdar A 2003 Appl. Phys. Lett. 83 2934

    [26]

    Alaie S, Goettler D F, Jiang Y B, Abbas K, Baboly M G, Anjum D H, Chaieb S, Leseman Z C 2015 Nanotechnol. 26 085704

    [27]

    Shrestha R, Li P, Chatterjee B, Zheng T, Wu X, Liu Z, Luo T, Choi S, Hippalgaonkar K, de Boer M P, Shen S 2018 Nat. Commun. 9 1664

    [28]

    Alaie S, Goettler D F, Abbas K, Su M F, Reinke C M, El-Kady I, Leseman Z C 2013 Rev. Sci. Instrum. 84 105003

    [29]

    Li D, Wu Y, Fan R, Yang P, Majumdar A 2003 Appl. Phys. Lett. 83 3186

    [30]

    Mavrokefalos A, Pettes M T, Zhou F, Shi L 2007 Rev. Sci. Instrum. 78 034901

    [31]

    Roh J, Hippalgaonkar K, Kang J, Lee S, Ham J, Chen R, Majumdar A, Kim W, Lee W 2010 3rd International Nanoelectronics Conference (INEC) January 3–8, 2010, Hong Kong, China p633

    [32]

    Lee S, Yang F, Suh J, Yang S, Lee Y, Li G, Sung Choe H, Suslu A, Chen Y, Ko C, Park J, Liu K, Li J, Hippalgaonkar K, Urban J J, Tongay S, Wu J 2015 Nat. Commun. 6 8573

    [33]

    Shi L, Li D Y, Yu C H, Jang W Y, Kim D, Yao Z, Kim P, Majumdar A 2003 J. Heat Transfer 125 881

    [34]

    An T H, Lim Y S, Park M J, Tak J Y, Lee S, Cho H K, Cho J Y, Park C, Seo W S 2016 APL Mater. 4 104812

    [35]

    Tang J, Wang H T, Lee D H, Fardy M, Huo Z, Russell T P, Yang P 2010 Nano Lett. 10 4279

    [36]

    Wingert M C, Chen Z C Y, Dechaumphai E, Moon J, Kim J H, Xiang J, Chen A R 2011 Nano Lett. 11 5507

    [37]

    Haras M, Lacatena V, Morini F, Robillard J F, Monfray S, Skotnicki T, Dubois E 2014 IEEE International Electr on Devices Meeting (IEDM) December 15–17, 2014, San Francisco, CA, USA p8.5.1

    [38]

    Lim J, Wang H T, Tang J, Andrews S C, So H, Lee J, Lee D H, Russell T P, Yang P 2016 ACS Nano 10 124

    [39]

    Holland M G 1963 Phys. Rev. 132 2461

  • 图 1  Si微/纳米带的制备流程 (a) SOI晶圆片; (b) 顶层硅刻蚀; (c) 埋氧层刻蚀; (d) 气态HF释放

    Figure 1.  Si micro/nanobelt preparation process: (a) SOI wafer; (b) top silicon etching; (c) buried oxide layer etching; (d) gaseous HF release.

    图 2  FIB切割过程 (a) 制造过程中的横截面图; (b) 悬浮Si的SEM图

    Figure 2.  Process of the FIB milling: (a) Cross sectional view of fabrication process; (b) SEM image of the released freestanding Si.

    图 3  微悬空结构的制备流程 (a) LPCVD生长低应力氮化硅; (b) 金属剥离制备蛇形电阻和引出电极; (c) RIE刻蚀氮化硅; (d) 悬空结构的最终释放

    Figure 3.  Micro-suspension structure preparation process: (a) LPCVD growth of low-stress silicon nitride; (b) metal stripping to prepare serpentine resistors and lead electrodes; (c) RIE etching of silicon nitride; (d) release of final suspended structure.

    图 4  悬空结构的SEM图 (a) 悬空结构整体的SEM图; (b) 悬空部分在52° 倾角下的近景SEM图; (c) 微设备的伪彩色SEM图

    Figure 4.  SEM images of the suspended structure: (a) SEM image of the whole suspended structure; (b) close-up SEM image of the suspended part at 52° inclination; (c) false colour SEM picture of the microdevice.

    图 5  (a) 纳米探针与Si微/纳米带接触过程; (b) 通过Pt金属焊接, 将Si微/纳米带从原本的位置转移走; (c) 纳米探针把Si微/纳米带转移到悬空结构上的过程; (d) 通过Pt金属焊接, 将Si微/纳米带固定在悬空岛两端; (e) 样品1, Si微/纳米带的宽度为2000 nm; (f) 样品2, Si微/纳米带的宽度为800 nm

    Figure 5.  (a) Contact process between nanoprobe and Si micro/nanobelt; (b) transfer the Si micro/nanobelt from its original position by Pt metal welding; (c) process of transferring the Si micro/nanobelt to the suspended structure by the nanoprobe; (d) fix the Si micro/nanobelt on both ends of the suspended island by Pt metal welding; (e) Sample 1, where the width of the Si micro/nanobelt is 2000 nm; (f) Sample 2, where the width of the Si micro/nanobelt is 800 nm.

    图 6  (a) Pt电阻随温度的变化; (b) Pt电阻随温度的相对变化, R0为303 K时Pt电阻的阻值

    Figure 6.  (a) Change of Pt resistance with temperature; (b) relative change of Pt resistance with temperature, where R0 is the resistance value of the Pt resistance at 303 K.

    图 7  (a) 样品电阻随温度的变化; (b) 样品电阻率随温度的变化

    Figure 7.  (a) Change of samples’ resistance with temperature; (b) change of sample resistivity with temperature.

    图 8  (a) 引线电阻随温度的变化; (b) 样品塞贝克系数随温度的变化. Seebeck系数的不确定性为5%

    Figure 8.  (a) Change of wires’ resistance with temperature; (b) change of samples’ Seebeck coefficient with temperature. The uncertainty of Seebeck coefficient is 5%.

    图 9  (a) 样品在不同温度下的热导率值, 插点为文献值; (b) 样品在不同温度下的ZT值, 其中热导率和ZT值的不确定性分别为8%和13%

    Figure 9.  (a) Thermal conductivity value of the sample at different temperatures, where the interpolation point is literature values; (b) ZT value of the samples at different temperatures. The uncertainty of thermal conductivity and ZT value are 8% and 13%, respectively.

    表 1  不同样品的尺寸参数

    Table 1.  Size parameters of different samples.

    宽度/nm厚度/nm长度/µm
    Sample 120002205
    Sample 28002203
    DownLoad: CSV
  • [1]

    Zhou Y, Guo Z, He J 2020 Appl. Phys. Lett. 116 043904

    [2]

    袁国才, 陈曦, 黄雨阳, 毛俊西, 禹劲秋, 雷晓波, 张勤勇 2019 物理学报 68 117201

    Yuan G C, Chen X, Huang Y Y, Mao J X, Yu J Q, Lei X B, Zhang Q Y 2019 Acta Phys. Sin. 68 117201

    [3]

    邹平, 吕丹, 徐桂英 2020 物理学报 69 057201

    Zou P, Lv D, Xu G Y 2020 Acta Phys. Sin. 69 057201

    [4]

    Snyder G J, Toberer E S 2008 Nat. Mater. 7 105

    [5]

    Vining C B 2009 Nat. Mater. 8 83

    [6]

    Sales B C, Mandrus D, Williams R K 1996 Science 272 1325

    [7]

    Kim H S, Liu W, Chen G, Chu C W, Ren Z 2015 Proc. Natl. Acad. Sci. U.S.A. 112 8205

    [8]

    Goldsmid H J, Douglas R W 1954 Br. J. Appl. Phys. 5 386

    [9]

    Zhao H, Sun X, Zhu Z, Zhong W, Song D, Lu W, Tao L 2020 J. Semicond. 41 081001

    [10]

    Cai X, Han X, Zhao C, Niu C, Jia Y 2020 J. Semicond. 41 081002

    [11]

    Castenmiller C, Zandvliet H J W 2020 J. Semicond. 41 082003

    [12]

    Boukai A I, Bunimovich Y, Tahir-Kheli J, Yu J K, Goddard W A III, Heath J R 2008 Nature 451 168

    [13]

    Hochbaum A I, Chen R, Delgado R D, Liang W, Garnett E C, Najarian M, Majumdar A, Yang P 2008 Nature 451 163

    [14]

    Zhang Y, Su Q, Zhu J, Koirala S, Koester S J, Wang X 2020 Appl. Phys. Lett. 116 202101

    [15]

    Pettes M T, Jo I, Yao Z, Shi L 2011 Nano Lett. 11 1195

    [16]

    Liu H, Yang C, Wei B, Jin L, Alatas A, Said A, Tongay S, Yang F, Javey A, Hong J, Wu J 2020 Adv. Sci. 7 1902071

    [17]

    Shrestha R, Luan Y, Shin S, Zhang T, Luo X, Lundh J S, Gong W, Bockstaller M R, Choi S, Luo T, Chen R, Hippalgaonkar K, Shen S 2019 Sci. Adv. 5 eaax3777

    [18]

    Choe H S, Prabhakar R, Wehmeyer G, Allen F I, Lee W, Jin L, Li Y, Yang P, Qiu C W, Dames C, Scott M, Minor A, Bahk J H, Wu J 2019 Nano Lett. 19 3830

    [19]

    Choe H S, Li J, Zheng W, Lee J, Suh J, Allen F I, Liu H, Choi H J, Walukiewicz W, Zheng H, Wu J 2019 Appl. Phys. Lett. 114 152101

    [20]

    Park J, Bae K, Kim T R, Perez C, Sood A, Asheghi M, Goodson K E, Park W 2021 Adv. Sci. 8 2002876

    [21]

    Zhao Y, Zheng M, Wu J, Huang B, Thong J T L 2020 Nanotechnol. 31 225702

    [22]

    Madarasz F L, Lang J E, Szmulowicz F 1981 J. Electrochem. Soc. 128 2692

    [23]

    Wada H, Kamijoh T 1996 Jpn. J. Appl. Phys. 35 L648

    [24]

    Asheghi M, Kurabayashi K, Kasnavi R, Goodson K E 2002 J. Appl. Phys. 91 5079

    [25]

    Li D, Wu Y, Kim P, Shi L, Yang P, Majumdar A 2003 Appl. Phys. Lett. 83 2934

    [26]

    Alaie S, Goettler D F, Jiang Y B, Abbas K, Baboly M G, Anjum D H, Chaieb S, Leseman Z C 2015 Nanotechnol. 26 085704

    [27]

    Shrestha R, Li P, Chatterjee B, Zheng T, Wu X, Liu Z, Luo T, Choi S, Hippalgaonkar K, de Boer M P, Shen S 2018 Nat. Commun. 9 1664

    [28]

    Alaie S, Goettler D F, Abbas K, Su M F, Reinke C M, El-Kady I, Leseman Z C 2013 Rev. Sci. Instrum. 84 105003

    [29]

    Li D, Wu Y, Fan R, Yang P, Majumdar A 2003 Appl. Phys. Lett. 83 3186

    [30]

    Mavrokefalos A, Pettes M T, Zhou F, Shi L 2007 Rev. Sci. Instrum. 78 034901

    [31]

    Roh J, Hippalgaonkar K, Kang J, Lee S, Ham J, Chen R, Majumdar A, Kim W, Lee W 2010 3rd International Nanoelectronics Conference (INEC) January 3–8, 2010, Hong Kong, China p633

    [32]

    Lee S, Yang F, Suh J, Yang S, Lee Y, Li G, Sung Choe H, Suslu A, Chen Y, Ko C, Park J, Liu K, Li J, Hippalgaonkar K, Urban J J, Tongay S, Wu J 2015 Nat. Commun. 6 8573

    [33]

    Shi L, Li D Y, Yu C H, Jang W Y, Kim D, Yao Z, Kim P, Majumdar A 2003 J. Heat Transfer 125 881

    [34]

    An T H, Lim Y S, Park M J, Tak J Y, Lee S, Cho H K, Cho J Y, Park C, Seo W S 2016 APL Mater. 4 104812

    [35]

    Tang J, Wang H T, Lee D H, Fardy M, Huo Z, Russell T P, Yang P 2010 Nano Lett. 10 4279

    [36]

    Wingert M C, Chen Z C Y, Dechaumphai E, Moon J, Kim J H, Xiang J, Chen A R 2011 Nano Lett. 11 5507

    [37]

    Haras M, Lacatena V, Morini F, Robillard J F, Monfray S, Skotnicki T, Dubois E 2014 IEEE International Electr on Devices Meeting (IEDM) December 15–17, 2014, San Francisco, CA, USA p8.5.1

    [38]

    Lim J, Wang H T, Tang J, Andrews S C, So H, Lee J, Lee D H, Russell T P, Yang P 2016 ACS Nano 10 124

    [39]

    Holland M G 1963 Phys. Rev. 132 2461

  • Citation:
Metrics
  • Abstract views:  399
  • PDF Downloads:  42
  • Cited By: 0
Publishing process
  • Received Date:  26 April 2021
  • Accepted Date:  16 May 2021
  • Available Online:  07 June 2021
  • Published Online:  20 September 2021

Fabrication and thermoelectric properties of Si micro/nanobelts

    Corresponding author: Wang Xiao-Dong, xdwang@semi.ac.cn
  • 1. Engineering Research Center for Semiconductor Integrated Technology, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China
  • 2. Research Center of Materials and Optoelectronics, College of Microelectronics, University of Chinese Academy of Sciences, Beijing 100049, China
  • 3. Beijing Institute of Quantum Information Science, Beijing 100193, China
  • 4. Beijing Semiconductor Micro/Nano Integrated Engineering Technology Research Center, Beijing 100083, China

Abstract: Currently, low-dimensional materials are a hot spot in the field of thermoelectric research, because the thermoelectric properties will be significantly improved after the low-dimensionalization of bulk materials. In a bulk material, its thermoelectric figure of merit ZT value cannot be increased by changing a single parameter, because the parameters of the material are interrelated to each other, which is not conducive to the research of internal factors and thus limiting the efficiency of thermoelectric material, but thermoelectric material on a micro-nano scale is more flexible to adjust its thermoelectric figure of merit ZT value. There are many different kinds of methods of implementing the low-dimensionalization of bulk materials. In this paper, size-controllable Si micro/nanobelts are prepared based on semiconductor micromachining and focused ion beam (FIB) technology, and the thermoelectric properties of Si micro/nanobelts of different sizes are comprehensively studied by the micro-suspension structure method.In this experiment, we find that the conductivity of doped Si micro/nanobelt is significantly better than that of bulk Si material, that as the width of the Si micro/nanobelt decreases, the thermal conductivity of the material decreases significantly, from 148 W/(m·K) of bulk silicon to 17.75 W/(m·K) of 800 nm wide Si micro-nanobelt, that the Seebeck coefficient of the material is lower than that of the corresponding bulkmaterials. The decrease of thermal conductivity is mainly due to the boundary effect caused by the size reduction, which leads the phonon boundary scattering to increase, and thus significantly inhibiting the behavior of phonon transmission in the Si material, thereby further affecting the transmission and conversion of thermal energy in the material. At 373 K, the maximum ZT value of the 800 nm wide Si micro/nanobelt reaches ~0.056, which is about 6 times larger than that of bulk silicon. And as the width of the Si micronanobelt is further reduced, the thermoelectric figure of merit ZT value will be further improved, making Si material an effective thermoelectric material. The FIB processing technology provides a new preparation scheme for improving the thermoelectric performances of Si materials in the future, and this manufacturing technology can also be applied to the low-dimensional preparation of other materials.

    • 热电(thermoelectric, TE)材料可以有效将热能直接转化为电能, 也可以通过电流用来对指定区域制冷. 热电器件具有无噪声、无振动、可靠性高、能够长期稳定工作和尺寸灵活的独特优点[1,2]. 热电材料的性能通常由无量纲的热电优值ZT来描述, ZT = S 2σT/κ, 其中S, σ, κT分别为塞贝克系数、电导率、热导率和绝对温度[3]. 最大化ZT值是困难的, 因为在块体材料中, S, σκ之间相互依赖, 相互制约[4]. 由于不能单一调节热参数, 块体材料的热电性能受到了严重的约束, 大多数体材料的ZT值都在1附近, 相应的能量转化效率约为10%或更低[5]. 经过理论计算, 当ZT值达到3时, 可以进行大规模的热电应用; 当ZT值达到4时, 热电之间的能量转换效率将高达30%[6-8].

      硅是地球上最丰富的半导体. 但是, 块体Si材料并不是一种理想的热电材料, 主要原因是其具有高的导热系数(κ = 148 W/(m·K), 300 K), 这导致了极低的热电转化效率. 研究表明, 块体材料低维化可以显著改善材料的性能[9-11]. 当块体Si被制备成直径远小于大部分声子平均自由程(phonon mean free path, MFP)的Si纳米线时, 热导率减小为块体的1/150左右, 使Si的热电优值从0.01 (块体Si)显著提高到了1左右(200 K), 与块体硅相比增大了100倍[12]; 相似地, Hochbaum等[13]发现表面粗糙的Si纳米线也可以显著抑制其热导率, 使其减小到约1.6 W/(m·K), 在室温下可以获得约0.6的热电优值. Si纳米线热电性能的显著提高, 主要原因是热导率的显著降低. 然而, Si纳米线具有一定的局限性. Si纳米线的机械强度较弱, 且热电性能受到样品表面形貌和直径的严重影响, 这不利于重复、大规模的应用.

      针对Si纳米线存在的问题, 本文报道了一种基于半导体微加工和聚焦离子束(focused ion beam, FIB)技术的快速制备方法, 成功制备出了尺寸可控的悬浮Si微/纳米带. 这些悬浮的Si微/纳米带表现出了良好的机械强度和低的热导率, 同时具备高的电导率, 这些优点使它成为TE设备的候选材料. 热电参数的测量是通过微悬空结构完成的, 该悬空结构是为了纳米线、纳米管和纳米膜的热参数测量而开发的[14-21]. 本文的实验结果表明, 对于800 nm宽的Si微/纳米带, 热导率值可低至17.75 W/m K. 热导率的降低主要是由于声子的边界散射, 边界散射效应显著抑制了Si材料中声子的热传输. 在不显著牺牲热电功率因子“S 2σ” 的情况下, Si微/纳米带在373 K时的ZT值达到了约0.056, 显著优于块体Si材料.

    2.   实验部分
    • 实验的起始材料为双抛的SOI (silicon on insulator)晶圆片. 器件层厚度为(220 ± 10) nm; 埋氧层的厚度为3 µm; 底层硅厚度为500 µm. 埋氧层相对较厚, 主要目的是为了后续悬空. 微加工制备之前, 首先用标准的RCA工艺对SOI晶圆片进行清洗. 微加工制备过程主要包括2次光刻、2次干刻蚀工艺、2次气态HF干法腐蚀和1次电子束蒸发生长金属Ti过程. 图1给出了Si微/纳米带制造过程的示意图. 图1(b)为顶层硅在深反应离子刻蚀(deep reactive ion etching, DRIE)后的形貌. DRIE完成后, 再进行一次套刻, 套刻的目的是为了把暴露出来的埋氧层完全刻蚀掉, 埋氧层的刻蚀是通过电感耦合等离子体技术, 刻蚀埋氧层的目的是为了减少气态HF释放的时间, 刻蚀后的示意图如图1(c)所示. 在制备过程中, 最关键的一步就是使Si微/纳米带完全悬空, 该过程采取了气态氟化氢(hydrogen fluoride, HF)释放而不是溶液湿法腐蚀, 是为了避免在释放过程中Si微/纳米带与衬底粘连, 气态HF释放后的结果如图1(d)所示.

      Figure 1.  Si micro/nanobelt preparation process: (a) SOI wafer; (b) top silicon etching; (c) buried oxide layer etching; (d) gaseous HF release.

    • Si微/纳米带制备成功之后, 接下来要通过FIB Ga+ 进行切割, 制备出不同宽度的Si微/纳米带. 在FIB Ga+离子切割之前, 需要对制备好的悬浮Si微/纳米带进行一些特殊的处理. 如图2(a)所示, 在FIB切割之前需要在悬浮的Si微/纳米带上生长一层薄的保护层, 该保护层能够最大限度地减少FIB系统中Ga+离子在切割过程对Si微/纳米带造成的掺杂影响, Ga+离子的掺杂会同时影响Si微/纳米带的电特性[22]和热性能[23,24]. 由于本文的目的是确定Si微/纳米带的宽度对其热导率、电导率和塞贝克系数的影响, 因此最小化外界因素带来的不确定影响是至关重要的, 所以在FIB Ga+切割之前必须要对悬浮Si微/纳米带进行相应的保护措施.

      Figure 2.  Process of the FIB milling: (a) Cross sectional view of fabrication process; (b) SEM image of the released freestanding Si.

      保护层的选择取决于很多因素, 比如: 实验过程中保护层必须相对容易添加和除去, 不能给实验造成困难和影响; 保护层的制备需要与样品的制备过程兼容, 保护层的引入和除去不能对待测样品造成影响, 造成实验误差; 保护层的厚度也有严格的限制, 保护层必须相对较薄, 如果保护层生长得太厚, 那么Ga+离子束将很难穿透保护层和下面的硅层, 无法正常完成对样品的切割加工; 还要考虑保护层的导电性问题, 如果保护层是绝缘的, 带电的Ga+离子轰击样品的表面时会受到严重的影响, 样品表面的电荷无法及时导出, 造成扫描电子显微镜(scanning electron microscope, SEM)图像的漂移, 无法对切割区域精准定位, 造成严重的切割偏差. 根据以上分析, 本次实验选取Ti金属作为切割的保护层, 如图2(a)所示. 在FIB Ga+切割之后, 再一次通过气态HF除去剩余的Ti保护层. 图2(b)分别给出了FIB Ga+离子切割前后的Si微/纳米带的SEM图.

    • 起始材料为(100)晶面的Si片, 厚度约375 µm. 在制备之前, 同样使用标准的RCA工艺对样片进行清洗. 微悬空结构的制备流程如图3所示. 主要包括: 3次光刻工艺、2次金属剥离工艺、1次低压化学气相沉积(low pressure chemical vapor deposition, LPCVD)氮化硅生长、1次等离子体增强化学气相沉积(plasma enhanced chemical vapour deposition, PECVD)氮化硅生长、2次反应离子(reactive ion etchin, RIE)干法刻蚀和1次四甲基氢氧化铵(tetramethylammonium hydroxide, TMAH)溶液湿法腐蚀.

      Figure 3.  Micro-suspension structure preparation process: (a) LPCVD growth of low-stress silicon nitride; (b) metal stripping to prepare serpentine resistors and lead electrodes; (c) RIE etching of silicon nitride; (d) release of final suspended structure.

      图3(a)所示, 首先需要在Si晶圆片表面采用LPCVD生长500 nm厚的SiNx薄膜, 作为后续的绝缘层和悬臂梁. 之后需要两次金属剥离工艺, 用来制备Pt蛇形电阻和Au引出电极. Cr/Pt蛇形电阻和Cr/Au引出电极都是通过磁控溅射法制备的, 厚度分别为5 nm/50 nm, 20 nm/200 nm, 剥离后的金属如图3(b)所示. 在金属制备完成之后, 将其放入PECVD腔室内约10 min, 在金属顶部生长约350 nm厚的SiNx保护层, 这样做的目的是为了防止在后续的TMAH腐蚀过程中Pt蛇形电阻和Au引出电极的脱落. 之后通过光刻和RIE技术定义氮化硅悬臂和悬岛的相对位置和具体尺寸, 刻蚀之后如图3(c)所示. 在氮化硅刻蚀完成之后, 将样片放入质量分数约25% 的TMAH溶液中进行湿法腐蚀, 反应温度控制在80 ℃. 腐蚀约6 h之后, 将其取出放置在IPA溶液中进行干燥, 这样可以避免悬空结构的断裂.

      制备完成的微悬空结构SEM图如图4所示, 悬空结构由两个相邻的500 nm厚的低应力SiNx悬岛组成, 悬岛的尺寸约为80 μm × 120 μm. 每个SiNx悬岛通过8个400 μm长、10 μm宽的悬臂与衬底相连. 每个SiNx悬岛上都长有50 nm厚的Pt蛇形电阻温度计, Pt蛇形电阻通过沉积在长氮化硅悬臂上的4条200 nm厚、5 μm宽的Au引线与衬底上150 μm × 150 μm Au键合Pad相连. 其中两根引线向Pt蛇形电阻提供电流, 另外两根引线用于测量Pt蛇形电阻上的电压降. 通过简单测量, 悬空结构下的沟槽深度约180 μm, 在一个四寸Si片上可同时制备出将近1000个微悬空结构.

      Figure 4.  SEM images of the suspended structure: (a) SEM image of the whole suspended structure; (b) close-up SEM image of the suspended part at 52° inclination; (c) false colour SEM picture of the microdevice.

    • 微悬空结构制备完成后, 需要将待测样品转移到悬空结构上, 使待测样品连接两个孤立的悬空岛, 以便测量其热电性能. 目前比较普遍的转移方法有液滴法[25]、纳米探针法等[21,26,27]. 液滴法转移有很大的局限性, 例如: 桥接两个相邻悬浮SiNx岛的样品是随机的, 充满了不可控性, 而且转移之后的微悬空结构特别脏; 纳米探针法可以精准地转移不同尺寸的样品, 具有很好的可控性, 转移之后的MEMS微悬空结构非常干净, 并且整个转移过程是在FIB高真空腔室内完成的, 不会造成额外的外界污染. 基于以上分析, 本文选择了纳米探针干法转移. Si微/纳米带转移的整个过程SEM图如图5(a)(d)所示.

      Figure 5.  (a) Contact process between nanoprobe and Si micro/nanobelt; (b) transfer the Si micro/nanobelt from its original position by Pt metal welding; (c) process of transferring the Si micro/nanobelt to the suspended structure by the nanoprobe; (d) fix the Si micro/nanobelt on both ends of the suspended island by Pt metal welding; (e) Sample 1, where the width of the Si micro/nanobelt is 2000 nm; (f) Sample 2, where the width of the Si micro/nanobelt is 800 nm.

      图5(a)给出了经过FIB Ga+ 切割加工之后的样品和纳米探针, 在转移过程中, 纳米探针需要精准地移动到悬空Si微/纳米带附近, 然后移动纳米探针与Si微/纳米带接触, 接触时纳米带会发生轻微的振动, 但是这些振动不足以使Si微/纳米带断裂. 接触之后, 通过Pt气体注入系统(gas injection system, GIS)进行焊接, 为转移做准备. 在纳米探针与Si微/纳米带焊接完成之后, 对Si微/纳米带两端进行切割, 如图5(b)所示. 在FIB切割过程中要小心, 尽量减少暴露的硅带被额外的离子注入, 因为这样可能会改变样品的热导率. 为了避免FIB切割过程中离子的注入问题, 采取最小的电流, 约为30 pA对Si悬空微/纳米带两端进行切割. 2013年, Alaie等[28]证明了通过FIB切割和转移样品之后, 不会对样品造成可检测的损伤. 在样品切割之后, 通过移动纳米探针使样品移动到微悬空岛的上方, 为接下来搭载做准备, 如图5(c)所示. 移动纳米探针将Si微/纳米带放置在悬空岛的两端, 并在局部沉积Pt金属, 将Si微/纳米带固定在悬空岛上. 当这一步完成后, 可以将纳米探针取出, 而使Si微/纳米带保持在悬空岛上, 如图5(d)所示. 其中Pt金属沉积可以改善Si微/纳米带与悬空岛引出电极之间的热接触[16,17,29-32]. 通过FIB纳米探针转移了两种不同尺寸的样品, 如图5(e)图5(f)所示, 以下简称为样品1和样品2. 在样品1和样品2热电参数测试完成之后, 通过SEM精确测量了样品的尺寸, 具体尺寸参数如表1所列.

      宽度/nm厚度/nm长度/µm
      Sample 120002205
      Sample 28002203

      Table 1.  Size parameters of different samples.

    3.   结果与讨论
    • 为了测量单个Si微/纳米带的热电参数, 需要将转移有Si微/纳米带的微悬空结构置于低温恒温器中. 所有的测量都是在小于6 × 10–6 Torr ($ 1\;\mathrm{T}\mathrm{o}\mathrm{r}\mathrm{r}\approx 133.322\;\mathrm{P}\mathrm{a} $)的高真空中进行的, 以消除热对流和热辐射损失, 使Si微/纳米带成为加热端悬岛和和测量端悬岛之间唯一的热传输通道. 利用微悬空结构电热法测量热电参数的细节已经在之前的文献中被详细报道[25,29]. 简单地说就是通过测量悬岛之间的热传输来表征Si微/纳米带的热导率.

      在进行热电参数测量之前, 首先需要对Pt蛇形电阻进行温度校准. 高温端和低温端Pt蛇形电阻的阻值测量都采用四线法. 温度变化范围为303—373 K, 图6给出了Pt蛇形电阻阻值随温度变化的结果. 实验测得高温端和低温端的Pt蛇形电阻温度系数分别为TCRH = 1.46 × 10–3 K–1TCRC = 1.14 × 10–3 K–1. 实验报道的体Pt的温度系数为3.927 × 10–3 K–1, 高于50 nm厚的磁控溅射的Pt金属薄膜, 体Pt与薄膜Pt金属之间的差异可以归因于不同的制备方法和材料形貌.

      Figure 6.  (a) Change of Pt resistance with temperature; (b) relative change of Pt resistance with temperature, where R0 is the resistance value of the Pt resistance at 303 K.

    • 待测样品的电阻率可以通过与样品相接触的4个Pt电极来测量, 这样可以消除寄生压降, 大大提高电阻率的测量精度. 图7为两种不同尺寸的Si微/纳米带在300—380 K温度范围内的电阻和电阻率的测量数据, 可以看出, 随着温度的上升, 电阻和电阻率都在逐渐增加. 由图7(a)可知, 样品1的电阻值低于样品2的电阻值, 即随着样品宽度的减小, 样品的电阻增大. 由图7(b)可知, 样品2的电阻率基本等于样品1的电阻率, 这主要是由于样品宽度从2 µm减小到800 nm时, 带来的尺寸效应不足以影响样品中电子的正常输运, 所以电阻率未受到影响. 随着温度的不断升高, 样品晶格振动散射增加, 导致迁移率的下降. 样品的迁移率和电阻率的表达式分别为

      Figure 7.  (a) Change of samples’ resistance with temperature; (b) change of sample resistivity with temperature.

      其中q为电荷量, p为载流子的浓度, m*为载流子的有效质量, Ni为载流子的浓度, AB为晶格振动和电离杂质散射的系数, µp为载流子的迁移率. 由(1)式和(2)式可知, 电阻率会随着温度的升高而增大, 这与图7(b)中电阻率的变化趋势是一致的.

    • 通过Pt蛇形电阻可以精确测量待测样品两端的温差, 由于温差的存在, 会在样品两端产生热电压, 即VTE = (SSSPt)/(THTC), 其中SPt为Pt金属的Seebeck系数, 由于在300 K时, SPt的大小通常为5 µV/K, 并且随着温度升高线性减小[33], 因此, 可以近似认为SSSPt $ \approx $ SS. 分别测量出热电压, THTC就可以得到待测样品的Seebeck系数. 图8为两种不同尺寸的Si微/纳米带在300—380 K温度范围内的塞贝克系数的测量数据. 通常, 塞贝克系数用Mott关系来描述[34]:

      Figure 8.  (a) Change of wires’ resistance with temperature; (b) change of samples’ Seebeck coefficient with temperature. The uncertainty of Seebeck coefficient is 5%.

      其中, kB是玻尔兹曼常数, h是普朗克常数, ${m}_{\mathrm{d}}^{*}$是态密度有效质量, p是空穴的浓度, T是温度. 塞贝克系数不仅受载流子浓度的影响, 还受态密度有效质量和温度的影响. 从图8可以看出, 样品1和样品2的塞贝克系数相差不是很大, 主要原因是样品1和样品2除了尺寸不同, 其他完全一致; 样品的塞贝克系数数值随着温度的上升具有上升的趋势. 实验中测量所得的最大Seebeck系数约为133.5 µV/K, 低于Tang等[35]所测的Seebeck系数, 主要原因是本实验所测Si微/纳米带的掺杂浓度(约3 × 1020 cm–3, P-type)比Tang等[35]所测的高1个数量级, 由(3)式可知, 随着掺杂浓度的升高, Seebeck系数会逐渐减小.

    • 除了电阻率和Seebeck系数外, 还需要测定Si微/纳米带的热导率. 微悬空结构电热法测量热导率的基本原理是, 给热端Pt蛇形电阻施加恒定的直流电流I, 热端悬空岛温度由T0升高到TH, 产生的焦耳热通过中间搭载的Si微/纳米带传到低温端悬岛, 使其温度从T0升高到TS. 通过测量Pt蛇形电阻计算对应的温度值, 进而计算出Si微/纳米带的热导率. 在热导率测试过程中, 会首先改变腔室的温度, 待腔室温度稳定之后再通电流进行相应的测试, 测试过程主要使用了Keithley2440数字源表和SR830锁相放大器. 由于MEMS微悬空结构电热法在测量过程中与衬底分离, 隔离了众多热传输通道, 显著地提高了样品热导率的测量精度, 微悬空结构电热法是目前低维材料热导率测试精度最高的方法[36].

      微悬空结构电热法测量低维材料热导率的过程中, 样品与衬底的接触热阻会严重影响测量结果. 在本文的实验中, 通过FIB GIS在Si微/纳米带与电极接触点的顶部沉积了Pt金属以加强电接触, FIB沉积可以显著降低接触热阻使接触热阻最小化[33], 接触热阻仅占样品热阻的10%或者更小, 所以在测试过程中认为所测总热阻本质上等于Si微/纳米带的热阻. 图9(a)所示为样品1和样品2在300—380 K温度范围内的热导率的测量数据. 图9(a)中的插点为文献报道的Si热导率值, 分别为56 [Haras, 2014][37], 10.23 [Tang, 2010][35]和1.8 W/(m·K)[Lim, 2016][38], 这些热导率值相对于块体Si有了显著的降低. 与块体硅的热导率相比, 本文测量的Si微/纳米带的最小热导率值减小为其1/9, 并且随着Si微/纳米带宽度的减小, 相应的热导率值逐渐减小. 这表明边界散射增强对Si微/纳米带中的声子传输具有很大的影响. 随着温度逐渐升高, Si微/纳米带的热导率先缓慢增加后快速减少, 并在340 K附近达到最大值. 在373 K时, 热导率值降到最小, 样品1和样品2的最小热导率值分别为40.9和17.75 W/(m·K). 为了理解热导率К的温度依赖性, 使用马蒂森规则来描述各种散射机制下的声子弛豫时间[39]. 总声子散射率τ –1可以近似表示为

      Figure 9.  (a) Thermal conductivity value of the sample at different temperatures, where the interpolation point is literature values; (b) ZT value of the samples at different temperatures. The uncertainty of thermal conductivity and ZT value are 8% and 13%, respectively.

      其中${\tau }_{\mathrm{P}\mathrm{h}\text-\mathrm{P}\mathrm{h}}$, $ {\tau }_{\mathrm{i}\mathrm{m}\mathrm{p}\mathrm{u}\mathrm{r}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{e}\mathrm{s}} $$ {\tau }_{\mathrm{b}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{y}} $分别是受Umklapp散射、杂质散射和边界散射限制的声子弛豫时间. 由于Si微/纳米带是通过相同的SOI晶圆制作的, 因此Si微/纳米带中杂质散射的影响并不会发生变化, 类似于整体散射. 然而, 来自Si微/纳米带边界的散射确实会影响整个声子的散射时间. Umklapp散射的特点是热导率随着温度的升高而减小. 由于声子在与其他声子散射之前会受到边界的阻碍, 因此在较高的温度下边界散射变得非常重要. 对于Si微/纳米带, Umklapp散射和杂质/边界散射之间的相互竞争产生的热导率峰值发生在340 K附近, 如图9(a)所示.

      由测试可知, 材料的热导率、电导率、塞贝克系数和温度有关. 把每个参数测量值代入ZT = S 2σT/κ公式中, 就可以计算出材料在不同温度下的热电优值ZT. 图9(b)分别给出了样品1和样品2在不同温度下的热电优值ZT. 通过图9(b)可知, 样品1和样品2的热电优值ZT随着温度的升高而增加, 并且样品2的热电优值远大于样品1, 主要原因是随着Si微/纳米带宽度的减小, Si微/纳米带的电阻率和塞贝克系数几乎不变, 但是, 热导率发生了显著的降低, 导致了热电优值ZT的显著增加. 在373 K时, 样品2的最高热电优值达到了0.056, 相比体硅提高了将近6倍.

    4.   结 论
    • 通过FIB技术制备了尺寸可控的高质量Si微/纳米带, 并通过纳米探针转移到微悬空结构上进行了热电性能的详细测试. 实验发现, 随着Si微/纳米带宽度的减小, 材料的热导率发生了显著的降低. 对于800 nm宽的Si微纳米带, 在373 K下测得的热导率κ约为17.75 W/(m·K), 此值远低于块体Si的热导率. 该结果表明, Si微/纳米带的热导率被强声子边界散射所抑制. 在不显著牺牲热电功率因子“$ S^2\sigma $” 的情况下, 800 nm宽的Si微/纳米带在373 K时的ZT值达到了约0.056. 我们认为, FIB快速加工制备Si微/纳米带的技术, 不仅为将来高热电性能Si材料的制备提供了新的方案, 也为其他材料低维化提供了新的思路.

Reference (39)

Catalog

    /

    返回文章
    返回