【作者】 李驰麟；

【导师】 傅正文；

【作者基本信息】 复旦大学 ， 物理化学， 2008， 博士

【摘要】 全固态薄膜锂电池一般是由正极薄膜、电解质薄膜和负极薄膜三部分组成,其薄膜材料与传统电池的粉体材料相比有其特殊性,并不是传统材料薄膜化的简单照搬,本文就是围绕着基于全固态薄膜锂电池的正极和电解质薄膜材料来展开论述的。对于正极薄膜材料,我们最初的思路是把传统的粉体材料薄膜化,并进行掺杂改性,以满足其在薄膜电池中的实际应用需求。首先,我们通过射频磁控溅射法成功制备了纳米晶的纯锂锰氧（LMO）薄膜,其在3V和4V区域表现出明显且对称的电化学活性,这与LiMn₂O₄粉体的电化学特性是一致的,但是由于在全固态环境中没有Mn²⁺离子的溶解,Li/LiPON/LMO薄膜电池在大的电位范围2-4.8V内有很好的循环稳定性,可逆容量为78μAh/cm²-μm。但是LMO在充放电过程中的具有1V电位突跃的两个平台被认为在实际应用中是不利的,所以我们通过在LMO中掺入过量ZrO₂以制备成LMZO薄膜来改性,ZrO₂由于其低的热传导性、化学惰性和强的硬度,被认为是可以抑制高温退火或长期循环后LMO薄膜晶粒生长的合适材料。Li/LiPON/LMZO薄膜电池表现出完全不同的电化学行为,消除了LMO两个独立平台间的1V突跃,但LMZO仍然具有与LMO一样的1.7个Li的插入量,可逆容量可维持在53μAh/cm²-μm。事实上,对于大多数的正极材料,为了获得具有高电化学活性的理想晶型,高温合成过程是不可避免的。但是对于我们的薄膜电池来说,高温过程不利于薄膜电池制备流程与一般的微电子工艺兼容,而且晶态结构薄膜的表面颗粒较大,表面较粗糙,不利于与固体电解质的界面匹配。所以我们希望开发出能低温沉积制备,且无需高温退火处理的非晶态正极薄膜,以满足大部分微电子器件有源集成的潜在需要,而且非晶态特点使薄膜具有平滑表面,可以保证薄膜电池具有长的循环寿命和平滑的电极/电解质界面,避免薄膜电池在长时间电化学过程中的界面退化甚至短路。其实,由于退火过程中电极表面晶粒生长的有效控制,上述LMZO薄膜已经体现出比LMO更平滑的电极/电解质界面,但是高温处理的避免和非晶态薄膜形式的目标并没有实现。受非晶态FePO₄粉体材料具有较好电化学活性的启发,我们首次用射频磁控溅射制备不非晶态的FePO₄和掺氮磷酸铁（FePON）正极薄膜,并分别将它们集成到全固态薄膜锂电池中。低温制备的FePO₄薄膜被证明具有更大的体积比容量,但是由于动力学限制,可逆容量只能维持在20μAh/cm²-μm左右。而在N₂下制备的FePON薄膜可得到大得多的体积比容量,首次放电容量为63μAh/cm²-μm,是FePO₄容量的3倍左右,从循环伏安曲线看,FePON容量的大幅提高是由于其在1.7V和2.4V区域出现一对额外的CV峰引起的。另外,拉曼光谱和XPS说明了FePON中N插入可能破坏了PO₄^3-聚阴离子框架。为了进一步考察N插入对薄膜电化学行为的改善,我们从动力学角度进行了系统研究,分别制备了基于FePON和FePO₄薄膜的三明治结构体系和全固态电池体系,并采用交流阻抗技术进行定量分析。从三明治结构可以得到,FePON和FePO₄薄膜的离子导电率分别为2.5×10^-8S/cm和7.5×10^-10S/cm,两者的电子导电率则分别为2.6×10^-8S/cm和2.7×10^-6S/cm,可以看出,FePON的离子和电子导电率更加趋于平衡,与FePO₄相比,掺氮使薄膜离子和电子导电率发生相反变化,这很可能是由N掺入产生了O缺陷造成的。从不同温度下的离子导电率可以得出FePON和FePO₄薄膜的活化能均为0.72eV,而电极/电解质界面的活化能相对更小,为0.25eV,暗示着非晶态薄膜层间的致密接触,有利于Li⁺在界面的传输。在全固态薄膜电池结构中,由于N的掺入,电池在不同电位下的电荷传输电阻大幅度下降,Li⁺在FePON和FePO₄薄膜中的平均扩散系数分别为1.7×10^-11 cm²/s和1.2×10^-13 cm²/s,说明了N掺杂有利于Li⁺在固态电极中的扩散。另一方面,我们知道传统的正极材料往往只有一个反应活性中心,当然也有一些材料具有多个活性中心,但是这些材料或者分子量较大,或者充放电时易发生结构坍塌,这些都会限制材料的质量比容量。但是对于应用于薄膜电池的薄膜正极材料来说,体积比容量才是最重要的性能指标,所以我们可以考虑一些虽然分子量大,但具有多反应活性中心的物质,以制备成高密度的致密薄膜来提高薄膜电池的体积比容量。本文我们选择了基于活性聚阴离子框架（WO₄）^2-的CuWO₄和LiFe（WO₄）₂作为多活性中心正极薄膜,它们都由射频磁控溅射方法制备。首先,经XRD和SEM表征,可认为三斜相的CuWO₄薄膜由20nm尺寸的纳米晶粒组成。Li/CuWO₄液态模拟电池的首次放电容量可达192mAh/g,在2.5V和1.6V区域可以分别观察到对应于Cu²⁺/Cu⁰和W⁶⁺/W⁴⁺的两个独立放电平台,但是可能由于在电化学过程中电极/电解液界面恶化或Cu离子溶解,CuWO₄薄膜电极会经历结构的变化及容量的衰减。基于CuWO₄在不同充放电状态下的TEM和SAED结果,我们提出了一个涉及取代反应的两步反应机理,认为（WO₄）^2-框架在还原分解后可重新形成,原来很好晶型的CuWO₄在电化学循环后可维持不变。然后,我们把CuWO₄薄膜正极成功组装到了以LiPON为固体电解质的全固态薄膜锂电池中,其首次放电容量可达145μAh/cm²-μm,由于在全固态体系中不存在界面恶化和Cu离子溶解等问题,所以在后续循坏中电极结构和电化学曲线变化不大。非晶态LiFe（WO₄）₂正极薄膜可同时实现上述多反应活性中心设计和低温沉积制备两大目标。Li/LiFe（WO₄）₂液态模拟电池首次放电容量可达198mAh/g,3V和1.5V区域的两个独立放电平台分别对应Fe³⁺/Fe²⁺和W⁶⁺/W^x+（x=4或5）,300次循环后,可逆放电容量仍可维持在110mAh/g左右。基于不同电位下的XPS和TEM测量,可以认为与Fe和W有关的两个氧化还原中心是电化学可逆的,且在最初的一些循环中非晶态的LiFe（WO₄）₂薄膜有逐步结晶化的趋势,从而导致1-2V电位区域内容量的连续衰减。LiFe（WO₄）₂薄膜的另一大优势则体现在其在大电流密度下的容量维持性和稳定性,其在50μA/cm²下的体积比容量稳定在60μAh/cm²-μm左右。固体电解质薄膜作为薄膜电池成败的关键,其研究显得尤为重要。目前可供选择的薄膜电解质类型并不是很多,而当前应用最为广泛的无机电解质薄膜材料是具有较高锂离子导电率、良好电化学稳定性的含氮磷酸锂薄膜（LiPON）。为了进一步提高LiPON的离子导电率,需要提高LiPON中N的掺入量,这就需要设法改进薄膜的沉积技术,我们就是通过在磁控溅射装置中加入电子回旋共振（ECR）等离子体来辅助沉积薄膜,以提高LiPON的含N量和离子导电率。在ECR功率为200W时N/P比最佳,为0.65,由交流阻抗谱分析,此条件下LiPON薄膜的电导率约为8^*10^-6S/cm。另外,传统的制备LiPON薄膜的射频磁控溅射技术存在沉积速率低、沉积面积小等缺点,且LiPON薄膜本身易潮解,所以我们采用电子束蒸发这一新方法来制备锂镧钛氧（LLTO）薄膜以取代LiPON,这一方法可大大增加薄膜沉积的效率。为了制备高性能的LLTO薄膜,电子束功率是至关重要的因素。600W电子束功率下的LLTO沉积薄膜具有最大的离子导电率,为1.8×10^-7S/cm,活化能为0.32eV。从XRD和SEM,在不同电子束功率下的薄膜都表现出非晶态结构,但是薄膜的表面形貌在不同功率有不同的特点,而XPS测试显示更高的功率使LLTO薄膜中有更多的La,这有助于薄膜离子导电率的提高。而基于非晶态LLTO薄膜电解质的全固态电池也已被成功制备,并在3-4.4V区域内表现出较好的电化学性能。当然,为了进一步提高固体微电池的容量,我们要尽量提高不同功能薄膜间的接触面积,而二维固体电池薄膜间的平面接触面积必然有限,所以,开发具有高比表面积的三维形貌电极材料正在成为研究热点,碳微电机械体系（C-MEMS）正是其中的一种流行材料。作为本论文的拓展,我们对CMEMS的一种三维模型—微网碳膜（CMNFs）进行了初步的研究,以期望它在三维全固态微电池领域有潜在的应用。这种材料由微机械加工技术制备,包括SU-8光刻胶的图形模式制定和两步高温分解过程,通过TEM表征,CMNFs为短程有序金刚石相和非晶态碳相的混合体。CMNFs的首次放电容量可达350μAh/cm²,在后续的循环中,可逆容量很好地维持在100μAh/cm²左右,与层状的石墨碳相比,CMNFs的类似电容行为的充放电曲线和宽的CV峰暗示了更复杂的电极反应过程。在100次循环后,原来的三维形貌可以很好地维持,没有出现结构坍塌或碳网破裂的情况,根据CMNFs图形在电解液渗透效应的自调整现象,我们给出了一个Li离子在CMNFs中可能的迁移路径,即在没有温度梯度的高温退火下,大部分Li离子的脱嵌路径是平行于CMNF微网边的横截面的。由于网状结构的连续性和节点固定等特点,CMNFs被认为可以提供许多令人感兴趣的信息,而作为一种三维电极材料,其下一步目标是探寻在三维固态微电池中应用的可能性。更多还原

【Abstract】 All-solid-state thin film lithium batteries usually consist of positive electrode, solid electrolyte and negative electrode thin films.Compared with traditional powder materials,thin film materials could show their particular properties in some aspects. The fabrication of thin film form of the traditional materials can’t simply satisfy all the demands in the fields of thin film batteries.This thesis will focus on the positive electrode and electrolyte thin film materials applied in all-solid-state thin film lithium batteries.In our initial investigation on positive electrodes,we have considered the thin film fabrication of traditional materials with and without doping for the practical application in thin film batteries.At first,we successfully fabricated nano-crystalline Li_xMn₂O₄（LMO） thin films by R.F.Sputtering.LMO thin film exhibited distinct and symmetrical electrochemistry at 3V and 4V regions,respectively,and it was accordant with the performance of LiMn204 powder electrodes.Due to the absence of Mn²⁺ dissolution in all-solid-state case,the thin film battery with Li/LiPON/LMO layers showed excellent cycling stability in a large voltage range of 2-4.8V,and the reversible capacity could be well kept at 78μAh/cm²-μm.However,in the charge/discharge curves of LMO batteries,the feature of two separated plateaus with 1V step was considered to be unfavorable for practical application.Thus,we tried to add abundant ZrO2 in thin film for the preparation of LMZO to resolve this problem. ZrO₂ characterized by low thermal conductivity,high chemical inertness and high hardness was viewed as a suitable material to maintain the small grain size of LMO after high-temperature annealing or subsequent hundreds of cycles.The thin film battery with Li/LiPON/LMZO layers exhibited utterly different electrochemical behavior from LMO.1.7 Li per LMZO could be inserted at the end of discharge as LMO,and the capacity of Li/LiPON/LMZO could be reversibly maintained at 53μAh/cm²-μm.In fact,for most of positive electrode materials,the high-temperature synthesis process is unavoidable in order to obtain the ideal crystallinity,which is favored for the highly active electrochemistry.However,the high-temperature treatment is not expected in the fabrication process of thin film batteries considering the compatibility with micro-electron techniques.Moreover,the well-crystallized thin films always show quite rough surface with apparent grain boundaries,which is unfavorable for the interface match with solid electrolyte layers.Based on the requirement of on-chip integration of thin film batteries in micro-devices,the amorphous positive electrode thin films are expected from our knowledge due to the low-temperature deposition and the absence of annealing process.The smooth surface of amorphous thin film could largely improve the cycling life and electrode/electrolyte interface,whose degradation even short circuit could be effectively eliminated.According to the above description,LMZO thin film has been demonstrated to show smoother interface, which should be resulted from the control of crystal grain growth during anneal. However,the aim for the fabrication of amorphous thin film without high-temperature treatment hasn’t been achieved.Enlightened by the reports of amorphous FePO4 powder materials with active electrochemistry,we fabricated the amorphous FePO₄ and nitrided iron phosphate（FePON） positive electrode thin films for the first time.In the measurement of Li/LiPON/FePO4 cells,the FePO4 thin films deposited at low-temperature showed higher volumetric rate capacity.However,due to dynamic limitation,the reversible capacity was merely kept at about 20μAh/cm²-μm.The higher capacity 63μAh/cm²-μm of as-deposited FePON thin film was obtained when thin film was deposited in pure N₂ ambient.From CV curves of FePON,a set of additional redox peaks located at 1.7V and 2.4V regions was observed,and it was associated with the great capacity enhancement.The Raman and XPS data indicated that the PO₄^3- polyanion framework might be destroyed by the N insertion in FePON thin film.For further investigation on the improvement of thin film electrochemistry after N insertion,we have used the AC impedance technique for quantitive study from a viewpoint of dynamics.Here,two systems based on FePO4 and FePON thin films, sandwich structures with a variety of layer number and all-solid-state battery structures,were adopted.The ionic conductivity of FePON and FePO₄ thin films estimated from sandwich structures was 2.5×10^-8 S/cm and 7.5×10^-10 S/cm, respectively,while the electronic conductivity of FePON and FePO₄ was 2.6×10^-8 S/cm and 2.7×10^-6 S/cm,respectively.The ionic and electronic conductivities of FePON were much more balanced than those of FePO₄.N insertion resulted in the reverse change of thin film ionic and electronic conductivities,which might rise from the oxygen defects.FePON and FePO₄ had the same activation energy of 0.72 eV estimated from the ionic conductivity of various temperatures.The activation energy of interface between electrode and solid electrolyte was relatively small（0.25 eV）, indicating the compact contact of amorphous film layers and the facilitated ion-transfer through interface.In all-solid-state battery structures,the charge transfer resistance at various potentials was greatly reduced by the doping of N.The average chemical diffusion coefficient of Li⁺ ion FePON and FePO4 was 1.7×10¹¹ cm²/s and 1.2×10^-13 cm²/s,respectively,indicating that the doping of N facilitated the lithium diffusion in solid electrodes.On the other hand,as known,traditional positive electrodes usually consist of one reactive center.Of course,some materials have more than one active couples. However,the possibly heavy weight or structural collapse during charge/discharge may limit the mass rate capacity.In the fields of thin film electrodes and batteries,the volumetric rate capacity should be the most important performance index.Therefore, some materials with multiple active centers,despite of heavy weight,should be paid enough attention in their thin film form.These compact thin films with high density are expected to improve the volumetric rate capacity of thin film batteries.In this section,we have choosed the materials based on active polyanion framework（WO₄）^2-. CuWO₄ and LiFe（WO₄）₂ positive electrode thin films with two active couples have been successfully fabricated by R.F.Sputtering.For CuWO4 thin films,through the characterization of XRD and SEM,the anorthic phase was made up of nano-sized crystal grains with 20 nm dimension.The initial discharge capacity of 192 mAh/g was achievable for Li/CuWO₄ cells,and two separated plateaus at 2.5 V and 1.6 V regions were detected due to the reactivity of both Cu²⁺/Cu⁰ and W⁶⁺/W⁴⁺.Because of the possible degradation at electrode-electrolyte interface and Cu ion dissolution in liquid electrolyte,CuWO4 electrode underwent gradually structural evolution accompanied with capacity degradation.Based on the TEM and SAED data at different charged/discharged states,a reasonable two-step reaction mechanism of CuWO₄ associated with displacement reaction was suggested.The（WO₄）^2- framework could be rebuilt after reductive decomposition,and well-crystallized CuWO₄ could be reproduced as well during electrochemistry.In Li/LiPON/CuWO₄ cells,the initial discharge capacity was as high as 145μAh/cm²-μm.With the absence of possible degradation phenomena at cathode-electrolyte surface and Cu ion dissolution in all-solid-state system,the electrode structure and electrochemical curves were well kept during the following cycling.Amorphous LiFe（WO₄）₂ positive electrode thin films are expected to achieve both the objects,the design of multiple reactive centers and the deposition fabrication at low-temperature.Li/LiFe（WO₄）₂ cells showed a large initial discharge capacity of 198 mAh/g,and two separated plateaus at 3 V and 1.5 V regions were observed due to the reactivity of two redox couples(Fe³⁺/Fe²⁺ and W⁶⁺/W^x+,where x = 4 or 5).A sustainable discharge capacity of 110 mAh/g could be maintained over 300 cycles. Based on the XPS and TEM data at different charged/discharged states,it has been demonstrated that two redox centers associated with Fe and W were electrochemically reversible.An unavoidable crystallization tendency in as-deposited LiFe（WO4）2 thin film was found after initially several charge/discharge cycles,with concomitant continuous capacity degradation in the potential range between 1 V and 2 V Another advantage of LiFe（WO₄）₂ thin film is its capacity retention under high current density, and a volumetric rate capacity of 60μAh/cm²-μm could be well stabilized under a current density of 50μA/cm².Solid electrolyte thin films are the crucial functional layers in the assembly of thin film batteries.However,the valuable thin film electrolytes have not been plentifully provided as expected.So far,the lithium phosphorous oxynitride（LiPON） thin film has been widely used due to its high ionic conductivity and large electrochemical stability window.In order to enhance the ionic conductivity of LiPON,it is necessary to increase the N insertion quantity in thin films and improve the thin film deposition techniques.Here,we used electron cyclotron resonance（ECR） plasma in R.F.Sputtering equipment to assist the LiPON deposition for increasing the ionic conductivity.The optimization condition was found to be at ECR power of 200 W.Under this fabrication condition,the molar ratio of N/P was estimated to be 0.65,and Li ionic conductivity was obtained to be 8.0×10^-6S/cm.In addition,the R.F.Sputtering technique,which is the best method for LiPON fabrication,has been considered to have some drawbacks,such as low deposition rate and small deposition area.Moreover,some intrinsic disadvantages of LiPON thin film, such as vulnerability to moisture in air,couldn’t also be neglected.So we used e-beam evaporation technique to fabricate lithium lanthanum titanate（LLTO） solid electrolyte thin-films instead of LiPON thin film.This fabrication method could greatly enhance the deposition efficiency.E-beam power was a crucial factor for the preparation of high quality LLTO thin films.The as-deposited LLTO thin film at e-beam power of 600 W had the highest ionic conductivity of 1.8×10^-7 S/cm with activation energy of 0.32 eV.All thin films prepared at different e-beam powers showed the amorphous structure but with distinct surface morphologies from XRD and SEM data.XPS measurement indicated that higher e-beam power could keep more La insertion into LLTO thin film.An all-solid-state cell was fabricated using the amorphous LLTO thin film as solid electrolyte and presented good cycle stability at electrochemical window of 3 V-4.4 V.In order to further enhance the capacity of solid micro-batteries,the enlargement of contact area between different functional layers is expected.However,the planar contact area of 2D solid-battery thin films is very limited.The development of 3D electrodes with high rate surface area is being paid much attention today.Recently,the carbon micro-electromechanical systems（C-MEMS） as 3D materials are fashionable. As the extension part of this thesis,the 3D carbon micro-net films（CMNFs） as a prototype of C-MEMS have been primarily investigated for the possible application in 3D solid micro-batteries.CMNFs were fabricated by micro-machining technology, consisting of SU-8 photoresist patterning by photolithography and two-step pyrolysis process.They structurally presented the mixture of both short-distance ordered diamond-like phase and amorphous carbon matrix by TEM.The first discharge capacity for CMNFs was calculated to be as high as 350μAh/cm²,and in the following cycles,the reversible capacity was well maintained at about 100μAh/cm². Compared with graphitized carbonous materials,the pseudocapacitance-like electrochemical behavior and wide CV redox peaks indicated complex electrode reaction with Li⁺.After 100^th cycle,the original 3D shape of CMNFs was well kept without the presence of structural collapse and network rupture.A reasonable Li⁺ moving pathway was concluded according to the size self-regulation of CMNF pattern due to the effect of electrolyte penetration.Under the absence of pyrolysis temperature gradients,most of Li could be intercalated or deintercalated paralleling to the cross-section of CMNF sides.Due to the features of continuity of network structure and fixing of nodes,the CMNFs are expected to provide much interesting information.For the next step,they are aimed for the possible application in future 3D solid micro-batteries.更多还原