碱性阻挡层抛光液各成分对CMP的影响

为实现图形片的全局平坦化,通过研究碱性阻挡层抛光液各成分对铜和介质(TEOS)去除速率的影响,遴选出一种碱性阻挡层抛光液。在此抛光液基础上添加不同质量分数的盐酸胍,对钽光片进行抛光,选出满足要求的抛光液,并在中芯国际图形片上验证此抛光液的修正能力。实验表明,当磨料质量分数为20%、盐酸胍质量分数为0.3%、I型螯合剂(FA/O I)体积分数为1%、非表面活性剂体积分数为3%时,钽和TEOS去除速率之和是铜去除速率的3.3倍,此种碱性阻挡层抛光液对钽的去除速率为42 nm/min,各项参数均满足工业要求。与商用酸性、碱性抛光液相比,该抛光液对碟形坑和蚀坑有更好的修正能力。     

新型碱性阻挡层抛光液的研究(英文)

为了达到Ta的高去除速率,通过实验改变碱性阻挡层抛光液中各组分的体积分数,研究各组分变化对阻挡层材料Cu和Ta抛光速率的影响。碱性阻挡层抛光液中,首先确定磨料及去离子水的比例,然后再改变盐酸胍和螯合剂等组分体积分数,得到Cu和Ta抛光速率的变化规律,便于选取此条件下阻挡层抛光液最佳配比。通过各种测试手段表征抛光液对12英寸(1英寸=2.54 cm)布线片碟形坑修正能力及铜布线表面粗糙度的影响。结果表明,Ta和Cu材料的抛光速率随各组分变化规律明显,当磨料体积分数为30%、盐酸胍体积分数为3%以及螯合剂体积分数为1%时,可以达到Ta和Cu的最大抛光速率差。优化后的新型阻挡层抛光液1 min内可以实现碟形坑的有效修正及粗糙度的明显降低。     

FA/O阻挡层抛光液在铜布线中的去除速率的选择性优化

在阻挡层平坦化的过程中,由于不同材料去除速率的要求,所以有必要研发一种阻挡层抛光液来改善Cu/barrier/TEOS(SiO2)速率的选择性来减小蚀坑和碟形坑,。本次实验中,研发了一种不含腐蚀抑制剂BTA的FA/O碱性阻挡层抛光液。它包含了20mL/L的FA/O螯合剂,5mL/L的表面活性剂,磨料浓度为1:5,pH值为8.0。通过控制抛光液的成分,有效的控制了不同材料的去除速率。最后,考察FA/O阻挡层抛光液对介电层材料的影响。在进行完CMP过程后,检测介电层材料的性能。结果显示,漏电流在皮安级,电阻2.4 kΩ,电容2.3pF,碟形坑和蚀坑都在30nm以下。介质层的各项指标完全满足工业化生产的需要。     

28nm集成电路铜互连弱碱性阻挡层抛光液性能北大核心

报道了一种用于技术节点为28 nm的集成电路多层铜互连阻挡层化学机械平坦化(CMP)的弱碱性抛光液(pH值为8.5),仅由平均粒径为20 nm的硅溶胶、FA/O多羟多胺螯合剂及非离子表面活性剂组成。实验结果表明,研发的弱碱性阻挡层抛光液对牺牲层SiO2、阻挡层Ta及金属互连线Cu等多种材料具有优异的速率选择性,去除速率一致性大于95%,并在28 nm多层铜布线片阻挡层CMP后具有较高的平坦化性能,分别获得了20 nm以下深度的碟形坑和蚀坑。通过电化学极化曲线测试表明,此弱碱性阻挡层抛光液能有效减少Cu和Ta之间的腐蚀电位差,并避免了铜互连结构Cu和Ta界面处电偶腐蚀的产生,这对提高集成电路芯片可靠性具有重要意义。     

无氧化剂条件下铜钴CMP去除速率的控制北大核心

利用不含氧化剂的碱性抛光液对铜和钴进行化学机械抛光,深入分析了抛光液组分包括硅溶胶磨料、FA/O螯合剂以及非离子表面活性剂对两种金属去除速率的影响规律及作用机理。实验结果表明,铜和钴的去除速率随着磨料质量分数的增加而升高,并且在磨料质量分数低于5%时钴的去除速率为20~30 nm/min,而铜的去除速率几乎为零;加入FA/O螯合剂可增强其与金属离子的络合,从而加快铜和钴的去除速率;非离子表面活性剂可以有效降低铜和钴的表面粗糙度。在抛光液各组分的协同作用下,可以达到两种材料的低表面粗糙度和高去除速率选择性。     

FA/O精抛液组分对Cu/Ta CMP去除速率的影响

研究了精抛液中各组分对Cu/Ta去除速率的影响.在分别考察磨料质量分数、Ⅱ型螯合剂、活性剂和双氧水等各组分对Cu/Ta去除速率的影响后,研发了磨料质量分数为2％,Ⅱ型螯合剂体积分数为0.35％,活性剂体积分数为2％,H2O2体积分数为2％的FA/O精抛液.并在MIT854布线片上进行了精抛测试.结果显示,精抛12 s之后,100-100线条处高低差从精抛前的76.7 nm降低为63.2 nm,而50-50线条处高低差从精抛前的61.3 nm降低为59.8 nm.并且,经过FA/O精抛液精抛后粗抛后产生的尖峰也有所减小,台阶趋于平坦.精抛后,100-100和50-50处的高低差均小于70 nm,能够达到产业化指标.     

不同表面活性剂对铜互连阻挡层CMP一致性的影响北大核心CSCD

为实现铜互连阻挡层化学机械抛光的高平坦化,研究了阻挡层抛光液中表面活性剂的作用。在抛光液中分别添加非离子型表面活性剂异辛醇聚氧乙烯醚(JFC-E)、本课题组研发的Ⅱ型活性剂以及复合表面活性剂,研究其对阻挡层CMP中铜和正硅酸乙酯(TEOS)去除速率及一致性的影响。实验结果表明,表面活性剂不仅可以改善Cu和TEOS的去除速率一致性和晶圆全局一致性,还可以降低铜表面的粗糙度,通过复配两种不同的表面活性剂,可以提高Cu和TEOS的去除速率,使Cu和TEOS的剩余厚度片内非均匀性(WIWNU)分别降低至1.39%和1.38%,铜表面粗糙度从5.25 nm降至1.51 nm。基于实验结果,研究并提出了表面活性剂影响抛光液润湿性,从而改善片内非均匀性和铜表面质量的机理。     

新型碱性抛光液各组分对铜化学机械平坦化性能的影响

随着集成电路特征尺寸的减小、晶圆尺寸的增大以及布线层的逐渐增多,加工晶圆过程中实现较高的材料去除速率、较小的片内非一致性(WIWNU)及较小的表面粗糙度已经成为铜化学机械抛光工艺的几大难点.采用正交实验法选取5组抛光液进行Cu CMP实验,系统研究了含有双氧水、脂肪醇聚氧乙烯醚(AEO)、FA/O Ⅰ型螯合剂与苯骈三氮唑(BTA)的碱性抛光液化学组分对铜去除速率、WIWNU的影响,并对铜CMP的各种变化规律做出机理分析.结果表明:采用pH值约为8.6,体积分数为3％的H202,质量分数为0.08％的非离子表面活性剂AEO与体积分数为1.5％的螯合剂的碱性抛光液,在12英寸(1英寸=2.54 cm)铜镀膜片抛光后有助于去除速率达到629.1 nm/min,片内非一致性达到4.7％,粗糙度达到1.88 nm.     

阻挡层CMP中表面活性剂对抛光效果的影响

研究了阻挡层在化学机械抛光过程中表面活性剂的作用。利用5种活性剂体积分数不同的抛光液对3英寸(1英寸=2.54 cm)Cu/Ta/SiO2光片进行抛光。通过抛光前后质量变化可得各晶圆片的抛光速率,再对12英寸布线晶圆片进行抛光,利用台阶仪测量抛光前后碟形坑大小的变化,最后利用原子力显微镜对抛光后布线晶圆片的表面形貌进行测试。研究表明抛光液中活性剂体积分数不同会引起抛光速率的变化,也会影响碟形坑的修正效果。当抛光液中活性剂体积分数达到2.0%时,修正值达到85.6 nm/min,优于其他活性剂体积分数时的修正值。另外,活性剂体积分数的增加有助于降低抛光后晶圆表面的粗糙度。活性剂体积分数小于3.0%时,粗糙度随着活性剂体积分数的增加而降低。这一发现可以对配制抛光液时活性剂体积分数的确定起到一定的参考作用。     

各组分体积分数对碱性抛光液存储时间的影响

主要研究了碱性抛光液各组分体积分数对其有效存储时间的影响。实验中每隔两个月测试了抛光液的pH值、平均粒径和Cu膜去除速率等参数随存储时间的变化值。研究表明:FA/O螯合剂体积分数是影响抛光液有效存储时间的主要因素,螯合剂体积分数越高抛光液的有效存储时间越短。在FA/O螯合剂体积分数较低时(8%),Cu布线碱性抛光液的有效存储时间在半年以上,基本能够符合产业化要求。SiO2磨料体积分数和非离子型表面活性剂体积分数是影响碱性抛光液有效存储时间的次要因素,对其有效存储时间影响不明显。     

Performance improvement in tensile-strained In/sub 0.5/Al/sub 0.5/As/In/sub x/Ga/sub 1-x/As/In/sub 0.5/Al/sub 0.5/As metamorphic HEMT

This paper proposes a In/sub 0.5/Al/sub 0.5/As/In/sub x/Ga/sub 1-x/As/In/sub 0.5/Al/sub 0.5/As (x=0.3-0.5-0.3) metamorphic high-electron mobility transistor with tensile-strained channel. The tensile-strained channel structure exhibits significant improvements in dc and RF characteristics, including extrinsic transconductance, current driving capability, thermal stability, unity-gain cutoff frequency, maximum oscillation frequency, output power, power gain, and power added efficiency.     

Very low-noise Al/sub 0.3/Ga/sub 0.7/As/Ga/sub 0.65/In/sub 0.35/As/GaAs single quantum-well pseudomorphic HEMTs

AlGaAs/GaInAs/GaAs pseudomorphic HEMTs with an InAs mole fraction as high as 35% in the channel has been successfully fabricated. The device exhibits a maximum extrinsic transconductance of 700 mS/mm. At 18 GHz, a minimum noise figure of 0.55 dB with 15.0 dB associated gain was measured. At 60 GHz, a minimum noise figure as low as 1.6 dB with 7.6 dB associated gain was also obtained. This is the best noise performance yet reported for GaAs-based HEMTs.<>     

A 12 /spl times/ 12 In/sub 0.53/Ga/sub 0.47/As-In/sub 0.52/Al/sub 0.48/As avalanche photodiode array

We report a 12 /spl times/ 12 In/sub 0.53/Ga/sub 0.47/As-In/sub 0.52/Al/sub 0.48/As avalanche photodiode (APD) array. The mean breakdown voltage of the APD was 57.9 V and the standard deviation was less than 0.1 V. The mean dark current was /spl sim/2 and /spl sim/300 nA, and the standard deviation was /spl sim/0.19 and /spl sim/60 nA at unity gain (V/sub bias/ = 13.5 V) and at 90% of the breakdown voltage, respectively. External quantum efficiency was above 40% in the wavelength range from 1.0 to 1.6 /spl mu/m. It was /spl sim/57% and /spl sim/45% at 1.3 and 1.55 /spl mu/m, respectively. A bandwidth of 13 GHz was achieved at low gain.     

Al/sub 0.3/Ga/sub 0.7/As/In/sub x/Ga/sub 1-x/As (0/spl les/x/spl les/0.25) doped-channel field-effect transistors (DCFET's)

The properties of both lattice-matched and strained doped-channel field-effect transistors (DCFET's) have been investigated in AlGaAs/In/sub x/Ga/sub 1-x/As (0/spl les/x/spl les/0.25) heterostructures with various indium mole fractions. Through electrical characterization of grown layers in conjunction with the dc and microwave device characteristics, we observed that the introduction of a 150-/spl Aring/ thick strained In/sub 0.15/Ga/sub 0.85/As channel can enhance device performance, compared to the lattice-matched one. However, a degradation of device performance was observed for larger indium mole fractions, up to x=0.25, which is associated with strain relaxation in this highly strained channel. DCFET's also preserved a more reliable performance after biased-stress testings.<>     

Morphological and electrical characterization of SixCryCzBv thin films

Si_xCr_yC_zB_v thin films with several compositions have been studied for integration of high precision resistors in 0.8 μm BICMOS technology. These resistors, integrated in the back-end of line, have the advantage to provide high level of integration and attractive electrical behavior in temperature, for analog devices. The film morphology and the structure have been investigated through transmission electron microscopy analysis and have been then related to the electrical properties on the base of the percolation theory. According to this theory, and in agreement with experimental results, negative thermal coefficient of resistance (TCR) has been obtained for samples with low Cr content, corresponding to a crystalline volume fraction below the percolation threshold.Samples with higher Cr content exhibit, instead, a variation of the TCR as a function of film thickness: negative TCR values are obtained for thickness lower than 5 nm, corresponding to a crystalline volume fraction below the percolation threshold; positive TCR are obtained for larger thickness, indicating the establishment of a continuous conductive path between the Cr rich grains. This property seems to be determinant in order to assure the possibility to obtain thin film resistors almost independent on the temperature.     

Highly linear Al/sub 0.3/Ga/sub 0.7/N-Al/sub 0.05/Ga/sub 0.95/N-GaN composite-channel HEMTs

We report an Al/sub 0.3/Ga/sub 0.7/N-Al/sub 0.05/Ga/sub 0.95/N-GaN composite-channel HEMT with enhanced linearity. By engineering the channel region, i.e., inserting a 6-nm-thick AlGaN layer with 5% Al composition in the channel region, a composite-channel HEMT was demonstrated. Transconductance and cutoff frequencies of a 1 /spl times/100 /spl mu/m HEMT are kept near their peak values throughout the low- and high-current operating levels, a desirable feature for linear power amplifiers. The composite-channel HEMT exhibits a peak transconductance of 150 mS/mm, a peak current gain cutoff frequency (f/sub T/) of 12 GHz and a peak power gain cutoff frequency (f/sub max/) of 30 GHz. For devices grown on sapphire substrate, maximum power density of 3.38 W/mm, power-added efficiency of 45% are obtained at 2 GHz. The output third-order intercept point (OIP3) is 33.2 dBm from two-tone measurement at 2 GHz.     

Comparison between Pt/TiO2/Pt and Pt/TaOX/TaOY/Pt based bipolar resistive switching devices

Nonvolatile memories have emerged in recent years and have become a leading candidate towards replacing dynamic and static random-access memory devices. In this article, the performances of TiO₂ and TaO₂ nonvolatile memristive devices were compared and the factors that make TaO₂ memristive devices better than TiO₂ memristive devices were studied. TaO₂ memristive devices have shown better endurance performances (10⁸ times more switching cycles) and faster switching speed (5 times) than TiO₂ memristive devices. Electroforming of TaO₂ memristive devices requires~4.5 times less energy than TiO₂ memristive devices of a similar size. The retention period of TaO₂ memristive devices is expected to exceed 10 years with sufficient experimental evidence. In addition to comparing device performances, this article also explains the differences in physical device structure, switching mechanism, and resistance switching performances of TiO₂ and TaO₂ memristive devices. This article summarizes the reasons that give TaO₂ memristive devices the advantage over TiO₂ memristive devices, in terms of electroformation, switching speed, and endurance.     

Broadband microwave noise characteristics of high-linearity composite-channel Al/sub 0.3/Ga/sub 0.7/N/Al/sub 0.05/Ga/sub 0.95/N/GaN HEMTs

We report broadband microwave noise characteristics of a high-linearity composite-channel HEMT (CC-HEMT). Owing to the novel composite-channel design, the CC-HEMT exhibits high gain and high linearity such as an output third-order intercept point (OIP3) of 33.2 dBm at 2 GHz. The CC-HEMT also exhibits excellent microwave noise performance. For 1-/spl mu/m gate-length devices, a minimum noise figure (NF/sub min/) of 0.7 dB and an associated gain (G/sub a/) of 19 dB were observed at 1 GHz, and an (NF/sub mi/) of 3.3 dB and a G/sub a/ of 10.8 dB were observed at 10 GHz. The dependence of the noise characteristics on the physical design parameters, such as the gate-source and gate-drain spacing, is also presented.