Forces acting between polishing tool and workpiece surface in magnetorheological finishing

Polishing characteristics and mechanism in magnetorheological planarization using a permanent magnetic yoke with translational movement

“…The normal force of 3-10 N in this process is comparable to those obtained from the conventional MRF process, for instance, 6-9 N reported by Miao et al [24] and 2-20 N reported by Schihaerl et al [25]. However, the polishing contact area of 2500 mm 2 using our developed PM yoke with a straight air gap was much larger than those using the conventional electromagnets, such as the areas of 36-54 mm 2 reported in [24] and 28 mm 2 in [25].…”

“…The normal force of 3-10 N in this process is comparable to those obtained from the conventional MRF process, for instance, 6-9 N reported by Miao et al [24] and 2-20 N reported by Schihaerl et al [25]. However, the polishing contact area of 2500 mm 2 using our developed PM yoke with a straight air gap was much larger than those using the conventional electromagnets, such as the areas of 36-54 mm 2 reported in [24] and 28 mm 2 in [25]. Therefore, the shear and normal stresses of ∼3.6-15 kPa and ∼1.2-4 kPa in our process are considerably smaller than the shear stress of ∼80-110 kPa in [24] and normal stresses of ∼160-190 kPa in [24] and ∼71-714 kPa in [25].…”

“…However, the polishing contact area of 2500 mm 2 using our developed PM yoke with a straight air gap was much larger than those using the conventional electromagnets, such as the areas of 36-54 mm 2 reported in [24] and 28 mm 2 in [25]. Therefore, the shear and normal stresses of ∼3.6-15 kPa and ∼1.2-4 kPa in our process are considerably smaller than the shear stress of ∼80-110 kPa in [24] and normal stresses of ∼160-190 kPa in [24] and ∼71-714 kPa in [25]. The significant difference in stress between the conventional processes and our process should be mainly attributed to the different field strengths and contact areas used.…”

“…This was attributed to the use of different magnetic sources in the two processes. In the conventional process [24,25] an electromagnet was used to produce a strong magnetic field, under which MR fluid formed a hard "core" near the periphery of the carrier wheel. The existence of the "core" resulted in a throttling effect and drastically increased the hydrodynamic pressure when the MR fluid passed through the work gap.…”

Experimental Study on Cluster Magnetorheological Variable Gap Dynamic Pressure Planarization Finishing

“…proved that the cluster magnetorheological variable gap dynamic pressure planarization finishing is feasible and effective. Key words：cluster magnetorheological planarization finishing；variable gap dynamic pressure；planarization finishing；surface roughness；material removal rate 0 前言 随着 5G 通讯、人工智能等高新技术的发展， 对光电晶片超光滑平坦化加工提出了更高的要求。 目前，光电晶片平坦化加工的主流技术仍然是以游 离磨料机械去除和化学反应复合的化学机械 抛 光 [1] ，其中晶片与抛光工具为刚性接触状态，磨料 粒径不均匀容易产生晶片表面、亚表面损伤和残余 应力等加工缺陷，同时导致抛光垫磨损、失效，严 重影响晶片的加工品质和加工成本。 磁流变抛光作为确定性抛光方法，磨料通过柔 性去除实现加工，具有去除函数稳定、加工过程可 控、加工表面不产生亚表面损伤等优点而被广泛 应用于光学加工 [2][3] 。FENG 等 [4] 利用磁流变抛光 抛光大型蓝宝石光 学窗口，获得 PV 约为 λ/15、表面粗糙度 RMS 为 0.8 Å 的超光滑表面。然而，经典磁流变抛光过程中 工件与"抛光缎带"为"斑点"局部接触，需要用 轨迹扫描的方式才可以加工整体表面，加工效率 低 [4] 。为了提高加工效率和质量，磁流变抛光垫与 工件"面-面"接触的集群磁流变平面抛光等方法被 提出 [6] ，利用该方法加工 2 英寸(1 英寸=2.54 cm)碳 化硅晶片 30 min， 表面粗糙度从 Ra72.89 nm 迅速降 低至 1.9 nm [6] ；在此基础上将静态磁场变换为多磁 极旋转动磁场，对单晶碳化硅、单晶硅片、钛酸锶、 磷化铟等基片进行抛光，获得了纳米级甚至埃米级 的表面粗糙度 [7][8][9] ，但抛光压力动态稳定性是制约集 群磁流变抛光效率的重要因素 [10] 。为研究磁流变抛 光过程中的抛光压力，SCHINHAERL 等 [11] 采用三 向测力仪测量在磁流变抛光过程中磁流变液形成的 缎带经过楔形间隙时对工件表面的作用力，发现当 使用 50 mm 的抛光轮时所测的法向力在 2～20 N 之 间，工件越靠近抛光轮，受到的作用力越大。MIAO 等 [12] 发现法向力与切向力随着金刚石磨料质量分数 的增加而增大，但随着金刚石磨粒浓度的进一步增 大最终保持在一个稳定值，随着压入抛光缎带深度 的增加，法向力与剪切力明显增大。梁华卓等 [13] 研 究了加工参数对抛光力的影响，发现抛光力 F n 和 F t 随转速的增大而增大， 随加工间隙的增大而减少， 随磨粒浓度和羰基铁粉浓度的增大而增大，随偏摆 幅度的增大抛光力略有增大。付有志等 [14] 人研究了 抛光盘结构对抛光力的影响，发现抛光压力随楔形 角和工作间隙的增大而减少，随楔形区宽度的增大 而增大。TAO 等 [15][16] 发现磁流变液在挤压作用下可 以提供较大作用力，其他学者也发现了该规律并提 出了磁流变液的挤压强化效应。但这方面的研究目 前主要侧重于理论分析…”