Wafer bonding of silicon wafers covered with various surface layers

2000 ◽  
Vol 86 (1-2) ◽  
pp. 91-95 ◽  
Author(s):  
M Wiegand ◽  
M Reiche ◽  
U Gösele ◽  
K Gutjahr ◽  
D Stolze ◽  
...  
2012 ◽  
Vol 2012 (CICMT) ◽  
pp. 000436-000440 ◽  
Author(s):  
S. Günschmann ◽  
M. Fischer ◽  
T. Bley ◽  
I. Käpplinger ◽  
W. Brode ◽  
...  

For the fabrication of a micro fluidic high pressure oil sensor (400 bar) based on an infrared transmission measuring principle the bonding of 2 mm silicon wafers is necessary. Conventional bonding techniques such as silicon fusion bonding or anodic bonding are not suitable for bonding thick and inflexible silicon wafers, because these techniques can not compensate for the wafer bow. We present a new bonding procedure for silicon substrates thicker than 1 mm using a silicon adapted LTCC tape as an intermediate leveling layer. The wafers are preprocessed by etching a nano structured silicon surface on the internal side. The silicon wafers are aligned and stacked with pre-structured green LTCC tapes by an optical stacking unit. During the hot isostatic lamination at 55 bar the structured LTCC tape is adjusted to the silicon. A subsequent pressure assisted sintering leads to a wafer bonding strength up to 5000 N/cm2. With the bonding technique it is possible to create cavities and channels between the thick wafers by the use of punched and laser cut LTCC. The fabrication steps of the sandwich build-up especially the sequential lamination and the optical adjusting procedure of the flexible (LTCC) and inflexible (2 mm Wafer) substrates will be explained in detail. A method to reduce the shrinkage and distortion of the green LTCC during handling is demonstrated. The distribution of the bonding and bursting strength of the single fluidic systems on a complete sandwich substrate is analyzed.


Author(s):  
Devasena Duraipandi ◽  
John M. Heck ◽  
Raymond K. Yee ◽  
Sang-Joon J. Lee

Wafer-level packaging of RF MEMS devices offers an attractive option to reduce packaging cost significantly and ensures hermetic encapsulation of devices. Low-temperature cofired ceramic (LTCC) cap wafers are particularly favorable because they can be pre-patterned with through-wafer vias for integrated electrical contacts and high-density packaging, at a much lower cost than silicon wafers with similar features. However, thermal expansion mismatch between ceramic and silicon wafers at high bonding temperatures induces thermal stresses at the interface, resulting in wafer curvature. For example, a 150 mm silicon wafer 675 μm thick with a ceramic cap wafer 500 μm thick has been measured to exhibit out-of-flatness displacement as severe as of 1.7 mm at the center. While the curvature can be reduced significantly using low-thermal-expansion ceramic, such materials are non-standard and require custom formulation. Furthermore, as the wafer diameter is increased, thermal expansion mismatch becomes more problematic. Therefore, it is desirable to address the problem using a geometrical approach in addition to optimizing the ceramic for wafer bonding applications. The present study applies finite element analysis (FEA) to examine the potential for reducing such curvature by introducing slots in the ceramic cap wafer. Two-level factorial design simulations involving five parameters were conducted to investigate the effect of slot parameters on wafer curvature, using 2-D plane strain simulation of wafer cooling from 300 °C to 25 °C. The five parameters investigated were cap wafer thickness, slot width, slot depth, slot separation, and slot orientation. The nonlinear temperature dependence of thermal expansion was also examined based on test data for the ceramic wafers. Furthermore, a 3-D finite element simulation was conducted to compare the 2-D results to overall impact on wafer distortion. FEA results were compared with experimental curvature measurements on sample wafers measured by coordinate measuring machining (CMM). Simulated results suggest that introduction of slots shows reduction in wafer curvature, and the displacement can be reduced by as much as 25% based on the geometric parameter values for slots in the cap wafer.


1995 ◽  
Vol 378 ◽  
Author(s):  
A. Laporte ◽  
G. Sarrabayrouse ◽  
M. Benamara ◽  
A. Claverie ◽  
A. Rocher ◽  
...  

AbstractThis paper presents the comparison of the structural and electrical characteristics of Si/Si bonded interfaces depending on whether the surface layers were rendered amorphous by high dose ion implantation prior to annealing or not. While the general structure of the interfaces is the same when the wafers are preamorphized more precipitates are seen in the interface along with a few extended defects propagating into the volume. The most striking difference between both procedures is that the Spreading Resistance profile is more complicated in shape and difficult to master in the case of preamorphized wafers. Careful TEM analysis shows that only in this case the interfacial region is stressed in contrast with the fully relaxed structure obtained by direct bonding of crystalline wafers.For these reasons, there is little chance that the preamorphization technique will benefit to the bonding procedure of direct Si wafers.


1987 ◽  
Vol 107 ◽  
Author(s):  
Kun-Young Ahn ◽  
Ulrich Gösele ◽  
Patrick Smith

AbstractThe conditions for the dissolution and disintegration of SiO2 layers between silicon wafers during direct wafer bonding are discussed in terms of two possible mechanisms. The calculated maximal thickness of a SiO2 layer which may be completely dissolved does not only depend on the bonding temperature and time but also on the starting concentration of interstitial oxygen in the silicon wafers. Finally, the influence of rotational misorientation of the two wafers on the behavior of the S1O2 layers is dealt with.


1998 ◽  
Vol 535 ◽  
Author(s):  
Stefan Bengtsson ◽  
Mats Bergh ◽  
Anders Söderbärg ◽  
Bengt Edholm ◽  
Jörgen Olsson ◽  
...  

AbstractMaterial integration for the formation of advanced silicon-on-insulator materials by wafer bonding and etch-back will be discussed. Wafer bonding allows combining materials that may not be possible to grow on top of each other by any other technique. In our experiments, polycrystalline diamond, aluminum nitride or aluminum oxide films with thickness of 0.1-5 µm were deposited on silicon wafers. Bonding experiments were made with these films to bare silicon wafers with the goal of forming silicon-on-insulator structures with buried films of polycrystalline diamond, aluminum nitride or aluminum oxide. These silicon-on-insulator structures were aimed to address self-heating effects in conventional silicon-on-insulator materials with buried layers of silicon dioxide. The surfaces of the deposited diamond films were, by order of magnitude, too rough to allow direct bonding to a silicon wafer. In contrast the deposited aluminum nitride and aluminum oxide films did allow direct bonding to silicon. Bonding of the diamond surface to silicon was instead made through a deposited and polished layer of polycrystalline silicon on top of the diamond. In the case of the aluminum nitride electrostatic bonding was also demonstrated. Further, the compatibility of these insulators to silicon process technology was investigated.


2017 ◽  
Vol 261 ◽  
pp. 93-100
Author(s):  
Ronald Allan S. de los Reyes

The current concept of grinding or abrasive machining involves the formation and removal of segmented strips of material termed chips from the surface of the solid. A novel cutting mechanism is hereby presented in this research study that suggests that the generation of chips from the surface does not occur but only a shearing process that splits material creating added surface features and textures in the silicon surface. This arises from the unique set of factors of abrasive grit size, thrust force, polishing speed, and polishing time that lead to phase transformations in the surface layers of the silicon wafers. Statistical analysis of the factor effects yielded results that show the surface roughness values, Ra and Rz, increasing without any appreciable change in the thickness of the silicon wafers. This can be attributed to the proposed cutting mechanism indicating that only in-plane surface shearing occurred due to the change of the silicon crystal structure from exhibiting brittle behavior to that of ductile mode of deformation. Moreover, experimental quantities of the specific energy for surface machining of silicon was calculated with an overall mean of 50.5 GPa. This is about 33% less than the currently accepted value and can be considered further evidence that polymorphic transitions to a softer material occurred rendering the surface layers more susceptible to longitudinal cutting deformation and fracture. A model based on the inverted spherical cap or spherical bottom geometry for the individual abrasive particle is also proposed, verified by a finite element method analysis simulation, that can mathematically describe this particular micromachining process.


1984 ◽  
Vol 75 ◽  
pp. 407-422
Author(s):  
William K. Hartmann

ABSTRACTThe nature of collisions within ring systems is reviewed with emphasis on Saturn's rings. The particles may have coherent icy cores and less coherent granular or frosty surface layers, consistent with thermal eclipse observations. Present-day collisions of such ring particles do not cause catastrophic fragmentation of the particles, although some minor surface erosion and reaccretion is possible. Evolution by collisional fragmentation is thus not as important as in the asteroid belt.


Author(s):  
P.E. Batson ◽  
C.R.M. Grovenor ◽  
D.A. Smith ◽  
C. Wong

In this work As doped polysilicon was deposited onto (100) silicon wafers by APCVD at 660°C from a silane-arsine mixture, followed by a ten minute anneal at 1000°C, and in one case a further ten minute anneal at 700°C. Specimens for TEM and STEM analysis were prepared by chemical polishing. The microstructure, which is unchanged by the final 700°C anneal,is shown in Figure 1. It consists of numerous randomly oriented grains many of which contain twins.X-ray analysis was carried out in a VG HB5 STEM. As K α x-ray counts were collected from STEM scans across grain and twin boundaries, Figures 2-4. The incident beam size was about 1.5nm in diameter, and each of the 20 channels in the plots was sampled from a 1.6nm length of the approximately 30nm line scan across the boundary. The bright field image profile along the scanned line was monitored during the analysis to allow correlation between the image and the x-ray signal.


Sign in / Sign up

Export Citation Format

Share Document