Effects of the Underlayer Surface State on the Interconnecting Aluminum Film Properties
- PDF / 1,010,529 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 76 Downloads / 165 Views
Before we perform 5000A Al-l.0%Si-0.5%Cu / 500A Ti metal depositions at 500 'C, 1000A oxides were thermally grown at 900 'C on silicon substrates. To investigate surface effect of the chemical vapor deposition (CVD) oxides on the subsequently depositing metal quality, a variety of CVD oxides such as ozone undoped silicate glass (0 3 -USG) and silicon-rich silicon dioxide (Si+,xOz.,) were deposited. For these oxides, we deposited 8000A Al-0.5%Cu / 1000A Ti metal depositions at 400 1C. We examined surface roughness of the each thin film layers by using non-contact mode atomic force microscopy (AFM). As shown in figure 1, reflectivity of the aluminum films at iline (365nm) showed a good linear relation with AFM root mean square (rms), and was used for indirect measuring aluminum surface roughness. Auger electron spectroscopy (AES) and X-ray photo-electron spectroscopy (XPS) were used to examine surface contamination on the thermal oxides. Surface state of the oxide underlayers was measured by "Sessile drop contact angle" method. To observe effects of the underlying oxide surface state on the metal micro-crystalline structure, preferred orientation for the each metal layers was measured by X-ray diffraction (XRD) method.
120
< to
100
*
80
-
~60 4020
0.8
0.82 0.84 0.86 0.88
0.9
0.92
Reflectivity Figure 1 AFM rms versus reflectivity of the 4500A Al-Cu / 500A Ti / thermal oxide stacks deposited at 500 'C RESULTS AND DISCUSSION Surface state variation of thermal oxides Figure 2 shows variation of the aluminum reflectivity on thermal oxides monitored for 4 days. This shows a cyclic fluctuation in aluminum surface quality. All hardware (chamber pressure, heater temperature and temperature uniformity, etc.) and utility (purity of utility gases, power, etc.) components for the sputtering system and the reflectivity measurement accuracy were very stable during this period. To observe effects of the oxide surface state on the aluminum surface quality, we performed aluminum layer depositions on the four sets of thermal oxide substrates. We deposited 5000A Al / 500A Ti layers at 500 'C on two thermal oxide wafers (each of them were selected from separate oxide wafer groups: oxidel-a and oxidel-b), and measured their reflectivity. For 40 hours, we repeated depositing aluminum layers on each two oxide wafers with random time-delay and monitored the time-delay effect after oxide growth. With time-delay for the oxide wafers after growth, we observed a clear ramp-up and steady state (approximately after 2 days after oxide growth) in reflectivity of the aluminum layers as shown in figure 3. We performed the same experiment for the other two oxide wafer groups (oxide2-a and 440
oxide2-b), which were grown 20 hours after the first two wafer groups. From these groups of wafers, we observed the same effect of time-delay. The variation in aluminum reflectivity showed a strong dependency on the surface state of thermal oxides. As shown in Fig. 4, surface state of the thermal and native oxides are gradually transformed from hydr
Data Loading...
