• PDF / 318,678 Bytes
• 6 Pages / 414.72 x 648 pts Page_size
• 74 Downloads / 288 Views

DOWNLOAD

REPORT

---

Mat. Res. Soc. Symp. Proc. Vol. 477 ©1997 Materials Research Society

evaluation of the effect on GOI of process variables like chemical types and concentrations, chemical bath sequence, chemical solution lifetime and wafer residence time have been performed. Surface metal concentration on the wafers at the end of the global cleaning sequence has been kept under control guarantying values < 5E 10 at/cm2. A subset of the experimental matrix has beeo addressed to the evaluation of surface polishing process in which the effect on GOI of intentional contamination with organic during the chemical mechanical process have been studied. Oxide integrity has been measured by applying a two steps constant current test [6] on doped poly gate MOS devices of 0.1cm^2 area. Thermal oxide of 215 A thickness was grown at 900C in dryO2+transLC. Results will be generally expressed as percentage of survived devices after 500msecs of ImA/cmA2 current injection (Y2). Additional techniques used to measure both silicon surface properties and chemicals concentration in solution are described here below. Automatic inspection of the wafer surface by laser scattering have been performed using WIS CR80 instrument counting Light Point Defects(LPDs) having size greater than 0.12 gim. It is well known[7] that COPs are revealed as LPDs in the small size bin channels (0. 12-0.2jim) Surface micro roughness has been measured by Digital Instrument Dimension 3000 Atomic Force microscope using tapping mode and exploring Igim X 1gm area. Total Organic Carbon (TOC) in DI water has been measured by a C02 Infrared analyser (Dhorman DC 180). Ozone concentration in solution has been detected by iodometric titration while NH40H and H202 concentrations have been determined by potentiometric titration. RESULTS A clear evidence of the importance of an acid oxidizing bath for organic digestion as initial stage of the cleaning sequence in affecting gate oxide quality is shown in fig. 1 where a comparison between weak oxidizing bath and strong acid oxidizing bath is performed.

ORGANIC RESIDUALS IMPACT

ON GOI 60 500

ik

' o

•

i-20

to

. LOW OROANIC REMOVAL EMCIENCY

I

2

3

4

5

8

7

8

9

HIGHORGANIC REMOVAL EFIE•rECyt

01o 1 121314151017181920

N. Or BATCH

Fig. 1: Organic residuals impact on gate oxide integrity, measured on Ppolished wafers

In particular, in case of the use of 03 in acqueous solution at room temperature, the GOI response 242

degrades on wafers cleaned at decreasing ozone concentration due to its depletion in the bath and consequent organic removal efficiency reduction as described in fig.2. ORGANIC RESIDUAL IMPACT ON GOI 45

40

0)

i35

03

2

3

4

5

6

7

8

0

10

II

12

OZONE LEVEL (PPM)

Fig. 2: GOI yield versus ozone concentration An "ad hoc" test has been performed at wafer polishing phase by using deionized water with high TOC content in a first sequence of polishing runs followed by a second series of runs in which a different deionized water loop with TOC content below detection limit has been used. The corresponding GOI