semi合集-English.pdf - 第6475页

SEMI G78-0699 © SE MI 1999 8 DIE Norm a lized Die Offset (No rmalized by Lo t) Averag e Die Of fset - Avera ge Lot Offs et [ACC URACY- TOTA L PROB E SYSTEM ERR OR] Thre e Sigm a Normalized Die Offset (No rmalized by Lo t…

SEMI G78-0699 © SEMI 19997

6.6.5 If the results of this method tu rn out to be

unfavorable, the user has the option of applying a more

detailed, enhanced data analysis application of this

method. That application is contained in Appendix 1.

The enhanced method uses the same data, but it

requires more manipulation of the data during data

analysis. The detail contained in the enhanced

algorithm will provide the user with greater insight as

to the cause of unfavorable results related to the probe

system under evaluation. A summary of the enhanced

method is shown in Figure 6.

6.7 Alternatives — What has been left undefined to

this point is availability of an automated approach to

data collection per the requirements of this method.

Regardless, data collection can be manually

accomplished. The manual process is laborious and

requires a video measuring system to make offset

measurements on the die pads, and a spreadsheet to

analyze the resulting data.

6.7.1 If manual operation is not desi rable or practical,

automated Probe Mark Analysis Systems do exist in the

market place as an item to be purchased. Providers are

also available to accept probed wafers and execute

probe mark analysis under contract as a service.

NOTE: A Final Note to the User of This Method — This

method provides sufficient data for sophisticated analysis of

probe system performance across die-to-die, wafer-to-wafer,

and setup-to-setup. It also provides a simple metric. It is easy

to make comparisons with simple "single number" metrics,

but that has the potential for oversight, distortion and

subsequent inappropriate comparisons. The test engineer,

working with the probe system supplier, must be the ultimate

judge of the applicability of this method and the correct

interpretation of the results.

DIE

σσ

Average Die Offset

Three sigma calculation of Die Offset Values

Die to Die X and Die to Die Y variation

[REPEATABILITY]

VISION SYSTEM or PROBE MARK ANALYSIS SYSTEM

ESTABLISH

PAD OFFSETS

Offset from the center of the bond pad to the center of the scrub mark for Lot-X and Lot-Y

Average Die Offset

Average all Pad Offsets by die

(Avg. of 108 Lot-X and Avg. of 108

Lot-Y values)

SETUP LOT

WAFER

Figure 5

Pad Offset and Two-Step Analysis Algorithm for Determining Probe System Error

SEMI G78-0699 © SEMI 1999 8

DIE

Normalized Die Offset

(Normalized by Lot)

Average Die Offset - Average Lot Offset

[ACCURACY- TOTAL PROBE SYSTEM ERROR]

Three Sigma Normalized Die Offset

(Normalized by Lot)

Three sigma calculation of

Normalized Die Offset Value

[REPEATABILITY]

Three Sigma Total Probe System Error Range

Average Lot Offset +/- Total Probe System Error

VISION SYSTEM or PROBE MARK ANALYSIS SYSTEM

ESTABLISH

PAD OFFSETS

Offset from the center of the bond pad to the center of the scrub mark for all pads analyzed

Average Die Offset

Average all Pad Offsets by

die

Normalized Pad Offset

(Normalized by Die)

Pad Offset – Avg. Die

Offset

Three Sigma Normalized

Pad Offset

(Normalized by Die)

Average Wafer Offset

Avg. of Avg. Die Offset

Three Sigma Normalized

Die Offset

(Normalized by Wafer)

Normalized Die Offset

(Normalized by Wafer)

Avg. Die Offset – Avg. Wafer

Offset

Normalized Wafer Offset

(Normalized by Setup)

Avg. Wafer Offset – Avg.

Setup Offset

Three Sigma Normalized

Wafer Offset

(Normalized by Setup)

Normalized Setup Offset

(Normalized by Lot)

Avg. Setup Offset – Avg.

Lot Offset

Three Sigma Normalized

Setup Offset

(Normalized by Lot)

SETUP LOT

WAFER

Average Setup Offset

Avg. of Avg. Wafer Offset

Average Lot Offset

Avg. of Avg. Setup Offset

Figure 6

Pad Offsets and 15-Step Analysis Algorithm for Determining Probe System Error

SEMI G78-0699 © SEMI 19999

7 Required Hardware Check list (see Figure 7)

 Wafer prober (if the test plan includes either hot or

cold chuck testing, the prober must be

appropriately equipped).

 A planarized and aligned probe card which

matches the selected wafers, accompanied by a

metrology tool printout of the ‘x’, ‘y’ and ‘z’

positions of all of the probe tips to be used for

probe mark analysis. Note: The probe card should

be measured for alignment, but not ”tweaked”,

immediately prior to and after the conclusion of

each subsequent test.

 Untested wafers

 A special “test program” with a fixed “test” time.

 Tester or PC that will run the “test program”. If a

PC is used to simulate the tester, it must be

equipped with the appropriate hardware for

communication with the wafer prober.

 A stopwatch, with 0.1 second resolution. The

stopwatch is to have “split” capability.

Note that the user of this method may choose at

their discretion to utilize any other time stamp

logging technology at their disposal, so long as the

computational overhead of this logging has no

effect upon the throughput of the prober.

8 Requirements (see Figure 7)

8.1 Qualified Prober — Before initiating the test

process, the wafer prober is to be certified by a

representative of the prober manufacturer to be fully

operational.

8.2 Qualified Personnel — The individual running the

wafer prober for the test procedure must be certified or

otherwise qualified to operate the prober being tested.

9 Test conditions (see Figu re 7)

9.1 Test Time — a fixed “device te st time” will be

used during these tests. The recommended value for

this test time is 1.0 seconds.

9.2 Accuracy and Repeatability vs. Throughput — The

person or manufacturer running this test may make a

choice (or choices) as to how they elect to balance

accuracy and repeatability with throughput, but all three

tests must be run simultaneously. This will generate a

set of numbers defining a particular prober's accuracy

and repeatability at a given throughput (or a throughput

at a given accuracy and repeatability). All adjustments

to the prober are to be made using standard, end-user-

adjustable settings, and all prober settings associated

with a particular throughput / accuracy and repeatability

combination are to be included in the final report.

9.3 Temperature — This test can b e conducted at any

temperature, however, numerous variable can distort

the results if run at a temperature other than ambient,

i.e. probe tip drift due to an increase or decrease in

temperature.

• Cautionary Note: If the tests are performed at

temperatures other than ambient, the user must use

the same probe card and wafer type for all tests

across all test platforms to ensure uniformity of

results.

9.4 Probe Card Type — All types of probe cards

(blade, peripheral / cantilever or vertical) may be used,

so long as they leave a visible scrub mark.

• Cautionary note: Some vertical probe cards exhibit

an inherent amount of ‘x’ - ‘y’ needle drift. The

user should verify the intrinsic repeatability of the

probe card before using it to characterize the

prober.

Prober

available?

Probe card

available?

"Test

Program"

available?

Stop-

watch

available?

Is Prober

"Qualified"?

Condition of

probe card

known,

documented?

Personnel

"Qualified"?

Hardware to

run "Test

Program"

available and

tested?

Untested

wafers

available?

7.0 8.2 7.0

9.5

7.0 7.0

7.0

8.1

Figure 7

Preparation