MIL- STD-883F 2004 TEST METHOD STANDARD MICROCIRCUITS - 第404页

MIL-STD-883F METHOD 2030 29 May 1987 4 FI GURE 2030-1. Voi d cri teria .

MIL-STD-883F

METHOD 2030

29 May 1987

3.6 Reports of records.

3.6.1 Reports of inspection

. For class level S devices, or when specified for other device class levels, the manufacturer

shall furnish inspection reports with each shipment of devices. The report shall describe the results of the ultrasonic

inspection, and list the purchase order number or equivalent identification, the part number, the date code, the quantity

inspected, the quantity rejected, and the date of test. For each rejected device, the part number, the serial number when

applicable, and the cause for rejection shall be listed.

3.6.2 Acoustic micrograph and report retention

. When specified, the manufacturer shall retain a set of the ultrasonic

images and a copy of the inspection report. These shall be retained for the period specified.

3.7 Examination and acceptance criteria

. In the examination of devices, the following aspects shall be considered

unacceptable die mounting, and devices that exhibit any of the following defects shall be rejected.

Voids: When imaging devices ultrasonically, certain types of mounting material may not give true representation of voids.

When such devices are inspected, the mounting shall be noted on the inspection report.

a. Contact area voids in excess of 50 percent of the total intended contact area.

b. A single void which exceeds 15 percent of the intended contact area, or a single corner void in excess of 10

percent of the total intended contact area (see figure 2030-1).

c. When the image is divided into four equal quadrants by bisecting both pairs of opposite edges, any quadrant

exhibiting contact area voids in excess of 70 percent of the intended quadrant contact area (see figure 2030-1).

In case of dispute, the percent of voiding shall be determined by actual measurement from the image.

4. SUMMARY

. The following details shall be specified in the applicable acquisition document:

a. Number of views, if other than indicated in 3.1.1.

b. Marking, if other than indicated in 3.2 and marking of samples to indicate they have been ultrasonically imaged, if

required.

c. Defects to be sought in the samples and criteria for acceptance or rejection, if other than indicated in 3.7.

d. Image and report retention, if applicable (see 3.6.2).

e. Test reports when required for class level B devices.

MIL-STD-883F

METHOD 2030

29 May 1987

FIGURE 2030-1. Void criteria

MIL-STD-883F

METHOD 2031.1

27 July 1990

METHOD 2031.1

FLIP-CHIP PULL-OFF TEST

1. PURPOSE

. The purpose of this test is to measure the strength of internal bonds between a semiconductor die and a

substrate to which it is attached in a face-bond configuration.

2. APPARATUS

. The apparatus for this test shall consist of suitable equipment for applying the specified stress to the

bonds. A calibrated measurement and indication of the applied stress in grams force (gf) shall be provided by equipment

capable of measuring stresses up to twice the specified minimum limit value, with an accuracy of ±5 percent or ±0.25 gf,

whichever is the greater tolerance.

3. PROCEDURE

. The test shall be conducted using the following procedure. All die pulls shall be counted and the

specified sampling, acceptance, and added sample provisions shall be observed, as applicable. The sample size number

and accept number specified shall determine the number of die to be tested (not bonds). For hybrid or multichip devices, a

minimum of 4 die or all die if four are not available on a minimum of 2 completed devices shall be used. All pull tests shall

be performed prior to the application of encapsulants, adhesive, or any material which may increase the apparent bond

strength.

When flip chips are bonded to substrates other than those in completed devices, the following conditions shall apply:

a. The sample of chips for this test shall be taken at random from the same chip population as that used in the

completed devices that they are intended to represent.

b. The chips for this test shall be bonded on the same bonding apparatus as the completed devices, during the time

period within which the completed devices are bonded.

c. The test chip substrates shall be processed, metallized, and handled identically with the completed device

substrates, during the same time period within which the completed device substrates are processed.

3.1 Testing

. The calibrated pull-off apparatus (see 2) shall include a pull-off rod (for instance, a current loop of nichrome

or Kovar wire) having a cross-sectional area of 75 percent, +3 percent, -5 percent of the chip surface area. The rod shall

make connection with a hard setting adhesive material (for instance, a cyanoacrylate or other adhesive possessing high

tensile strength) on the back of the flip chip. The substrate shall be rigidly installed in the pull-off fixture and the pull-off rod

shall make firm mechanical connection to the adhesive material. The die shall be pulled without shock, within 5° of the

normal at a rate of 500 grams ±100 grams per second, until the die separates from the substrate. When a failure occurs,

the force at the time of failure, the calculated force limit, and the failure category shall be recorded.

3.2 Failure criteria

. Any flip-chip pull which results in separation under an applied stress less than

500 kg/in

x average solder bump area (in

) x number of solder bumps shall constitute a failure.

3.2.1 Failure category

. Failure categories are as follows: When specified, the stress required to achieve separation and

the predominant category of separation or failure shall be recorded.

a. Silicon broken.

b. Lifted metallization from chip.

c. Separation at bond-chip interface.

d. Failure within bond.