Transcript 開啟檔案

版權所有 翻印必究
日
期： 2 0 1 1 . 0 5 . 02
指導老師：林克默、黃文勇 博士
學
生： 陳
立
偉
2016/4/11
STUT 太陽能材料與模組實驗室
1
版權所有 翻印必究
Outline
1.
2.
3.
4.
5.
6.
Introduction
Reliability test plan
Power cycling
thermal shock
Vibration testing
Conclusions
2016/4/11
STUT 太陽能材料與模組實驗室
2
版權所有 翻印必究
1. Introduction
A Directive, endorsed by the European
Parliament, on “Waste from electrical and
electronic equipment” can be expected which could
include a ban on lead based soldering in electronics
in the near future.
The Japanese electronics industry has already
indicated they will phase-out lead in consumer
electronics in the coming years, which could have
worldwide implications.
2016/4/11
STUT 太陽能材料與模組實驗室
3
版權所有 翻印必究
Earlier in the project it was decided to target
alloys with melting points in the range 190-220°C,
and ternary alloys of Sn, Cu, and Ag, with small
additions of Sb and Bi were investigated. It was
recognised that other alloys such as eutectic
SnBi58, SnAg3.5, or SnCu0.7 could also be
considered, but not for a wide application range.
2016/4/11
STUT 太陽能材料與模組實驗室
4
版權所有 翻印必究
The scope of this paper is confined to the
assessment of the reliability of mainstream
electronics surface mount reflow assemblies based
on the ternary SnAg3.8Cu0.7 (Tm = 217°C) solder
alloy. In the following SnAg3.8Cu0.7 is referred to
as SAC, while SPA denotes Sn62PbAg2.
2016/4/11
STUT 太陽能材料與模組實驗室
5
版權所有 翻印必究
2. Reliability test plan
• According to Ref. [4] and the IPC-SM-785
guidelines for accelerated reliability testing of
surface mount solder attachments, the most
common reliability threat comes from stressrelaxation based (thermal) fatigue damage.
• However, in Ref. [5] is indicated that
environmental temperature variations do not
represent inservice conditions,
2016/4/11
STUT 太陽能材料與模組實驗室
6
版權所有 翻印必究
• Other tests such as thermal and mechanical shock
and vibration fatigue In Ref. [6] is explained that
the fast thermal mechanical excursions without
enough hold time to induce creep and stress
relaxation effects in the solder joints.
2016/4/11
STUT 太陽能材料與模組實驗室
7
版權所有 翻印必究
3. Power cycling
3. 1. Test setup
• A 64 channel multiplexer system was built for intest (not continuous) monitoring of electrical
continuity and resistance. Each standard single
layer FR4 PCB board contained two daisy
chained circuits of four 1206 100 Ohm resistors
with Sn100 metallisation and one daisy chained
100 gullwing leads, 0.65 mm pitch. MQFP
plastic package with a copper leadframe and
Sn85/Pb15 metallisation.
2016/4/11
STUT 太陽能材料與模組實驗室
8
版權所有 翻印必究
• Each batch underwent a number of power cycles;
i.e. 0, 1000, 3000 or 5000 cycles and consisted of
different board metallisation and solders (see
Table 1.
• SAC-X where X is a proprietary elemental
addition. The same all-round flux (which
combined high-speed printability with low void
formalion) and reflow conditions were chosen for
the lead-free boards, while a standard reflow
profile for the SPA boards was used.
2016/4/11
STUT 太陽能材料與模組實驗室
9
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
10
版權所有 翻印必究
• An IR camera and thermocouples were used to
assess the temperature distribution over the board
and components as power was applied. 50 W
power dissipation in the 18 MQFP dice was
required to reach a temperature of 100°C in the
leads. The 1206 resistors reached a maximum
temperature of 100°C for 60 W of applied power.
2016/4/11
STUT 太陽能材料與模組實驗室
11
版權所有 翻印必究
• The temperature cycle was selected to go from
ambient to 100°C with a dwell time of 15 min to
allow sufficient stress relaxation. This selection
was based on guidelines for accelerated testing
for a consumer product use environment [7].
Ramp up and cool down rates were about
25℃/min, which resulted in a 30-min cycle. Fig.
1 shows the actual solder joint temperature
profiles for both MQFP100 sides and a 1206
resistor.
2016/4/11
STUT 太陽能材料與模組實驗室
12
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
13
版權所有 翻印必究
3.2. Test results
• MQFPs, which did not experience wirebond
failures had no electrical resistance increases
over 5000 cycles No electrical failures were seen
for the 1206s up to 4000 cycles, but between
4000 and 5000 cycles, four boards, including
SPA, indicated resistance increases and interrupts.
2016/4/11
STUT 太陽能材料與模組實驗室
14
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
15
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
16
版權所有 翻印必究
• Before ageing the SPA resistor joint on OSP-Cu
showed a eutectic structure with the Pb phase
finely dispersed in the Sn phase. The SAC on
immersion Sn solder body (before power cycling)
showed a Sn-rich matrix of large ß-Sn phase
dendritic globules. with a dispersed fine eutectic
microstructure of Cu6Sn5 and Ag3Sn and a
Cu6Sn5 intermetallic layer at the pad/solder
interface. This type of microstructure was
evident in all SAC resistor joints before ageing.
2016/4/11
STUT 太陽能材料與模組實驗室
17
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
18
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
19
版權所有 翻印必究
• At 1000 cycles, cracks had developed in all 1206
resistor joints. SPA on OSP-Cu resistor joints
after 1000 power cycles showed that cracks
initiate under the component termination and
progress along the grain boundaries between the
Pb-rich and Sn-rich areas of the joint, indicating
intergranular cracking. Coarsening of the
microstructure is also evident and the crack
propagates through the coarsened region.
2016/4/11
STUT 太陽能材料與模組實驗室
20
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
21
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
22
版權所有 翻印必究
4. thermal shock
• Thermal shock testing was carried out whereby
the rapid temperature changes induced shear and
tensile stresses in the solder joints. For this
relatively short environmental test twin baths
were used and consisted of cycling between 20°C and +125°C, with 3 min dwells at both
temperature extremes and a cycle time of 7 min .
Boards were cycled at 1000, 2000 and 3000
cycles.
2016/4/11
STUT 太陽能材料與模組實驗室
23
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
24
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
25
版權所有 翻印必究
5. Vibration testing
• frequency band between 100 and 2000 Hz for up
to 60 mm per axis. No electrical, mechanical or
microstructural degradation was observed.
Longer test durations, up to 8 h per axis in x and
y direction (see Fig. 15) still did not result in
significant solder joint damage.
• Random vibration testing for 6 h in the axis
direction showed daisy chain resistance increases
of up to 10 times the original values.
2016/4/11
STUT 太陽能材料與模組實驗室
26
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
27
版權所有 翻印必究
For the final tests
1. after assembly
2. after 6 h random vibration (RV) in z direction
(between 50 and 2000 Hz)
3. after 500 thermal shock (TS) cycles (cycled
between -40°C and 125°C with dwell time 15 mm
and cycle time 30 mm)
4. combined random vibration (2) and thermal
shock(3)
2016/4/11
STUT 太陽能材料與模組實驗室
28
版權所有 翻印必究
2016/4/11
STUT 太陽能材料與模組實驗室
29
版權所有 翻印必究
• In correspondence with previous test results with
eutectic Sn37Pb solder, the random vibration test
on its own has no mechanical degradation effect
on the SPA assemblies, but it has some fatiguing
effect on the SAC assemblies. The combined
effect of thermal shock and random vibration
showed that SAC has slightly better fatigue
resistance then, but comparable strength to that
of SPA.
2016/4/11
STUT 太陽能材料與模組實驗室
30
版權所有 翻印必究
6. Conclusions
•初步可靠性的測試結果，表示使用結合各種
元件與基板金屬化與SAC焊料有可能取代傳統
的鉛基系回流焊接。
•現有回流焊接設備無必需的重大修改，但必
須修改最低回流焊溫度至225℃
•振動測試對於元件與基板的接合度具有強大
的破壞性。結合熱衝擊和振動試驗可以加速
疲勞壽命，但還要需進一步的研究。
2016/4/11
STUT 太陽能材料與模組實驗室
31
版權所有 翻印必究
• 熱衝擊和隨機振動試驗表明，耐疲勞性之間
的差異，含鉛和SAC焊點無顯著差異，但電
路板的金屬表面影響也會影響其焊接效果。
2016/4/11
STUT 太陽能材料與模組實驗室
32