Well,
they relate, but lower CTE and higher Tg are not necessarily always the best.
If we assume that the reliability issue is stress due to different thermal
expansion between copper and the dielectric, leading to cracks, then we have to
look for a solution that minimizes the difference in expansion between copper
and dielectric over the range from room temperature (e.g. 20C) to the solder
shock temperature, let’s say 288C in the case of lead-free processing.
With copper having a CTE of approximately 17ppm/C, copper will expand (if
unrestrained) 4,556ppm when heating from 20C to 288C (i. e. 268x17).
Let’s look at three cases:
Case 1:
Room Temp (C)
|
ΔT (RT to Tg)
|
Tg
|
ΔT (Tg to SST )
|
Solder Shock Temp (SST)
|
|
|
20
|
150
|
170
|
118
|
288
|
|
|
|
CTE below Tg (α1)
|
|
CTE above Tg (α2)
|
|
|
|
|
35
|
|
70
|
|
|
|
|
Expansion (ppm) from RT to Tg
|
|
Expansion (ppm) from Tg to SST
|
|
Total expansion (ppm) from RT to SST
|
|
|
5,250
|
|
8,260
|
|
13,510
|
Case 2:
Room Temp (C)
|
ΔT (RT to Tg)
|
Tg
|
ΔT (Tg to SST )
|
Solder Shock Temp (SST)
|
|
|
20
|
150
|
170
|
118
|
288
|
|
|
|
CTE below Tg (α1)
|
|
CTE above Tg (α2)
|
|
|
|
|
20
|
|
100
|
|
|
|
|
Expansion (ppm) from RT to Tg
|
|
Expansion (ppm) from Tg to SST
|
|
Total expansion (ppm) from RT to SST
|
|
|
3,000
|
|
11,800
|
|
14,800
|
Case 3:
Room Temp (C)
|
ΔT (RT to Tg)
|
Tg
|
ΔT (Tg to SST )
|
Solder Shock Temp (SST)
|
|
|
20
|
130
|
150
|
138
|
288
|
|
|
|
CTE below Tg (α1)
|
|
CTE above Tg (α2)
|
|
|
|
|
20
|
|
70
|
|
|
|
|
Expansion (ppm) from RT to Tg
|
|
Expansion (ppm) from Tg to SST
|
|
Total expansion (ppm) from RT to SST
|
|
|
2,600
|
|
9,660
|
|
12,260
|
And the winner is: Case 3 (expansion closest to that of copper), even though
Case 3 has a lower Tg than Case 1 or 2. Comparing Case 1 and 2: Case 2 has a
lower CTE below Tg than Case 1, but leads to a larger expansion differential
than Case 1. The lesson here is that we have to consider both CTE below Tg and
above Tg.
By: Roberto Tulman
Posted: September 5, 2008 4:50 AM
The copper in-plane with the layers is not relevant as CTE of the dielectric is constrained by the glass cloth.
The copper plated inside the holes has a minimum elongation factor of 15%...
In the Lead-free world it will be more important to look at the Td than at the Tg.
Roberto Tulman
www.eltekglobal.com