IPC-4552.pdf - 第18页
IPC-4552 Proposal -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0 0.375 0.75 1.125 1.5 1.875 2.25 2.625 3 3.375 3.75 4.125 4.5 4.875 5.25 5.625 6 6.375 6.75 7.125 7.5 7.875 8.25 8.625 9 9.375 9.75 time i n seco nds mN/mm 2 mins 4 …
IPC-4552 Proposal June 2001
and calibration files for ENIG measurements to using an x-ray tube that is too high in energy that blasts its
way through the gold layer.
In an attempt to overcome some of these issues and to gain an understanding of the correct way to measure ENIG,
the committee enlisted the aid of Veeco, one of the industry suppliers of XRF equipment. Veeco has at its
headquarters a very expensive XRF unit, costing in excess of $500,000 that would be used as the de-facto
measurement tool. As a comparison to this reference unit would be a modern, thin film measurement capable XRF
unit typically found in a PWB shop or at incoming inspection at the OEM, price range of $50,000.
For the XRF measurement test, thirty measurements would be taken at predetermined pad locations per panel,
(fifteen on each side of the board). The data generated by each XRF machine would be compared to determine if
they were statistically the same. Refer to Figure 3.
The results from the studies were as follows:
1) There is a difference between chemical suppliers of ENIG as regard deposit thickness rates and
maximum thickness achievable under the conditions of this test.
2) The requirements of some OEM’s / EMS’s for 10 microinches of gold cannot be met under the
conditions of the test.
3) The range of nickel deposited would fall on average between 100 and 200 microinches.
4) The capability of a “working mans” XRF as compared to a state of the art XRF for ENIG measurement
is good with their being a very small statistical difference in the values recorded.
Minimum Gold Thickness Requirements:
Once the thickness of the deposits were verified the most important question could begin to be answered. While the
deposit is multifunctional in nature, something that is covered in the specification, the primary use is one of
soldering and as solderability preservative. What would be the minimum thickness of gold needed to afford shelf life
for upwards of one year? As mentioned above, the plating rate study was performed on wetting balance coupon
panels.
Solderability Testing:
The solderability testing would be carried out on a wetting balance using a test flux complying with ROL0 of the J-
Std-004 specification – i.e. a test flux. For each test, ten (10) samples would be tested and the average value of these
ten (10) readings would be reported. Therefore ten (10) tests were performed on the two- (2) minute plating time
group, ten (10) on the 4 minute plating group etc. A total of 100 soldering tests per vendor tested would be
performed for the “as received” condition. It was decided that it would not be practical to test all five suppliers given
the time constraints that the committee was working to, so as a compromise two vendors were chosen at random
from the groups: Vendor D and Vendor E. Fig 5 & 6, show thickness distribution. Figs 7 & 8 show wetting balance
as a function of dwell time in the gold bath.
The test protocol was as follows:
Each panel contains 27 strips of test coupons with 14 coupons per strip.
The coupons were scored and routed to enable them to be tested more easily and consistently.
The coupon is 25.4 mm wide, 1.6 mm thick and has a wettable area of 36 mm that is made up from twelve (12) 3mm
wide pads and is made from acid copper plated FR4 material Fig 4.
Each coupon is dipped into the test flux for five (5) seconds and excess flux is removed by touching the sample to a
piece of lab paper
The sample is allowed to dry completely prior to placing into the specimen holder of the wetting balance.
An automatic dross skimmer cleans the solder pot surface as the test begins.
The sample is dipped at 90° degrees incident to the solder with no preheat.
The depth of immersion is 0.5mm and the dip time is 10 seconds.
The wetting balance used has the ability to discriminate between the wetting forces produced by the sample (which
are small) compared to the forces that are produced by the buoyancy of the coupon which are large.
Page 15 of 10
Download From http://bbs.infoeach.com
Download From http://bbs.infoeach.com
IPC-4552 Proposal
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0
0.375
0.75
1.125
1.5
1.875
2.25
2.625
3
3.375
3.75
4.125
4.5
4.875
5.25
5.625
6
6.375
6.75
7.125
7.5
7.875
8.25
8.625
9
9.375
9.75
time in seconds
mN/mm
2 mins
4 mins
6 mins
8 mins
10 mins
12 mins
14 mins
16 mins
18 mins
June 2001
Page 16 of 10
Fig 7: Wetting times as a function of plating dwell times for Vendor E, 90 days old
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0
0.375
0.75
1.125
1.5
1.875
2.25
2.625
3
3.375
3.75
4.125
4.5
4.875
5.25
5.625
6
6.375
6.75
7.125
7.5
7.875
8.25
8.625
9
9.375
9.75
time in seconds
mN/mm
2mins
4 mins
6 mins
8 mins
10 mins
12 mins
14 mins
16 mins
18 mins
20 mins
Fig 8: Wetting times as a function of plating dwell times for Vendor D, 90 days old
Download From http://bbs.infoeach.com
Download From http://bbs.infoeach.com
IPC-4552 Proposal June 2001
RESULTS:
As Received:
Due to the extensive XRF study on these panels, a time lag of approximately 90 days occurred between the time
they were plated and the time they were tested. For ease of publishing this document, only a small sample of the data
is presented here, the reader may request the more complete package from the author or from the IPC. The results
from the 90 day old samples showed very little in the way of performance differences between the two minute
plating dwell time samples (average of one (1) microinch gold) compared to the twenty minute plating dwell time
samples (average of four (4) microinches gold). See Fig 7 for Vendor E and Fig 8 for vendor D.
After Aging:
Since the introduction of the alternate finishes to tin/lead, the search for a suitable accelerated shelf life predictor has
been underway.
Steam aging: Steam aging the shell life predictor for Tin/Lead systems does not work for alternate finishes. ENIG
will produce non-solderable results from as little as one-hour exposure to steam; yet members of the committee have
anecdotal evidence of year plus shelf life being common and not an issue.
Temperature and Humidity (85°C/85% R.H): Members have tested at 85°C/85% R.H., 65°C/92% R.H., 160° C dry
bakes, all with varying degrees of impact on the solderability of the deposit. While not having a correlation for any
of these accelerating factors, the committee agreed upon 85°C/85% R.H to see if we could make a distinction in
soldering performance between the thinner (one microinch gold) versus the more typical (four microinch gold)
deposits.
The results from the wetting balance studies showed only a marginal improvement in performance for the thicker
deposits over the one microinch deposit. There was an increase in wetting time and the slope of the curve to
maximum wetting forces was affected more for the one microinch deposit. In both cases however the overall wetting
forces were reduced as a result of the exposure to 85°C/85% R.H. However it should be noted that the increase in
wetting times would not cause any issues in assembly of the product as they were shorter than the typical contact
times in a wave and did not have the benefits of either preheat or wave dynamics. Since the majority of ENIG coated
PWB’s are assembled using paste reflow techniques that have time above liquidus of 30 plus seconds as a
minimum, the increase in wetting times would have no significant effect on assembly.
Actual Aging:
At the time of writing this paper, the author came upon the original set of coupons from the test that were lying,
unprotected, in a tote bin in my office. These samples were now 240 plus days old since plating, (I knew there was a
reason I don’t clean my office). Samples from the two (2) minute plating dwell time (one microinch average gold
thickness) from Vendor’s D & E were again tested. The wetting times and performance remain excellent. It should
be noted that the effect of eight (8) plus months of unprotected storage was less than the effect of 18 hours of
85°C/85% R.H conditioning!
The author has for many years been conducting a solderability study in real time on all surface finishes using the
same test coupon and flux. As a point of interest I took the curves produced by the exposure to 18 hours of
85°C/85% R.H conditioning and attempted to match the curves with that from real time testing of
ENIG. The closest approximation would be 575 days of unprotected storage. The fact that a one microinch deposit
of gold can afford such a level of solderability protection is very impressive.
Page 17 of 10
Download From http://bbs.infoeach.com
Download From http://bbs.infoeach.com