Creep Corrosion Test Using iNEMI FoS Chamber

Chair: Prabjit Singh (IBM) iNEMI WebinarSeptember 26, 2018

Please contact Haley Fu (haley.fu@inemi.org) if you would like to help us define further work on corrosion mitigation.

Listen to the webinar recording: http://bit.ly/2zM6FXSThis link will be active for up to six months following the webinar.

Download whitepaper and full project report

White Paper: A Cost-Effective & Convenient Approach to Creep Corrosion Testing (August 2018)

Project Report: Qualification Test Development for Creep Corrosion(August 2018)

n Prabjit Singh and Larry PalmerIBM Corporation, Poughkeepsie, NY, USA

n Simon Lee and Geoffrey Tong The Dow Chemical Company, Taipei and Hong Kong

n Dem Lee and Jeffrey Lee iST-Integrated Service Technology, Inc., Taiwan

n Karlos Guo and Julie Liu Lenovo (Beijing) Limited Corporation, Beijing, China

n Chen Xu Nokia, Murray Hill, NJ, USA

n Haley Fu iNEMI, Shanghai, China

Authors

3

Agenda

§ Background§ Industry available corrosion test§ iNEMI Flowers of sulfur (FoS) chamber design§ Test chamber conditions and their control§ Test procedure§ Round robin tests§ Effect of PCB storage time§ Role of solder mask edge§ iNEMI FoS test effectiveness § FoS chamber as a general purpose corrosion chamber

4

Background

• Poor creep corrosion resistance of the Pb-free PCBAs (flux, surface finish, etc.)

• Expanding electronics market in geographies such as Asia, which typically have high humidity as well as sulfur-bearing gaseous pollution, led to increased corrosion-based failures

• To eliminate creep corrosion, industry has initiated significant effort to better understand its cause and mitigation, including technical groups from ASHRAE, IPC, ISA and iNEMI.

2018-09-05 Restricted 5

Gaseous corrosivity levels ISA 71.04 standardSeverity

level Reactivity levelISA 71.04-2013

Copper corrosion rate, Å/month Silver corrosion rate, Å/monthG1 mild <300 <200G2 moderate 300-1000 200-1000G3 harsh 1000-2000 1000-2000GX severe >2000 >2000

ASHRAE thermal guidelines for IT equipment – 4th edition

ASHRAE recommended corrosion monitoring locations in data centers

n Cu and Ag metal foils should be placed in front of the rack at ¼ and ¾ height from the floor.

n After one month exposure, the Cu and Ag foils are returned to the lab to measure the corrosion product thickness using coulometric reduction.

X1/4 height

X3/4 height

ASHRAE survey: Cu corrosion rate

n The open and solid data points are for data centers with and without reported corrosion-related hardware failures.

n Copper corrosion rate is NOT a good predictor of hardware corrosion because of the overlap of Cu corrosion rates for data centers with and without corrosion-related failures.

0

200

400

600

800

1000

1200

10-Aug-10 18-Nov-10 26-Feb-11 6-Jun-11 14-Sep-11

Date

Copp

er c

orro

sion

rate

, ang

stro

ms/

mon

th

n The open and solid data points are for data centers with and without reported corrosion-related hardware failures.

n Silver corrosion rate is a good predictor of hardware corrosion because of no overlap of the silver corrosion rates for data centers with and without corrosion-related hardware failures.

ASHRAE survey: Ag corrosion rate

0

200

400

600

800

1000

1200

1400

1600

1800

2000

10-Aug-10 18-Nov-10 26-Feb-11 6-Jun-11 14-Sep-11

Date

Silv

er c

orro

sion

rate

, ang

stro

ms/

mon

th

Previous Study at iNEMI: Creep Corrosion Project Using Mixed Flowing Gas (MFG) Test (2009-2012)

• Phase 1: Industrial survey• Phase 2: Primary factors influencing creep

corrosion • Phase 3: Investigation of factors that

influence creep corrosion on PCB (H2S = 1200 ppb; NO2 = 200 ppb; Cl2 = 20 ppb; SO2= 200 ppb; 40℃, RH 70-75%, 20 days)

Qualification test for creep corrosion using FOS (2013-2017)

• The project explored a more cost-effective and convenient approach to creep corrosion testing that suppliers could use to satisfy customers that their products would survive moderately corrosive environments.– Phase 1: Test chamber design & preliminary test (2013-2014)– Phase 2: 6 test runs to study test setup effectiveness (2014-2015)– Phase 3: The effect of relative humidity on creep corrosion and

finalizing the creep corrosion qualification test (2016-2017)

MFG vs FoS: Key Difference

MFG FoSCorrosive Agents Used H2S, NOx, SO2, Cl2 S (and Cl2 in iNEMI FoS)

Test parameter monitoring

H2S, NOx, SO2, Cl2, RH, T, air flow rate RH, T

Test parameter controlall parameter can be

independently controlled and adjusted

T, RH, S concentration can be controlled

Corrosion products sulfide, sulfate, chloride, oxide, nitrate

sulfide, oxide (and chloride in iNEMI FoS)

Equipment and maintenance cost high low

Equipment setup complex simpleTest chamber size large moderate

Air flow laminar and uniform non-laminar

MFG: Most realistic and controlled simulation of real world environments.iNEMI FoS: Inexpensive and easy to maintain.MFG and iNEMI FoS are complementary.

Industry Standard Using FoS Method

• There are several industry standards using FoS method as below:

12

Standard ASTM B809 EIA-977 iNEMI FoSTemperature 50oC 60oC/105oC 50oC

Humidity 85% � 11%-90%

Corrosive Gas S8 S8 S8 + Cl2

Air Flowing Factor X X O

Major Mode Cu2S… Ag2S… Cu2S…

Corrosion ModePore/Pitting

CorrosionSulfur Corrosion Creep Corrosion

Feature

Industry

Metal Porosity

Golden Finger,

Connector

Passive, Switch PCB/PCBA

Test board used for developing the creep corrosion test

• Test boards are 140x110-mm by 1-mm thick. The FR4 epoxy used was compatible with Pb-free soldering conditions.

• 15 areas on the board with various traces and vias with various pitches designed to monitor the creep corrosion during test.

13

L8

L10

L11

L1

L4

L2

L13

Top Side (Reflow) Bottom Side (Wave Soldering)

L6 L5

L12L14L15

L9L7L3

iNEMI FoS chamber conditions

• An acrylic cube with a paddle wheel can hold 8 boards rotating at 20 rpm

• 50oC chamber temperature• 200 g sulfur• %RH controlled by salt solution in the

11 to 90% range• 100 ml 8.25% sodium hypochlorite

Clorox® is used to introduce chlorine gas that would work in synergy with sulfur

• Cu and Ag foils are mounted for Cu and Ag corrosion rate measurement

• The corrosion rate of Cu and Ag should be 500-1500 and 400-1200 Å/day, respectively. The rate of formation of AgCl should be 50-500 Å/day.

14

iNEMI FoS chamber

15

• The challenge is to have the Clorox release chlorine gas into the chamber but not upset the relative humidity.

• The saturated salt controls the relative humidity to its deliquescence value.

• The moisture coming off the Clorox is throttled and forced to flow over the saturated salt.

Sulfur content, g0 100 200 300 400 500

Cor

rosi

on ra

te, Å

/day

0

200

400

600

800

1000

1200

1400

Influence of sulfur content

Red dots are for Cu corrosionBlue dots are for Ag corrosionOpen dots are from mass gainSolid dots are from coulometric reduction

NH4NO3 saturated salt solution, 100ml Clorox, 50oC, 20 rpm.

Sulfur content, g0 100 200 300 400 500

AgC

l gro

wth

rate

, Å/d

ay

0

50

100

150

200

16

• We settled on 200 g sulfur.

Influence of humidity

FOS with 200g sulfur, 100ml Clorox, 50oC, 20 rpm.

17

The circular data points are from IBM; the triangular data points from iST; and the

square data points are from Lenovo. The solid data points are corrosion rates obtained

using coulometric reduction technique and the open data points are from mass-gain

technique.

AgCl formation rate

Corrosion rates in FoS chamber loaded with 6 test PCBs and one set of 25x50-mm copper and silver foils.

18

Chlorine gas content

13-18% RH 74-80% RH

19

• Chlorine gas that would work in synergy with sulfur• Chlorine is very soluble in water.• Chloride concentration in chamber is influenced by the saturated

salt solution and by adsorbed moisture on surfaces.

Silver corrosion rates in the FoS chamber as a function of time and Clorox dilution

• Daily AgCl and Ag2S formation rates during 5-day runs. – 100% Clorox plots are shown with solid dots.– 50% Clorox plots are shown with open dots.

20

FOS with 200g sulfur, NaNO2 saturated salt solution, 100ml Clorox in various dilutions, 50oC, 20 rpm.

Effect of %RH on Creep Corrosion After 10 Days In FoS Chamber

21

Table 1. Severity of creep corrosion after 10 days in FOS chamber at 50oC, with 100 ml household bleach, 1-mm setup gap and paddle wheel rotating at 20 rpm.

SaturatedsaltsolutionsandRH%

ZnCl213-30%

MgCl227-35%

NaNO255-59%

NH4NO352-68%

NaCl69-77%

KCl74-80%

%ofcreepsiteswithcreepcorrosion HC LC HC LC HC LC HC LC HC LC HC LC

IBM

ImAg

OR9 2 13 0 1 9 0 0 0 0 0 1

Lenovo 13 0 3 5 - - 0 5 - - 6 7iST 9 5 0 0 - - 0 0 0 0 0 0IBM

RO9 2 12 1 2 9 0 0 0 0 0 1

Lenovo - - 5 4 - - 4 2 - - - -iST 10 5 0 0 - - 0 1 0 0 0 0IBM

ENIG

OR0 0 12 1 0 0 0 8 0 0 3 1

Lenovo 2 3 10 4 - - 10 3 - - 7 7iST 0 0 0 2 - - 0 9 0 6 3 6IBM

RO0 0 12 1 0 0 0 8 0 4 5 2

Lenovo - - 7 6 - - 9 4 - - - -iST 3 1 2 4 - - 3 8 1 4 1 8IBM

OSP

OR0 0 2 3 0 0 0 0 0 0 0 0

Lenovo 5 5 0 3 - - 0 0 - - 0 0iST 3 1 0 0 - - 0 0 0 0 1 0IBM

RO0 0 3 4 0 0 0 0 0 0 0 0

Lenovo - - 0 2 - - 0 1 - - - -iST 1 8 0 1 - - 0 0 0 0 0 0

Ø 50oCØ 100ml household bleachØ Rotating at 20 rpmØ IBM used 1-mm Clorox

setup gap; Lenovo used 0.5mm/1mm; iST used 1mm/3mm gap.

Ø Sulfur content variation among three sites: 180g –800g

Influence of pre-baking

Pre-bake is necessary to avoid the VOCs coming off the PCBs from contaminating the limited-volume FoS chamber.

Creep corrosion occurred only on pre-baked PCBs from a lot that suffered creep corrosion in the field.

Unbaked card A Unbaked card B

Pre-baked card C Prebaked card D

22

iNEMI FoS Test procedure

• Prebake test PCBs at 100ºC in flowing nitrogen gas for 24 hours to eliminate the impact of volatile organic compounds (VOCs).

• Load paddle wheel with seven test PCBs plus one board on which Cu and Ag foils are mounted for Cu and Ag corrosion rate measurements. (Mounting board must be of a material that does not excessively absorb the gases and moisture in the chamber.)

• Expose PCBs to three five-day runs at 50ºC, each at different levels of relative humidity:

– Run 1: 31% RH using MgCl2 saturated salt solution

– Run 2: 59% RH using NaNO2 saturated salt solution

– Run 3: 81% RH using KCl saturated salt solution

• Copper and silver corrosion rates and rate of formation of AgCl measured using coulometric reduction and mass gain methods. Photograph and record the corrosion failures.

23

Pass/Fail Recommendation

• The acceptable extent of creep corrosion should be defined and accepted by the PCB manufacturer and procurer.

24

Rank 0 Rank 1 Rank 2 Rank 3 Rank 4

No Corrosion Local Corrosion

Light Creep

Corrosion

(<1/3 gap)

Moderate Creep

Corrosion

(>1/3 gap)

Heavy Creep

Corrosion

Oxidation or

tarnishing is OK.

Corrosion limited

to pads or holes

only, no spread.

Corrosion spreading

from pads or holes,

but not bridging

between features.

Corrosion spreading

from pads or holes,

but not bridging

between features.

Severe corrosion,

bridging between

features.

Pass Pass Pass Fail Fail

1st round robin test

Table 9.2c: Corrosion map after the 3rd 5 days with KCl saturated salt solution. The relative humidity range across 3 test sites was 61-78%.

Site Finish Flux Location on board (Grey is for no corrosion, yellow for corrosion; pink for light creep corrosion and dark red

for heavy creep corrosion. The resistance readings are in KΩ.) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

A

ImAg

OR Open Open Open Open Open Open Open Open 57 Open Open Open Open N/A N/A

B Open Open Open 300 Open Open Open Open Open Open Open Open 520 N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A A

RO Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

B 0.2 Open 0.13 Open Open Open 58 Open Open Open Open Open 720 N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A A

ENIG

OR 0 0.02 0.15 0.02 Open 0 Open 0.01 0.01 72.58 0.1 Open Open N/A N/A

B 240 Open Open 210 Open Open Open Open Open Open Open Open 800 N/A N/A C 0.068 Open Open 0.04 Open Open Open 0.045 Open Open Open Open Open N/A N/A A

RO 0 0 0.06 0.01 Open 0.01 Open 0.01 0.01 0.53 0.06 0.5 Open N/A N/A

B Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open 0.7 Open Open Open Open Open Open Open Open Open N/A N/A A

OSP

OR Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

B Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A A

RO Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

B Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

2nd round robin test

Table 9.5c: Corrosion map after the 3rd 5 days with KCl saturated salt solution. The relative humidity range across three test sites was 68-83%

Site Finish Flux Location on board (Grey is for no corrosion, yellow for corrosion; pink for light creep corrosion and dark red for

heavy creep corrosion. The resistance readings are in KΩ.) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

A

ImAg OR

29.4 Open Open 6.67 Open Open 7.21 0.886 3 Open 7.7 Open 122.3 N/A N/A B 1.42 8.4 0.3 1.27 230 2.6 0.05 86 66 630 6.3 Open 0.38 N/A N/A

B* Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open 0.695 Open Open 0.142 1.917 4320 Open Open Open 0.298 N/A N/A A

ImAg RO

1.262 Open Open 5.6 Open 55.9 21.6 180.3 1.042 45.0 Open Open 36.5 N/A N/A B 2.34 1.41 7.15 6.65 16.4 7.3 1.52 47.1 50 1150 42 Open 5.6 N/A N/A

B* Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open 0.615 Open Open 0.113 0.186 0.973 21.37 Open Open 0.146 N/A N/A A

ENIG OR 0.650 Open Open 44.59 Open Open 7.22 4.988 Open Open Open Open Open N/A N/A

B 0.53 Open Open Open Open Open 1900 Open Open 900 Open Open Open N/A N/A C 0.02 Open Open 0.083 Open Open Open Open Open Open Open Open Open N/A N/A A

ENIG RO 96.7 Open Open 326.9 Open Open Open Open Open Open Open Open Open N/A N/A

B 0.24 Open Open 5.9 Open Open 0.1 Open Open Open Open Open Open N/A N/A C 0.04 Open Open 0.147 Open Open Open Open Open Open Open Open Open N/A N/A A

OSP OR Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

B Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A A

OSP RO Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

B 15.4 Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A C Open Open Open Open Open Open Open Open Open Open Open Open Open N/A N/A

Worst case creep corrosion at Company A

1st round robin test run 2nd round robin test run

MgCl2 NH4NO3 KCl MgCl2 NaNO2 KCl

ImAg OR

ImAg RO

ENIG OR

ENIG RO

OSP OR

OSP RO

Effect of storage time

• Fresh PCBAs from two companies had suffered creep corrosion in the iNEMI FoS test.

• The PCBAs were from lots that had suffered creep corrosion in the field.

• After 2-year storage, the sister PCBAs from these lots no longer suffered creep corrosion in the iNEMI FoS test

Fresh Fresh2-year storage 2-year storage

Role of solder mask edge

• ImAg PCB soldered with NR330 flux from FOS run with 13-19% relative humidity.

• ENIG PCB soldered with EF8000 flux from FOS run with 74-80% relative humidity.

6th FOS test run: iNEMI creep corrosion examples

All the locations on the ENIG finished PCBs, soldered with organic acid and with rosin flux, tested at the 4 companies, suffered severe creep corrosion in 5 days of exposure in FOS chamber at 50oC with 40-ml Clorox in a 100-ml beaker at the 81% relative humidity afforded by saturated KCl solution.

L2 SMD L2 NSMD L8 SMD L8 NSMD

Control Mid-P, no flux,no treatment

Example of HCC

High-P, RO flux, PT3 treatment

Example of LCC

Mid-P, RO flux, PT3 treatment

iNEMI creep corrosion testing of various ENIG surface finish PCBs

Reproducing corrosion in lab

Creep corrosion in iNEMI FoSchamber at 50oC and 81%RH.

32

n Ag serpentine thin films, 770-nm thick, were evaporated on 15x15 mm SiO2 on Si dies.

n The Ag thin films were covered with the conformal coating under test.

n The corrosion rates of the conformally coated Ag thin films were measured using the electrical resistance approach.

n 4-point resistance measurements were made using wires connected to the thin film terminations with silver-laden epoxy.

Experimental procedure for characterizing conformal coatings

Fluoropolymer coatings (1-3 µm thick) in dry FOS chamber at 60oC

Bare Ag

Time, days0.0 0.2 0.4 0.6 0.8 1.0

Ag fi

lm th

ickn

ess,

nm

0

200

400

600

800

FP04

FP08

Fluoropolymer coating FP04 in dry FOS chamber at 60oC: 1X = 1-3 µm thick

hours0 20 40 60 80 100 120

Ag fi

lm th

ickne

ss, n

m

0

200

400

600

800

Bare Ag

1X

2X and 4X

Summary

• The iNEMI FoS corrosion chamber is inexpensive, easy to maintain and can be used for creep corrosion qualification testing of PCBAs.

• The temperature, relative humidity, sulfur and chlorine content in the chamber air can be reasonably well controlled.

• The test has been very successful in reproducing creep corrosion on printed circuit boards (PCBs) from lots that suffered creep corrosion in the field.

• It’s potential to be used as a general-purpose corrosion chamber for other applications/studies, e.g. characterization and ranking of conformal coatings, corrosion test for surface-mount multi-layer ceramic capacitors with soft terminations which may have corrosion-related failures in sulfur-bearing environments. Future work is considered to explore the usage of this FoS chamber for other purpose.

36

www.inemi.org

Q&A