White Residue Formation on Printed Circuit Board Assemblies Mulugeta Abtew Global Technology Solutions, Sanmina-SCI Corporation TM Chan Engineering Manager KS Moo Senior Process Engineer KH Teoh Process Engineer
Sanmina-SCI Penang, Malaysia
Abstract
Experimental and analytical approach was undertaken to investigate and correct white residue and corrosion failures found on multiple PCBAs. SEM and EDX evaluation of the failure sites revealed the presence of chloride, sodium and carbon. Ion Chromatography test of residue found on the failure sites also revealed the presence of 0.01 – 0.75 µg/in² chloride and 0.35 µg/in² bromide. The result of Fourier Transform Infrared Spectroscopy (FTIR) analysis identified the contaminant as flux residue coming from both aqueous and no-clean flux used in the assembly process. The field failure was successfully duplicated at the factory. Furthermore, response of both aqueous and no-clean flux to white residue formation was evaluated. Experimental results revealed that solder joints of active devices tend to produce white residue independent of ionic contamination level when both moisture and bias are applied to the site. The combination of experimental results and duplication of the
failure at the manufacturing site indicated that device powerup during PCB testing and inadequate moisture control were the primary causes of the failure. New testing and device power-up schemes were developed and implemented. Introduction
The phenomenon of white residue formation on printed circuit assemblies (PCA) after soldering is one of the most complex problems in electronic assembly(1). As shown in Figure 1, white residue is usually observed on PCA surfaces and on or around solder joints, and is encountered with a wide range of materials used in PCB assembly, including rosin fluxes, cleaning solvents, water-soluble fluxes, circuit board resins and epoxies, solder mask and component materials. The formation mechanism and the chemistry of white residue are very complex(2) with unpredictable outcomes. Consequently, a complete understanding of the reaction process is very difficult. The residue is known to
White Residue Formation on Printed Circuit Board Assemblies
occur suddenly with no known change of process parameters or material. Though the residue is commonly white, it also exhibits pale-yellow, blue or gray appearances.
Figure 1. White residue on failed PCA surface
The soldering and cleaning of electronic products involve a significant number of materials with varying chemical and physical properties. A typical PCA soldering process may involve many chemical species including acids (abietic, neoabietic, dehydroabietic, palustic, pimaric and isopimaric), esters (resin acid, fatty acid), diterpene, alcohol, hydrocarbons, wax, halides, epoxy, lead, tin, and many others. The byproducts of the reactions of these various chemicals subjected to soldering temperature are also as varied and complex. White residue is commonly associated with the presence of metallic salts of Pb, Sn and Cu forming a tenaciously adherent film, which is virtually insoluble in water, alcohol or in any known cleaning solvent used for PCA cleaning. The various metal salts that may appear after soldering and their solubility in aqueous solution(3) is listed in Table 1. Considering the chemical complexity of PCA soldering in general and white residue formation in particular, it is very difficult to isolate specific causes for white residue formation. However, based on experience and experimental studies, a number of possible causes for white residue formation reported in the literature(3) are summarized in Table 2.
Table 1. Solubility of post-soldering metal salts in aqueous solution
Compound
Solubility (g/100 cc)
Color
CuCl2
70.600
Green
CuCl
0.006
Green
CuBr2
Soluble
Black
SnCl2
Slightly Soluble
White
SnBr2
83.900
Pale Yellow
PbCl2
85.200
White
PbBr2
1.000
White
PbCO3
0.800
White
Copper Resinate
Insoluble
Green
Tin Resinate
Insoluble
Tan
Generally, white residue formation both in rosin-based noclean fluxes and water-soluble fluxes is often encountered in relatively high-temperature (> 215° C) soldering conditions of PCAs with large thermal mass where excess heat is likely to be generated. White residue formation is also known to occur after PCAs are cleaned in an aqueous cleaning system where the water temperature is relatively high (> 60° C). However, the presence or formation of white residue during PCA functional test and device power-up is not a common occurrence. In this report the results of experimental evaluations performed to investigate and characterize the formation of white residue observed with both no-clean and water-soluble fluxes during functional test and device power-up of PCAs is discussed.
2
White Residue Formation on Printed Circuit Board Assemblies
Table 2. Summary of possible causes for white residue formation
Possible Cause
Mechanisms
Oxidized Rosin Residue
Formation of insoluble white residue of peroxides and ketocompounds as a result of reaction with abietic acid (unsaturated double bond) in the presence of high heat.
Polymerized Rosin Residue
Polymerization of rosin due to exposure to high temperature and tin oxide serving as a catalyst, forming a tenaciously adherent white film insoluble in common solvents.
Hydrolyzed Rosin Residues
Hydrolization of the rosin acids due to absorbed moisture, creating an insoluble white residue after soldering.
Laminate, Flux and Solder Residue
Flux polymerization initiated possibly by the presence of excess epichlorohydrin in the laminate and tin oxide serving as a catalyst.
Solder and Rosin Flux Residue
Formation of insoluble metal salts (tin and lead abietates and pimarates) as a result of the reaction between solder or component terminations with carboxylic acid in the flux.
Solder and Halide Activators
Formations of white residue of metal halide salts of tin and lead chloride as a result of reaction between molten solder and halide activators in the flux.
Laminate and Flux Residue
When a laminate is under cured, brominated phenol in the curing reaction readily decomposes at soldering temperature (> 135° C), generating free bromide ions, which in turn react with Pb, forming a white residue of lead bromide.
Aqueous Cleaning Residue
White film residue tin oxide can be generated in aqueous cleaning systems where the alkaline saponifier concentration is too high (pH > 11).
PCA Failure Summary
A series of functional failures were encountered on multiple PCAs under service conditions. Preliminary investigation and inspection of the PCAs revealed white residue deposition around the failure sites. As shown in Figures 1 and 2, the white residue was found on the surface of the PCA and on and around solder joints. Furthermore, it was noted that the white residue was found around devices that get relatively hot during operation. The failed PCAs were soldered with a no-clean solder paste and flux. Scanning Electron Microscopy (SEM) images and Energy Dispersive X-ray (EDX) elemental spectrum of the white residue found at the failure sites are shown in Figures 3 and 4, respectively. In addition, Ion Chromatography (IC) results of the white residue found at the failure sites are also presented in Table 3.
Figure 2. White residue on device leads
When compared to a known good PCA, both the SEM and IC results of the failed PCAs indicated the presence of a high level of chlorine in the white residue with clear evidence of corrosion. Based on the EDX spectrum, it is highly likely that the composition of the white residue may be PbCl2 or PbCO3. Based on information from the suppliers, both the flux and solder paste used on the failed PCAs were halide free. Therefore, the source of the chlorine is not clear. However, it is highly likely that the chlorine might have come either from the Hot Air Solder Leveling (HASL) finish of the PCB substrate, from component terminations or from both. Comparative EDX results between a known good assembly and a failed assembly are listed in Table 4.
3
White Residue Formation on Printed Circuit Board Assemblies
Figure 3. SEM image of white residue
Table 3. EDX elemental results between failed and known good PCAs
Element
Failed PCA
Known Good PCA
Spot 1
Spot 2
Spot 3
Spot 1
Spot 2
C
16.12
42.40
21.65
46.73
71.09
O
5.44
25.22
17.90
13.68
18.09
Na
35.48
3.72
21.22
-
-
Cl
33.62
0.78
11.03
-
-
Al
-
1.74
-
0.36
0.24
Pb
9.34
8.28
8.50
11.89
-
Pb
-
-
-
-
-
Sn
-
13.75
16.74
10.14
0.94
Figure 4. EDX Spectrum of white residue
Background
During the PCA soldering process, the metals Sn, Pb and Cu experience oxidation, and their oxides react with the various acids from both the rosin and organic acid of the water-soluble flux, forming metallic salts. Lead carbonate and lead chloride are metallic salts commonly formed during PCBA soldering(4). These compounds are typically white in color. Lead carbonate, 2PbCO3Pb(OH)2 is a result of the reaction between PbO and carbonate. The carbonate is formed by the dissolution of CO2 from the atmosphere in moisture (H2O). Lead chloride is a result of the reaction between Pb and chlorine. Generally, both lead and tin carbonate and chloride can form. However, lead carbonate or chloride are more likely to be formed during soldering than tin carbonate or tin chloride because the solder surface is generally lead reach and the reaction process is thermodynamically(5) more favorable for lead reaction than tin. In the presence of moisture or relatively high humidity conditions (> 75% RH), a corrosion loop can be formed as a result of the reaction between PbCl2 and CO2 in the presence of moisture. The byproduct of the reactions between PbCl2, CO2 and H2O (moisture) are lead carbonate and hydrochloric. The acid then continues to react with metallic Pb in the solder to form more lead chloride(5).
4
White Residue Formation on Printed Circuit Board Assemblies
It has been shown that temperature and humidity are two critical parameters that control the reaction kinetics of corrosion failures in general and white residue formation in particular(6). The effect of temperature and humidity on failure mechanisms as a result of physiochemical reactions follows the Arrhenius Equation(5). [Equation 1]
where, kr = reaction rate ko = constant B = the activation energy, Ea, of the reaction, divided by the Gas Constant, K T = absolute temperature in °K The effect of relative humidity (RH) on the reaction can be expressed by an exponential equation [Sinnadurai]: [Eq. 2]
(
)
where
k o1 and C are constants The relationships expressed in Equations 1 and 2 show that reaction rates scale up with higher temperature and higher humidity conditions. The activation energy for white residue formation can also be determined experimentally from Equation 1. Assuming that the effects of temperature and humidity on the reaction rate are independent, the acceleration factor (AF) for temperature and humidity can be computed by combining Equation 1 and Equation 2. [Eq.3]
⎡ ⎛ 1 1 AF = exp ⎢ B⎜ − ⎢⎣ ⎜⎝ Tlife Ttest
⎞ 2 2 ⎟ + C RH test − RH life ⎟ ⎠
(
Expected Life Time = AF × Test time to fail The value of the constants B and C are experimentally determined and are case specific. However, for general estimations, the following values can be used as suggested by Sinnadurai(5): B ≈ 7000K, equivalent to 0.6 eV C ≈ 4.4 × 10-4
⎛−B⎞ k r = k o exp⎜ ⎟ ⎝ T ⎠
k H = k o1 exp C.RH 2
[Eq. 4]
⎤
)⎥ ⎥⎦
The subscript “test” refers to the laboratory test conditions and “life” the actual service conditions. Expected service life is estimated or predicted by multiplying the laboratory test time to fail by the acceleration factor as shown in Equation 4.
Using Equation 4, one can predict the service life of a PCA under defined temperature and humidity conditions in relation to corrosion and white residue formation. Experimental Approach
Based on evaluation of the failed assemblies, experimental approach was used to simulate the field conditions and to duplicate the failure. The objective of the experiment was to identify the parameters that caused or influenced the formation of the white residue on the failed assemblies. Four PCAs were assembled and soldered following the normal PCB assembly process, namely solder paste application, component placement, reflow soldering, cleaning1, hand load, wave soldering, cleaning and in-circuit test (ICT). Two of the PCAs were built with a no-clean chemistry while the other two were built using aqueousbased chemistry. Once the PCAs passed ICT, the assemblies were thoroughly inspected under a 30× microscope for any trace of white residue. Then, two assemblies, one from the no-clean and one from the aqueous batch, were randomly selected for further investigation after it was verified that the assemblies were free of any trace of white residue. The two assemblies were then functionally tested under normal environmental conditions where the humidity level was maintained at 55% RH. Power was applied on the PCA during functional test and the locations on the boards where white residue had formed on the previously failed PCAs were closely monitored for white residue formation. The remaining two PCAs were also functionally tested with full power on. However, in this case the humidity level was uncontrolled and drops of de-ionized water were intentionally added around the vicinity of the powered-on device leads. Progress was monitored on predefined locations (where white reside had been observed on previously failed PCAs) via visual inspection under a 20× microscope. 1
Applicable only to PCAs built with water-soluble chemistry
5
White Residue Formation on Printed Circuit Board Assemblies
To measure the level of contamination on both incoming PCBs and components, a random sample of PCBs and components were selected from stock and Ion Chromatography evaluation was performed following IPC-TM-650 test methods.
Table 5: Ion Chromatography result for bare PCB samples
Sample
Ref Area No
Ionic Contamination Chloride (Cl)
Sulphate (SO4)
Concentration (µg/in²)
Results and Analysis
Ion Chromatography results for incoming PCB and components are shown in Tables 4 and 5, respectively. The level of ionic contamination found on both incoming PCBs and components appears to be significantly below the maximum allowable ionic contamination levels for PCAs as shown in Table 4. The IC results generally show only the average level of contamination without any indication about the distribution of the contamination on the PCA. This means it is possible for a much higher level (> 4 µg/in²) of ionic contamination to be present on the PCAs concentrated on a specific location than what is shown in Tables 5 and 6. A high level of ionic contamination concentrated in a few locations may have caused failure in combination with applicable temperature and humidity conditions. It is not known what level of ionic contamination can actually induce corrosion failures or initiate white residue formation. However, one study has shown that as low as 4 µg/in² of chloride ion concentration will cause corrosion failure at 60° C temperature and 80% RH(6). Marginal Risk
1
2
3
1
0.3673
<0.0253
0.1888
2
0.1348
<0.0122
0.0920
3
0.1087
<0.0122
0.1207
1
0.0992
<0.0051
0.0154
2
0.1123
<0.0051
0.0297
3
0.0590
<0.0051
0.0088
1
0.1025
<0.0170
0.0614
2
0.0451
<0.0122
0.0166
3
0.0167
<0.0051
<0.0051
Table 6: Ion Chromatography result for component samples
Sample
Ionic Contamination Chloride (Cl)
Bromide (Br)
Sulphate (SO4)
Concentration (µg/in²)
Table 4. Allowable ionic contamination level for PCAs2
2
Bromide (Br)
Flux Type
Low Risk
High Risk
RMA
<8.0 µg/in²
8.0 – 10.0 µg/in²
>10 µg/in²
Aqueous
<6.0 µg/in²
6.0 – 8.0 µg/in²
>8 µg/in²
No-clean
<3.0 µg/in²
3.0 – 4.5 µg/in²
>4.5 µg/in²
4
0.7346
<0.0263
6.6983
5
0.1105
<0.0430
0.0301
6
0.0167
<0.3584
<0.3584
The ionic contamination levels are determined by IC test.
6
White Residue Formation on Printed Circuit Board Assemblies
Figure 5. White residue formed 3 minutes after power-on
Figure 6. White residue formed 5 minutes after power-on
No white residue formation was observed on the first batch of PCAs functionally tested under controlled humidity environment. However, formation of white residue was observed on the batch of PCAs functionally tested under uncontrolled humidity environment. The white residue on the failed PCAs started to form at two locations on affected devices two minutes after power-on. Then, more residue started to form and spread over the PCA surfaces. Figures 5 and 6 show the extent of white residue formed after 3 and 5 minutes of power-on, respectively.
Conclusion
The white residue initially occurred only on powered-on active devices, later spreading to passive components. But it was also noted that white residue did not form on all powered-on devices. Those powered-on devices that did not have white residue forming on or around their leads appear to be devices that had low power input (< 1.5 W). It was not known what temperature the components or component leads have to reach after power-on for the white residue to start forming. However, it was clear that white residue formed only on devices that were powered on and have relatively higher input power (> 3.0 W). This implies that the extent of input power to a device and the resulting rise in temperature during operation not only affect device performance but may also create favorable conditions for corrosion and white residue formation. It should also be noted that with increasing interconnect density, this may be a critical problem that warrants design considerations in the area of power management and device temperature control.
Experimental investigation was performed to identify the cause of field failure on multiple PCAs. White residue and corrosion spots were identified at failure sites. Ion Chromatography and SEM/EDX analysis of the residue on the failure site indicated the presence of a high level of chloride. The failure was successfully duplicated at the factory. White residue formation was encountered during device power-on under high-humidity conditions (> 75% RH). Experimental results also indicated a correlation between input power to a device and white residue formation, provided high-humidity environment. Specifically, white residue formation was observed on device locations that had an input power greater than 3.0 W. No actual temperature measurements of devices or device leads were conducted while the devices were in operating conditions. But, it is suggested that the rise in temperature of the device body during operation as a result of relatively high input power may be contributing to corrosion and white residue formation given a humid environment. New testing and device power-up schemes were developed and implemented to alleviate the problem. It was also noted that in addition to moisture and contamination control during assembly, white residue formation is related to the design and testing of PCAs in the area of power management and device operating temperature. Further studies are suggested to correlate input power and device temperature rise and to identify the device body temperature regions that trigger corrosion and white residue formation in high-humidity environments.
7
References
1. Zou, Y. Z., “White Residue on Soldered Printed Circuit Assemblies,” National Physical Laboratory, 1988. 2. Tautscher, C.J., “The Contamination of Printed Wiring Boards and Assemblies,” Omega Scientific Services, Bothell, WA (1976). 3. Bernier, F.D., “The Nature of White Residue on Printed Circuit Assemblies,” Technical Report, Kester Solder, 1988.
4. Ellis, B.N., Cleaning and Contamination of Electronic Components and Assemblies, Electrochemical Publications Limited, (1986). 5. Lea, C., After CFCs?, Options for Cleaning Electronics Assemblies, Electrochemical Publications Limited, (1992). 6. “Corrosion of Indium Based Solders,” Amdhal Technical Bulletin, 1990.
