Reflow solder flux residue and humidity interaction: investigation using real PCBA component designs | Scientific Reports

Jul 02, 2025

Scientific Reports volume 15, Article number: 22496 (2025) Cite this article

This paper investigates the climatic reliability of realistic PCBA component geometries soldered using reflow process, from the perspective of the interaction between component design parameters, reflow residue and humidity. The study was carried out using electrochemical characterization, electron microscopy, and X-ray imaging. Custom-designed PCBA test cards with dummy components have been exposed to harsh climate after reflow soldering. A methodology of combining Leakage current and electrochemical impedance measurements has been used to characterize PCBA performance. Changes in flux morphology and any associated negative effects as a function of humidity interaction have been connected to component connector design parameters, to achieve a holistic understanding of flux system-component suitability and humidity interaction.

Electronics and electronic devices are fast attaining omnipresence and near-complete integration in all sectors of human life. Daily life is incomplete and almost unimaginable without interaction with some kind of high-power device like communication and connectivity infrastructure (cell-phone towers), vehicular transport (electric vehicles and trains) and energy systems which run on very high voltages (> 300 V). These devices run on increasingly complex, dense and functionally powerful electronic systems, with an expectation of uninterrupted and consistent performance worldwide despite climatic variation. Increasing operational voltages create reliability risks due to higher electric field (E = V/d). High-humidity conditions can introduce larger amounts of moisture in the system, which create electrolytic cells between oppositely biased components. Conductive and hygroscopic ionic residues on the PCBA surface from the manufacturing process can accelerate formation of this water layer and the water layer conductivity, which further increases the reliability risk when combined with the high electric field. Formation of water film between pads can generate leak current affecting functionality of the device. Metal dissolution due to leak current and associated local conditions of pH can further result in the formation of electrochemical migration (ECM), characterized by formation of metallic dendrites which create short circuits. High current leakage ensues, which can cause intermittent and eventually complete operational failure of the device1,2.

The high computing capability and packing density of modern electronics is enabled by great inroads into design and packaging of the components that are attached to the PCBAs. Mobile phones and computers, along with other electronic control systems have, over the past few years, mainly used three groups of PCBA-level packages- Area array packages (Ball Grid Array/BGA type), Leadless packages like Quad-Flat No Lead (QFN), Quad-Flat Packages (QFP) used to package microprocessors in larger devices, and Wafer-level Packages (WLP)3. The state of the art now also includes System In package (SiP) components, which are often connected to the PCBAs through different layouts of BGA connectors, as is in the case of Embedded Wafer Level Ball Grid Array (eWLB), Flip-Chip packages and Wafer-Level Chip Scale Packages (WLCSP)4,5. These electronic components are assembled using reflow soldering, which relies on a solder paste consisting of solder powder, fluxing agent (a combination of resin/rosin base and weak organic acids for oxide removal from connection surfaces) and other chemicals to enable better flow and delivery of paste to the PCB surface through stencil-printing. Wave soldering is another option, but SMT using reflow is now the preferred method due to an overall cleaner process (wave solder flux is sprayed all over the board, as opposed to localized flux delivery) owing to the low standoff height (elevation of the package base from the board surface) of modern electronic components which can often range from 200 to 300 μm for larger BGAs6, to as low as 50 μm for some QFNs6,7. The market trend is to use ‘No-clean’ fluxes that require no post-manufacturing cleaning step to save time and cost, as the acidic and hygroscopic fluxing compounds either escape the system through outgassing (due to volatility) during the soldering process or are encapsulated in the resin/rosin base. These are different from water-soluble fluxes which are more aggressive, and require cleaning after soldering. However, a side-effect of the low standoff height is that the outgassing process is often left incomplete due to obstruction of gas flow by both the low height and the connector design and layout (size, distribution, presence of additional heat-sink and grounding lugs, etc.). This leaves a higher quantity of flux residue contained in the encapsulant under the component8, which has been observed to spread, change morphology and release hygroscopic activator acids that eventually create highly saturated electrolyte thin films under the component, and bring forth all the associated reliability concerns even at low relative humidity conditions9,10,11.

The current state of research on the interaction of electronic systems and humidity includes more fundamental work on the formation of water film on a PCBA surface12, the behaviour of different flux activator acids13,14 and amines in humid conditions15,16, corrosion studies of various solder alloy compositions9,17,18, and reliability studies of wave solder flux systems19. However, these studies have largely been carried out on standard interdigitated electrode test PCBs, which are “open” systems as they do not include a component. Other studies dealt with the reliability of solder flux systems, which involved the use of lab-scale test PCBA designs with a significantly larger pitch distance and standoff height than the industrial norm10,11. Studies of solder reliability in more realistic component-based systems (BGA, QFN) exist, but the focus has been on physical failure methods like effect of moulded external encapsulation20 and joint quality, made up of factors such as void formation, fracture mechanics21,22,23, thermal shock resistance24,25, vibrational ageing26, and the like. Another type of study has been done by Bixenman and McMeen, which looks at the effect of component standoff on measured SIR values8. There has not been systematic research into the combined interactions of different flux compounds, different climatic conditions and electronic components with more realistic manufacturing and design conditions, as found in the field. Such a work would be a combination of many of the research themes that constitute the current state of the art.

This study included test cards consisting of a total of 5 PCBA components [FBGA (with and without thermal lug), QFP, QFN, Capacitors) commonly found in devices. Three sets of test PCBAs were fabricated using 3 different flux systems to investigate combined effect PCBA design, flux chemistry, and reflow process on humidity robustness. Exposure to a high humidity environment with varying temperature and electrical measurements (Direct Current and Alternating Current-based (Electrochemical impedance)) have been carried out on variations of these components, which include 3 different flux systems (in order to study the effect of different flux chemistries and activator chemistry in flux systems), both in the presence and absence of a dummy components. PCBA with dummy components provide way of understanding flux residue formation below components depending on stand-off height and size. In addition to the humidity testing, tested PCBAs were analysed using SEM and X-ray imaging to study the changes in flux morphology after testing, and also identify any component design- specific abnormalities that might arise as a result of flux choice.

All experiments were conducted on custom fabricated test PCBA cards which had the possibility for electrical connection to the pads (Fig. 1). The body of the PCB is made of FR-4 material (Isola 370 h laminate with Taiyo PSR4000BN solder mask), and the connectors were made of Cu with no added surface finish. As the Fig. 1 shows, the test cards are intended to simulate the types of electronic components commonly used in electronic systems. The pads are connected in an interdigitated manner much like those found on a IPC B-24 test board, to enable electrochemical monitoring of the system under applied bias. These also serve as the solder points for a simulated dummy “component”.

Test Vehicle used, showing the process of reflow soldering and component surface mounting. (a)Test card with bare Cu connectors (b)Reflow-soldered test card, without component, to directly study changes in flux morphology under humid conditions (c) Test card with reflow-soldered dummy components.

Dummy components used in this case is made of a smaller piece of FR-4 material with the same solder mask and surface finish as the base PCB, on an identical pattern of non-electrically connected pads on face down part of the component. Test cards representing a total of 5 different types of components chosen for specific reasons, as shown in Table 1.

An application-based choice of solder flux systems (Table 2) was made for this study, in line with the “functional” point-of-view of the study. The frame of reference was solder alloy SAC305 (96.5% Sn, 3.0% Ag, 0.5% Cu), which made it that the flux system was the only differentiating factor, a choice which is often made in industry, dependent on the application. The words “solder paste”, “flux system” and “reflow solder flux” are henceforth used interchangeably in this study. These were all of the “ROL0” type, where “RO” refers to a Rosin base, and “L0” refers to a negligible Halide content in the fluxing compound. The three flux systems contained Succinate and Glutarate activator compounds were identified using ion chromatography. SP1, a Type-4 solder paste, has good spreadability and finished joint quality, which makes it common industrial choice for a variety of applications. SP2, a Type-4 solder paste is a known harsh-condition flux system choice for large-pitch applications, and SP3, also a Type-4 solder paste, with its low spreadability and clean joint formation, was the fine-pitch flux system representative. The solder pastes were printed on the base PCB using a Stentech 5 mil (127 μm) SLIK stencil (laser-cut) attached to a Speedline MPM Momentum Printer, after which a JUKI FX3-KE-2080 L pick-and place equipment was used to mount the dummy components. Reflow soldering was carried out using the reflow profile attached to Table 2, in a Heller 1913 MK III reflow oven (Air). This profile was made as an average of the OEM-specified reflow profiles (all of which were similar in structure and temperature levels) for the three solder pastes, in order to have a basis of comparison between flux residue quantity and behaviour.

The test PCBAs were placed in an ESPEC PL-3KPH climatic chamber (tolerances for temperature and relative humidity are: ±0.3 °C/±2.5%RH in the range − 40 °C to + 100 °C/20%RH to 98% RH) for testing- The climate profile in Fig. 2 was chosen to achieve aggressive moisture loading and unloading, along with a soak function. The combination of rapid temperature variation and high humidity stresses the system, as the increasing air moisture content with temperature from at 25 °C/95%RH to 40 °C/95%RH and finally to 60 °C/95%RH creates a lag between PCB temperature and chamber air temperature, which forces a large amount of moisture from the surrounding air to condense on the PCB surface, thus creating a sufficiently high-water film thickness to cause reliability issues such as leakage currents and electrochemical migration. The soak time is to intensify the effect of the increased water film thickness, and the profile cycles back down to 25 °C/95%RH to unload moisture from the surface, and study if the “shock and soak” phase brings about any changes in the system behaviour in terms of flux-humidity interaction. This was facilitated by electrical measurements.

Temperature Profile used for testing.

A Biologic VSP Potentiostat was used to carry out electrical measurements on the test PCBAs during climate exposure. A combination of Electrochemical Impedance Spectroscopy (EIS) and Chrono-Amperometry (CA) was selected to be able to simultaneously track changes in the system behaviour (resistive and capacitive behaviour) during climate exposure, and the more operationally critical leakage current measurement (Fig. 3). This method was based on an implementation in the work of Lauser et. al27. A series of trials (Table 3) was conducted to determine the appropriate parameters of these electrical measurement techniques, in order to reduce noise during EIS measurement and avoid any interference with the leakage current measurement. The expectation was to obtain a mirror-image effect between the measured leakage current values and impedance values.

Alternating AC-DC method used for electrical characterization of the PCBAs in the study. Figure adapted from27.

The first measurement regime attempted was a coupled AC-DC measurement, which was intended to be a fully combined version of the usual discrete measurement methods used more traditionally, which have included an EIS measurement for a few days, followed by a DC leakage current measurement for a few days. The coupled method involved a continuous 10 V DC signal to have a running leakage current measurement, over which a 10mV AC signal was applied intermittently to measure impedance. This proved to be impractical in praxis, as the dynamic nature of water film formation and flux-water film interaction, combined with the inherent sensitivity of low-amplitude EIS, made for a high Signal-to-Noise-Ratio (SNR) during data collection. The next step was to increase the AC amplitude, to reduce sensitivity and noise. This improved the SNR, but impedance response was not as pronounced. The coupled method was thus rejected. The next methodology was an alternating AC-DC measurement as opposed to directly coupled, to avoid the noise observed at lower frequencies. As Fig. 3 shows, the regime involved a ca. 3 min EIS measurement, followed by a 3 min leakage current measurement to provide the closest possible correlation between impedance and current values. Similar to the coupled method, 10 mV AC amplitude was tried as the first parameter and rejected due to unsatisfactory impedance response with climatic variation. The selected parameters were an alternating 500 mV AC (100mHz–100 kHz sweep) and 10 V applied DC, with a 1 min settling time between loops. This regime gave a low SNR, and good impedance-leakage current correlation.

Analysis of the flux systems was also carried out, to have a better understanding of the humidity behaviour. The testing rubric followed other studies on flux-humidity interaction10,11,28, which have determined that the resin/rosin base of the flux, which acts as an encapsulant for ionic fluxing agent residues (Weak Organic Acids- WOAs), opens up with sustained humidity interaction, thus releasing the ionic residues on to the PCBA surface and exacerbating corrosion issues. Solder paste was printed and reflowed on separate test PCBs with higher pitch distance and square pads (chosen for better imageability and extraction convenience, which was also used in a lab-scale study of reflow flux-humidity interaction11), which were then subjected to humidity exposure. SEM imaging was carried out before and after humidity exposure, along with a local water-based extraction from the PCB surface (using 200 µL of ultra-pure Millipore water (electrical conductivity of ca. 4 µS/cm), allowing 120 s of dissolution time before extraction), followed by identification and quantification of ionic content in an Ion Chromatograph).

Optical inspection of test PCBs was carried out on a Keyence VHX-6000 series Digital Microscope, to visually investigate morphological changes to the residues during the course of testing. The Digital microscope 1/1.8” sensor has a Virtual Pixel resolution of 1600*1200, and a magnification range of 20X-200X. Complementary component-level flux morphology was imaged using a FEI Quanta AFEG 250 SEM, with an acceleration voltage range of 200 V-30 kV, Maximum Horizontal Field Width of 5 mm at 10 mm analytical working distance, 200nA maximum probe current, and an obtainable magnification range of 14x-1000000x. An Everhardt-Thornley SED (ETD- Secondary electron) with rated resolution of 1 nm at 30 kV in high vacuum was used as primary imaging detector.

X-Ray imaging was carried out on the Zeiss XRADIA 520 Versa, which has a spatial resolution of 0.7 μm achieved from a sealed-transmission source with voltage range 30–160 kV, maximum power output of 10 W, and a detector with variable objectives of 0.4X, 4X and 20X. The samples were cut to a size close to the component footprint. Zeiss Scout + Scan software was used to program a source voltage of 80 kV, with source power of 7 W was chosen for imaging the samples. Pixels with size of 13.5 μm were binned to a factor of 2 to achieve optimum contrast and imaging time. Further contrast and gamma adjustments were made to improve image quality, along with the practice of reducing source distance and increasing detector distance to take clear, magnified X-ray images at 4X objective, as 20X objective produced images with low quality.

The Fig. 4 shows the variation of measured electrical impedance and leakage current during the climate exposure. The leakage current, in all cases follows the climatic profile, and shows a step-wise increase in value when the temperature step increases (from 25 to 40 °C-60 °C at the time points 24 h, 48 h and 36 h respectively- refer the climate profile in Fig. 2) although there are variations in the degree of increase between different component-solder flux combinations An additional trend is seen in the case of measured impedance, which in all cases shows a ‘mirror’ effect of variation i.e. decrease of impedance, compared to the leakage current.

Typical indicative plots showing the electrical measurement (Impedance and Leakage current vs. time) made during climate exposure. (a) SP1-B52 Board combination; (b) SP2-B52 Board combination; (c) SP3-B52 Board combination. Impedance has been plotted at 148 mHz, thus representing the resistive regime of the electrochemical system, which should make for the closest comparison to leakage current (inversely proportional).The blue, green and red arrows point to the “peak” leakage current values, which have been considered as the basis to compare operational reliability in an upcoming section.

The overall variation/sensitivity of the flux systems to changing temperature varied; variation of leakage current and impedance with temperature were most pronounced in the case of SP1 for all components. SP2 showed the best electrical response sensitivity to temperature in the case of the FBGA244 and QFP80 components; SP3 had the least temperature-based variation in electrical response for all components.

Ion Chromatography showed the content of the three flux systems (Table 4). SP1 and SP2 contained a combination of Succinic and Glutaric acids (shown by the presence of succinate and glutarate ions), and SP3 contained only Succinic acid. SP1 had the overall highest amount of acid extracted, followed by SP2 and SP3 respectively. However, the main differentiating factor is the higher amount of glutaric acid, of which SP1 has almost 4 times of SP2.

Figure 5 shows a broad overview of the change in flux residue morphology after exposure to humidity, which shows different degrees of spreading of flux residue (marked by the yellow arrows) of the three fluxes. SP1 showed the highest spreading of flux residue, followed by SP2 and SP3- these are in keeping with the expected behaviour of these flux systems, and explain the trends in electrical response to changes in temperature seen in Fig. 4. Figure 6 shows SEM micrographs of flux residue changing after exposure to the climatic test. Here, SP1 shows the highest level of “opening up” of flux residue (holes as highlighted using yellow circle), followed by SP2 (cracks shown inside green circle) and finally SP3 (tiny holes and larger closed deformations marked in blue circle). Figure 5 shows highest spreading of SP1 flux residue to bridge oppositely located and biased connectors, followed by SP2 and SP3 respectively.

Optical images of spreading of SP1, SP2 and SP3 flux residues after reflow soldering and humid exposure, as carried out on test PCBs similar to the ones used in11. The yellow-marked area in the reference image shows the area where images (a), (b) and (c) were taken.

SEM Micrographs of flux residue before and after humid exposure. These micrographs are taken from the same spots as marked by the yellow arrows in Fig. 5. The yellow, green and blue markings respectively show the large “opening up” of SP1 residue encapsulant, slightly smaller change in SP2, and no “opening up” of SP3 residue encapsulant after humid exposure.

Figure 7 shows peak ILC related to individual components derived from humidity test results in Fig. 4. The plot in Fig. 7 shows a few interesting trends in the performance of the various surface mount components showing combined effect of stand-off height, component size, flux used, and reflow process. The “Peak ILC” Y-axis values refer to the highest current measured across the entire 120 h test cycle (marked by the blue, green and red arrows in Fig. 4), with an intention of identifying any current levels in the vicinity of, or as close as possible to 0.01 µA, which is a current level above which dendrites due to electro-chemical migration have been observed in other studies30,31).This current level implies that the water layer has a sufficient thickness and electrical conductivity (due to dissolved metal ions from the anode, and dissolved ions from surface contamination) to allow transport of the metal ions from anode to cathode under electric field, and redeposition at the cathode to form dendrites. This be taken as a critical point beyond which the leakage currents can be expected to increase greatly both due to formation of a conductive metallic dendrite and the conductive water layer itself.

Plot showing the variation of measured peak leakage currents (across the entire 120 h measurement period) for the different Component-Solder flux combinations. Pitch distance in µm. Values marked by arrows in Fig. 4 have been chosen for this comparison.

Overall trends seen in Fig. 7 include the difference in measured leakage current for the CAPS, QFP80, QFN48 and FBGA244-GND with different solder pastes. SP1 showed the highest leakage current in case of the Capacitors, QFP80, FBGA244-GND and QFN48 components, followed by SP2 and SP3. A salient trend of connector design was the effect of the centre lug in the FBGA244-GND component. This component has a similar ball count to FBGA256 in a footprint size similar to FBGA572, but produced high leakage current values in the range of 1 µA. Component pad pitch distance also affects leakage current, as visible by the QFN showing higher leakage current than the FBGA 244-GND and QFP80 components in case of SP2 and SP3.An outlier.

Indicative SEM micrographs showing spreading of flux residue on FBGA244 and QFN component patterns after humidity exposure; (a1-a4)-SP1, (b1-b4)-SP2, (c1-c4)-SP3 The SEM micrographs have been taken on unpopulated boards which have been reflow-soldered with the same flux systems, for the sole purpose of illustration as removal of components without damaging the residues underneath proved difficult. The presence of components will only exacerbate the issue of spreading of flux, as has been shown in other work10,11,28. The red square-marked areas indicate the regions where the SEM micrographs were obtained after testing.

The effect of flux spreading is illustrated by Fig. 8, where the movement of flux residue from one region can be directly correlated to the trend in leakage current performance measured during testing. Moving from one BGA ball to another in a1 and a2, and from the centre lug to neighbouring connectors for the QFN in a3 and a4, SP1 shows the highest spreading of thick flux residue to cover different connector pads, characterized by the yellow-marked outcropping in Fig. 8 (a1), and in Fig. 8 (b1). The amount of flux residue is also an important consideration, and SP1 visually shows the highest amount, followed by SP2, and SP3 showing a significantly smaller amount of flux residue. This fits with the observational trend in residue spreading on the test PCBs in Fig. 5.

X-ray imaging was done to search for any abnormalities in solder joint formation under the components. Figure 9 shows the solder joint morphologies under the different components. As indicated by the yellow, green and blue arrows in Fig. 9a1, a2, a3 respectively, the morphology of the BGA572 solder joints is visible. The appearance is fainter than other imaging methods, as the metal powder in the flux residue absorbs lesser amount of x-rays due to being mainly polymeric in nature. Only areas with thicker accumulation of uncured solder flux or residue (containing metal powder) would show on the image, and the darkest features in the image will represent the metallic parts (solder joint in this case) due to the highest X-ray attenuation (effect of radiographic contrast32,33).

Typical X-Ray images, indicative of the joint and flux residue morphology under the different tested dummy components. These X-ray images are only intended to illustrate differences between the components and flux systems, and correlate to the SEM micrographs in Fig. 8.

Typical differences in solder joint morphology before and after testing of the FBGA244-GND component, showing the extension/spread of some kind of substance (marked in yellow, possibly semi-cured solder paste)from the GND/center lug.

Figure 10 shows the state of the GND/centre lug soldered with SP1, before and after testing. This is an exploded view of Fig. 9 (a1), and shows the spread of some faintly-visible residue outwards from the GND/Centre lug after testing.

Typical differences in solder joint morphology between the centre and corner of the BGA572 component, showing the effect of component warpage during reflow.

The X-ray imaging of the samples brought about an interesting feature of BGA connector design highlighted by the difference in the size and general morphology of solder joints in the vicinity of the corners of the BGA component. As Fig. 11 shows, the solder ball joints in the centre of the component have a more uniform, circular shape that fills the area of the connector. However, on moving to the corner, the solder joints appear irregular, with some unevenness and graining that are visible in the red-marked figure inset of Fig. 10. Upon closer inspection of the Fig. 9a, b, c, differences in the corner solder joint morphology can be seen, with SP3 showing the “smallest” size of circle, with an annular ring visible that is the image of the copper PCB connector pad- SP2 shows a slightly better-filled but still irregular solder joint morphology SP1 shows well-formed and evenly-formed circular solder joints.

As seen in Fig. 4, the use of impedance allows for tracking of changes in residue characteristics under the components, and the use of the low-frequency resistive regime can track changes in ionic resistance of the water layer under components (released from flux residue). From an electrical perspective as well, this fulfills the intention of the combined method of measurement, as it provides a good simultaneous tracking of the leakage current variation with the variation in resistive impedance (reduction in resistive regime of impedance translates to increase in leakage current, and vice-versa).

The variation in leakage current levels of capacitors, and the other tested components as well with the cyclic nature of temperature range, is due to a lag in PCBA surface temperature variation as compared to air (due to lower thermal conductivity of the polymer surfaces). This leads to condensation and water layer formation34,35, which eventually allows access of moisture to regions under electronic components that contain flux residues. Extended exposure to moisture leads to a reduction in Tg (Glass transition temperature) of the polymer encapsulant component in solder flux due to hydrolysis36. Tg reduction causes softening and “opening up” of the encapsulation to release ionic, hygroscopic flux residues like Weak Organic Acids (WOAs)10,11, which increases the conductivity of water film12and subsequent leakage currents10,11,28,35,37,38,39,40.

As shown in Table 4, SP1, SP2 and SP3 contained varying concentrations of the WOAs Succinic acid and Glutaric acid as fluxing agents. As Table 5 shows, glutaric acid takes up significantly more moisture than succinic acid13,29,41,42 as characterized by the significantly higher solubility, lower deliquescence humidity and saturated moisture sorption level of glutaric acid. The higher solubility also points to a more acidic nature than succinic acid, as indicated by the lower pKa1 value (3.7643 vs. 4.2144 for succinic acid- pKa1 considered from the Marvin method45 as the carboxylic acid group is of interest here). This points to the phenomenon where flux systems containing more glutaric acid will form a thicker water film of higher conductivity (high solubility, high dissociation to release H+ ions) than succinic acid at any given time point for ambient conditions above the dRH point of 25 °C/84.7%RH29,46.

Spreading of the flux residue increases the propensity for quick failures, as the residue can easily bridge to oppositely-biased connectors in the vicinity, followed by opening up to release hygroscopic ionic residues on to the PCBA surface that form a thicker, more conductive water layer and associated failure mechanisms like leakage currents and electrochemical migration10,11,28,37. The highest spreading and opening up of SP1 (Fig. 5 (a) and Fig. 6 (a2) respectively), combined with highest overall acid content and concentration of glutaric acid in the flux formulation, should point to SP1 having the worst electrical performance (High ILC) for most component types, followed by SP2 and SP3.

As shown in Fig. 7, the high ILC values for Capacitor components can be attributed to a combination of low (lowest of all components) standoff height, small size (enabling quick interconnection of the terminals by condensing water) and mounting in series. The water layer resistance across all components of varying sizes in series would be added up, resulting in an interconnection due to water layer on a single component from the series reflecting as an increased leakage current across the entire series, since the net resistance would reduce. The capacitors are thus primarily used to compare flux performance, and further analysis was carried out on the 4 other components owing to their increased design complexity. Looking at the effects of component design as well, the capacitors soldered with SP1 having the highest leakage current of all flux-component combinations is thus in agreement with performance expectations.

The higher ILC of the FBGA244-GND component compared to the FBGA572 in case of SP1 and SP2 can be attributed a dual effect of the GND/centre lug part of the component, which necessitates a higher volume of solder paste (as the square lug is filled in by division into 4 smaller squares) while also causing a structural flow restriction to outgassing volatiles from under the component during reflow. The trend of flux spread for SP1 and SP2 in the close-up SEM micrographs (Fig. 8) (and Fig. 5 as well) clearly illustrates this increased flux quantity, and the higher ILC value of SP1-soldered FBGA244-GND shows an electrochemical incompatibility between this type of component layout and flux system. The low ILC value for SP3 shows that a low-spread, low-residue flux system is more electrochemically suited to this component layout.

As seen in Fig. 7, QFN components also show high ILC values. The QFN component has among the lowest average standoff height of all components by design (Table 1), which increases the amount of entrapped residue under the component due to lack of sufficient flux outgassing during reflow soldering. This agrees with work done in two studies; one by Bixenman et al. which shows reduction in Surface Insulation Resistance measured on electronic components like QFP, QFN and capacitors, with reducing standoff height8, and the other by Li. et al., which showed a decrease of measured impedance value with reduction in standoff height of a reflow-soldered dummy component28. This effect of pitch distance on electrochemical reliability was shown by the work of Bahrebar et al. where an increasing leakage current and reduced time to failure was observed with reducing pitch distance30,31In this case as well, the high ILC of SP1 shows its incompatibility with this component type.

The lower ILC values shown by QFP80 component for SP2 and SP3 can be attributed to the higher pitch distance and standoff height (Table 1). SP1 showing a high ILC value can be attributed to the high spread of the flux residue, which can easily bridge the larger pitch distance. The location of connectors at the edge of the component also helps in flux outgassing during reflow soldering, which when combined with the higher standoff height, will reduce trapping of residues under the component and yield better overall reliability.

An outlier from the discussed trend was the FBGA572 component, where SP3 showed the highest ILC values. X ray imaging (Figs. 9, 10 and 11) revealed inconsistencies in the joint formation, which can be attributed to warpage or tilting of the plastic BGA body due to thermal gradients between the outer edges and centre of the component during the reflow process. Warpage would lead to increase or decrease of standoff height at the edges of the component. Figure 10 showed differences in joint appearance between edges and the centre of the component. A properly formed solder joint would be mostly circular in shape, and display a darker colour in the X-ray image due to the presence of solidified solder alloy- this was visible in the centre of the component. Smaller and irregular solder joints would indicate a lifting of the components above the expected standoff height towards the corners. This would force the solder paste melt to accommodate in the z-direction due to surface tension in the liquid state. In case of lower spreadability solder flux systems like SP2 and SP3, the low wetting cannot accommodate for warpage, giving a smaller x-y footprint in the X-ray image. Extending this, SP1 with the highest spreadability will easily accommodate warpage. From an electrochemical reliability logical perspective, higher standoff height due to warpage would allow for increased ingress of moisture under the component, which can potentially activate even a relatively smaller amount of flux residue that would be left behind due to somewhat better outgassing that the increased gap would inadvertently allow. If ideal solder coverage is maintained, larger components would perform better as ambient moisture would not achieve sufficient ingress to open up the polymer matrix of the entirety of the solidified flux residue, thus allowing it to keep the ionic flux residues contained10. This could explain the reversed ILC trend observed in the FBGA572 component.

The inconsistencies in joint morphology observed during X-ray imaging (Figs. 10 and 11) can be connected to the behaviour of flux during soldering. SP2 and SP3 being low-spread, less aggressive fluxes, the annular ring-like appearance of the joint in SP3- Fig. 9 (c) point to the possibility of joint formation occurring with incomplete wetting of the entire connector pad surface, called a Non-Wet Open (NWO)47. In case of NWO, component warpage during reflow results in sticking of solder paste to the component solder bumps away from the PCB surface, followed by complete reflow and solidification only on the component side. The irregular joint morphology (rough solder ball edges) for SP2- Fig. 9 (b) might point to a lower degree of NWO, or a joint failure phenomenon called Head-in Pillow (HIP) (the sticking, reflow and solidification of the PCB surface pad-level solder paste to the pad itself instead of with the component-level solder bump, which leads to an unformed, irregular joint)48,49,50. It must be noted that the relatively higher activity of the SP1 flux would also contribute to overall better wetting of the component bumps and PCB surface pads, thus improving joint quality, though at the expense of climatic reliability. This was shown by the well-formed joints under the FBGA572 (Fig. 9 (a)) indicate proper wetting of the bump and pad during reflow, along with adaptation to any possible component warpage. The choice of flux system/solder paste will thus affect the nature of joints formed, and the impact of warpage and tilting during reflow soldering.

An interesting effect of the GND/centre lug on warpage and tilting can also be observed in case of the FBGA244-GND and QFN48 components, where the lug likely provides a stabilising factor due to functioning as a heat sink, which can reduce reflow warpage of the components. This would of course bring forth the effects of other factors such as overall standoff height, solder amount and pitch distance, as has been mentioned earlier in this section. The FBGA244-GND component shows, as visible in Fig. 9a1, a pool of faintly-visible residue (marked by a yellow circle in Fig. 10, which could be either spilled solder flux or uncured/semi-cured solder paste) moving outwards from the GND/centre lug to the ball joints observed after testing. Similar effects have been observed in case of the QFP80 component (Fig. 9a2 and a3) as well, where small outcroppings are observable from the connectors, reducing the effective pitch distance in the specific areas. Darker-coloured outcroppings are likely improperly formed solder joints, which in this case overshoot the connector footprint, signifying a lowering of that part of the component from the designed standoff height, which causes the solder mass to be squeezed outwards (this can also cause complete connector shorting, in the worst case)51. This behaviour when combined with the high flux activity and spreadability of SP1, fits the electrical measurements showing overall higher leakage currents of SP1. An important feature of the QFP80 component is the large footprint and connector pitch distance between relatively small connectors located only at the edges of the component- this would allow for better outgassing of flux during soldering, leading to less reliability issues- reflected in the overall comparable (SP3) and even slightly lower (SP1 and SP2) measured leakage current for the QFP80 component despite its much larger size compared to the QFN and FBGA components. These interactions between flux behaviour and component design parameters shows the necessity for matching the intended use case with the right kind of flux system.

This study investigated the interaction and climatic reliability of different reflow flux systems with different commonly used electronic components. Functional analysis of the components by electrical measurement (leakage current) was combined with surface characterization methods of Scanning Electron Microscopy and X-ray imaging, to study the change in flux morphology with humid exposure, and the possible resultant during soldering. The conclusions are as follows:

Component design parameters such as connector layout and standoff height strongly impact overall component performance and reliability; the lowest standoff height of capacitors correlated to the highest leakage current under humidity exposure, followed by a similar effect of lower pitch distance in case of QFN48 increasing the leakage current.

Variation trend of leakage current was mirrored in Impedance measurements, showing promise for EIS as a non-destructive replacement for chronoamperometry (leakage current measurement) and SIR measurement in reliability testing.

Inclusion of thermal lugs as part of the component connector architecture increases flux quantity under components, along with obstruction to outgassing of flux during reflow, which can cause higher leakage currents during humidity exposure. The thermal lugs can conversely also provide structural support and act as a heat-sink (which is the intended function). This reduces occurrence of sharp thermal gradients, which minimize component tilting and warpage (which often introduce issues like uneven component standoff height and higher access of moisture at the lifted sides.

Flux system choice based on component design parameters (size, connector layout and standoff height), use-case (expected operational climates, packaging etc.) and outgassing possibilities (often determined by type of components in the vicinity, and their aspect ratios) is critical to component reliability. High-spreading flux systems with softer residues (higher tack) can improve reliability of components expected to undergo warpage. Smaller, lower standoff and pitch components would be suited to low-spreading fluxes with high residue hardness due to lower wettability requirements owing to smaller overall dimensions.

The datasets used and analysed during the current study are available from the corresponding author (Anish Rao Lakkaraju- [email protected]/[email protected]) on reasonable request.

Ambat, R., Conseil-Gudla, H. & Verdingovas, V. Corrosion in electronics, Encyclopedia of Interfacial Chemistry: Surface Science and Electrochemistry, pp. 134–144. https://doi.org/10.1016/B978-0-12-409547-2.13437-7 (2018).

Ambat, R. & Conseil-Gudla, H. Improving intrinsic corrosion reliability of printed circuit board assembly, Proceedings of the IEEE 18th Electronics Packaging Technology Conference, EPTC 2016, pp. 540–544. https://doi.org/10.1109/EPTC.2016.7861538 (2016).

Plieninger, R., Dittes, M. & Pressel, K. Modern IC packaging trends and their reliability implications, Microelectron. Reliab. 46 (9–11), 1868–1873. https://doi.org/10.1016/j.microrel.2006.08.008 (2006).

Wang, H., Ma, J., Yang, Y., Gong, M. & Wang, Q. A review of System-in-Package technologies: application and reliability of advanced packaging. Micromachines (Basel). 14 (6), 1149. https://doi.org/10.3390/mi14061149 (2023).

Institute of Electrical and Electronics Engineers, Reliability/CPMT, I. E. E. E., /ED Singapore Chapter., and P. & & Society, M. T. Components, 2010 12th Electronics Packaging Technology Conference: (EPTC 2010) : Singapore, 8–10 December 2010. IEEE (2010).

Wood, P. Rework Challenges for Leading Edge Components BGA, QFN and LED in Today’s Fast Moving Industry, in IPC APEX Proceedings, Iircuit Insight. Accessed: May 20, 2024. [Online]. Available: https://www.circuitinsight.com/programs/54959.html (2024).

Ghaffarian, R. Body of Knowledge (BOK) for Leadless Quad Flat No-Lead/Bottom Termination Components (QFN/BTC) Package Trends and Reliability. [Online]. Available: http://nepp.nasa.gov (2014).

Bixenman, M., & McMeen, M. Electrochemical reliability as a function of & component standoff, in Pan Pacific Microelectronics Symposium, pp. 1–9. https://doi.org/10.23919/PanPacific48324.2020.9059382 (2020).

Li, F. Solder Flux Chemistry and Climatic Reliability of Electronics: Optimization of Flux Chemistry for Robust Performance (Technical University of Denmark, 2021).

Piotrowska, K., Li, F. & Ambat, R. Transformation of reflow solder flux residue under humid conditions. Microelectron. Reliab. 123 https://doi.org/10.1016/j.microrel.2021.114195 (2021).

Lakkaraju, A. R., Conseil-Gudla, H. & Ambat, R. Study of interaction between reflow solder flux and humidity in relation to failures in electronics. IEEE Trans. Compon. Packaging Manuf. Technol. https://doi.org/10.1109/TCPMT.2024.3369076 (2024).

Article Google Scholar

Piotrowska, K., Jellesen, M. S. & Ambat, R. Water film formation on the PCBA surface and failure occurrence in electronics, Proceedings – 2018 IMAPS Nordic Conference on Microelectronics Packaging, NORDPAC, pp. 72–76. https://doi.org/10.23919/NORDPAC.2018.8423854 (2018).

Piotrowska, K., Verdingovas, V. & Ambat, R. Humidity-related failures in electronics: effect of binary mixtures of weak organic acid activators, J. Mater. Sci. Mater. Electron. 29 (20), 17834–17852. https://doi.org/10.1007/S10854-018-9896-0 (2018).

Piotrowska, K., Li, F. & Ambat, R. Thermal decomposition of binary mixtures of organic activators used in no-clean fluxes and impact on PCBA corrosion reliability, Solder. Surf. Mount Technol., 32 (2), 93–103. https://doi.org/10.1108/SSMT-05-2019-0020 (2020).

Shi, Y., Xiang, W. & Tolla, B. The Role of Organic Amines in Soldering Materials, in IPC APEX Expo Conference Proceedings, San Diego, CA: Circuit Insight (2015).

Li, F., Piotrowska, K., Jellesen, M. S. & Ambat, R. Alkanolamines as activators in no-clean flux systems: investigation of humidity robustness and solderability. J. Mater. Sci.: Mater. Electron. 32 (4), 4961–4981. https://doi.org/10.1007/s10854-020-05235-0 (2021).

Noh, B. I. & Jung, S. B. Behaviour of electrochemical migration with solder alloys on printed circuit boards (PCBs). Circuit World 34 (4), 8–13. https://doi.org/10.1108/03056120810918060 (2008).

Yao, Z. et al. Effect of Cu addition on the microstructure and mechanical properties of Sn-58Bi-0.5Ag solder alloys. J. Electron. Mater. 51, 3552–3559. https://doi.org/10.1007/s11664-022-09601-y (2022).

Article ADS CAS Google Scholar

Li, F., Jellesen, M. S. & Ambat, R. Comparative study of tripropylamine and naphthylamine as additives in wave solder flux: investigation of solderability and corrosion effects. J. Mater. Sci.: Mater. Electron. 33 (13), 10234–10250. https://doi.org/10.1007/s10854-022-08012-3 (2022).

Kulkarni, R. et al. Mar., Reliability study of electronic components on board-level packages encapsulated by thermoset injection molding, J. Manuf. Mater. Process. 4 (1). https://doi.org/10.3390/jmmp4010026 (2020).

Yang, Y., Lu, H., Yu, C. & Li, Y. Void formation at the interface in sn/cu solder joints. Microelectron. Reliab. 51 (12), 2314–2318. https://doi.org/10.1016/j.microrel.2011.06.026 (2011).

Bušek, D. et al. Flux effect on void quantity and size in soldered joints. Microelectron. Reliab. 60, 135–140. https://doi.org/10.1016/j.microrel.2016.03.009 (2016).

Synkiewicz, B. K., Skwarek, A. & Witek, K. Voids investigation in solder joints performed with vapour phase soldering (VPS). Solder. Surf. Mt. Technol. 26 (1), 8–11. https://doi.org/10.1108/SSMT-10-2013-0028 (2014).

Article CAS Google Scholar

Wang, P., He, D., Cao, V., & Su, J. Low temperature solder interconnect reliability and potential application in enterprise computer and automotive electronics, in Proceedings of SMTA International, Rosemont, IL (2018).

Vandevelde, B., Gonzalez, M., Limaye, P., Ratchev, P. & Beyne, E. Thermal cycling reliability of SnAgCu and SnPb solder joints: A comparison for several IC-packages, Microelectron. Reliab. 47 (2–3), 259–265. https://doi.org/10.1016/j.microrel.2006.09.034 (2007).

Arabi, F., Gracia, A., Delétage, J. Y. & Frémont, H. Effect of thermal and vibrational combined ageing on QFN terminal pads solder reliability. https://doi.org/10.1016/j.microrel.2020.113883ï (2020).

Lauser, S. Implementation of Electrochemical Impedance Spectroscopy (EIS) for Validation of Humidity Robustness of PCBA Design Elements ( PhD, Technical University of Denmark, 2022).

Li, F., Lakkaraju, A. R., Jellesen, M. S. & Ambat, R. Effect of flux activator in reflow process related flux residue on the climatic reliability of surface-mount electronic devices, J. Mater. Sci.: Mater. Electron. 34 (16). https://doi.org/10.1007/s10854-023-10708-z (2023).

Piotrowska, K., Ud Din, R., Grumsen, F. B., Jellesen, M. S. & Ambat, R. Parametric Study of Solder Flux Hygroscopicity: Impact of Weak Organic Acids on Water Layer Formation and Corrosion of Electronics, J Electron Mater. 47 (7), 4190–4207. https://doi.org/10.1007/S11664-018-6311-9 (2018).

Bahrebar, S. & Ambat, R. Time to failure prediction on a printed circuit board surface under humidity using probabilistic analysis. J. Electron. Mater. https://doi.org/10.1007/s11664-022-09668-7 (2022).

Bahrebar, S. & Ambat, R. Investigation of critical factors effect to predict leakage current and time to failure due to ECM on PCB under humidity. Microelectron. Reliab. 127, 114418. https://doi.org/10.1016/J.MICROREL.2021.114418 (2021).

Murphy, A., Chieng, R. & White, S. Radiographic contrast. Radiology Reference Article, Radiopaedia.org, Radiopaedia.org. https://doi.org/10.53347/rID-58718

Preece, J. W. & Webber, R. L. Toward a better Understanding of radiographic contrast. Oral Surg. Oral Med. Oral Pathol. https://doi.org/10.1016/0030-4220(82)90397-8 (1982).

Article Google Scholar

Piotrowska, K. Water Film Formation on PCBA Surface Investigation of Aspects Contributing To Premature Corrosion Failures and Safety Measures for Electronics Reliability Improvement (Technical University of Denmark, 2018).

Piotrowska, K. & Ambat, R. Residue-Assisted Water Layer Build-Up under Transient Climatic Conditions and Failure Occurrences in Electronics. IEEE Trans Compon Packaging Manuf Technol. 10 (10), 1617–1635. https://doi.org/10.1109/TCPMT.2020.3005933 (2020).

Mikhailov, E., Vlasenko, S., Martin, S. T., Koop, T. & Pöschl, U. Atmospheric Chemistry and Physics Amorphous and crystalline aerosol particles interacting with water vapor: conceptual framework and experimental evidence for restructuring, phase transitions and kinetic limitations, [Online]. Available: www.atmos-chem-phys.net/9/9491/2009/ (2009).

Dominkovics, C. & Harsányi, G. Effects of flux residues on surface insulation resistance and electrochemical migration, in ISSE 2006–29th International Spring Seminar on Electronics Technology: Nano Technologies for Electronics Packaging, Conference Proceedings, pp. 206–210. https://doi.org/10.1109/ISSE.2006.365387 (2006).

Capen, B., Bixenman, M., Mcmeen, M. & Fowler, E. SIR Glass Test Vehicle Designed to Characterize Process Materials, in Proceedings of SMTAI International, Minneapolis, MN (2021).

Minzari, D., Jellesen, M. S., Møller, P. & Ambat, R. On the electrochemical migration mechanism of Tin in electronics. Corros. Sci. 53 (10), 3366–3379. https://doi.org/10.1016/j.corsci.2011.06.015 (2011).

Tanaka, H. Factors leading to ionic migration in lead-free solder, Accessed: Mar. 10, 2024. [Online]. Available: https://www.test-navi.com/eng/report/pdf/FactorsLeadingToIonicMigrationInLead-freeSolder.pdf (2002).

Apelblat, A. & Manzurola, E. Solubility of oxalic, malonic, succinic, adipic, maleic, malic, citric, and tartaric acids in water from 278.15 to 338.15 K. J. Chem. Thermodyn. 19 (3), 317–320 (1987).

Article ADS CAS Google Scholar

Verdingovas, V., Jellesen, M. S. & Ambat, R. Influence of sodium chloride and weak organic acids (flux residues) on electrochemical migration of tin on surface mount chip components. Corrosion Eng. Sci. Technol. 48 (6), 426–435. https://doi.org/10.1179/1743278213Y.0000000078 (2013).

Tutone, M., Lauria, A. & Almerico, A. M. Theoretical determination of the pK a values of betalamic acid related to the free radical scavenger capacity: comparison between empirical and quantum chemical methods. Interdiscip Sci. 8 (2), 177–185. https://doi.org/10.1007/s12539-015-0101-3 (2016).

Comuzzo, P. & Battistutta, F. Acidification and pH control in red wines, in Red Wine Technology, pp. 17–34. https://doi.org/10.1016/B978-0-12-814399-5.00002-5 (Elsevier, 2018).

Szegezdi, J. & Czimadia, F. Prediction of dissociation constant using microconstants, in 227th National Meeting of the American Chemical Society, Accessed: May 28, 2024. [Online]. (2004). Available: https://docs.chemaxon.com/display/docs/attachments/attachments_1814016_1_Prediction_of_dissociation_constant_using_microconstants.pdf

Verdingovas, V., Jellesen, M. S., Rizzo, R., Conseil, H. & Ambat, R. Impact of hygroscopicity and composition of solder flux residue on the reliability of PCBA under corrosive conditions, in Proceedings of EUROCORR 2013, APA, [Online]. Available: http://www.eurocorr2013.org/ (2013).

Bastow, E. In Soldering, Understanding the Cause and Cure for NWO (Non-Wet Opens)., Indium Corporation Blogs. [Online]. (2024). Available: https://www.indium.com/blog/understanding-the-cause-and-cure-for-nwo-non-wet-opens.php. Accessed 29 January 2024.

Scalzo, M. Addressing the Challenge of Head-in-Pillow Defects in Electronics Assembly, in Proceedings of IPC EXPO Technical Conference, pp. 1133–1139. (2010).

Zhao, Z. et al. Effects of package warpage on head-in-pillow defect. Mater. Trans. 56 (7), 1037–1042. https://doi.org/10.2320/matertrans.MI201404 (2015).

Article CAS Google Scholar

Seelig, K. Head-in-Pillow BGA defects. Circuits Assembly. 19 (12), 28–31 (2008).

Google Scholar

Peterson, Z. Component Warpage Causes in a PCB, www.resources.altium.com. [Online]. Available: https://resources.altium.com/p/printed-circuit-board-defects-dealing-with-pcb-component-warpage. Accessed 10 March 2024.

Download references

The authors would like to acknowledge the financial support of the ELMAC project and thank the involved project partners and for their support during this study. The authors would also like to thank CreCon Industrial Consortium for Climatically Reliable Electronics (CreCon) for partial financial support. The authors would also like to thank Mr. Mark McMeen from STI Electronics Inc., Hunstville, AL, for the support in manufacturing the test PCBAs.

Center for Electronics Corrosion (CELCORR), Technical University of Denmark, Kongens Lyngby, Denmark

Anish Rao Lakkaraju, Helene Conseil-Gudla & Rajan Ambat

Magnalytix Inc, Nashville, TN, USA

Mike Bixenman

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

A.R.L. and R.A. wrote the main manuscript text. A.R.L. prepared all figures.H.C.G. and M.B. contributed to planning, sample preparation, experiments and materials characterization.All authors reviewed the manuscript.

Correspondence to Anish Rao Lakkaraju.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Lakkaraju, A.R., Conseil-Gudla, H., Bixenman, M. et al. Reflow solder flux residue and humidity interaction: investigation using real PCBA component designs. Sci Rep 15, 22496 (2025). https://doi.org/10.1038/s41598-025-05969-z

Download citation

Received: 08 August 2024

Accepted: 05 June 2025

Published: 02 July 2025

DOI: https://doi.org/10.1038/s41598-025-05969-z

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative