Printed circuit boards (PCBs) are built using ever increasing combinations of materials. One aspect of the various materials and processes used is cleanliness. Typically, when we discuss the cleanliness of a PCB, we are discussing the absence of harmful residues or contaminants. These harmful residues or contaminants can be classified into two major categories: ionic and non-ionic. Ionic residues are those materials that, in presence of moisture, disassociate into negatively and positively charged species. Once this occurs, the conduction of the resulting solution increases. Non-ionic residues are the organic species that can remain on a PCB after production. These species are typically polymers, oils, or greases.

Each of these classes of species can impact the functionality and reliability of electronic devices on which they are present. Typically, ionic contaminants are the species that cause the most concern. The two most common failures due to ionic contamination are corrosion and dendrite growth, see Figure 1 and 2. Both of these conditions can cause device failure. The most common sources of ionic contamination are flux, cleaning fluids (e.g. tap water), and plating chemistry residue (from the surface finish). The EMPF has observed several cases of each of these contributing to failure of a fielded electronic device. Figure 2. is an image of a pin that was corroded due to improper removal of a water-soluble flux. This ionic residue is believed to be behind the catastrophic field failure of a customer device sent to the EMPF for analysis seen in Figure 3. In this case, it is believed that ionic contamination lead to a dendrite growth, which in turn lead to a short that pulled an excessive amount of current through the connector, causing the destruction observed.

Non-ionic contamination is not usually of great concern by most electronic device manufacturers, but this does not mean that it does not affect reliability. The most common failure modes due to non-ionic contamination are reduced solderability, lack of connectivity and sensor malfunction. Since non-ionic contaminates are nonconductive, they can disrupt the flow of electricity through these connectors if they are present on an edge card connector or inside a socket. The EMPF has observed several problems with insulating films present on gold edge connectors.

Test Methods
Due to the fundamental differences of these contaminants, they require different methods of testing. The most common test applied is a derivative of the resistivity of solvent extract, or ROSE test. The most basic version of this test employs a solution into which the PCB is placed, the ionic contaminates dissolve into the solution, and the resulting change in conductivity is measured using a conductivity meter. Most assembly houses use an automated version of this test. These include OmegaMeter, Ionograph, ZeroIon, and Contaminometer. The premise behind all of these is the same: calibrate with a known concentration of salt solution, place the PCB inside the test cell, measure the resulting change in conductivity. This response is generally reported as a number of sodium chloride (NaCl) equivalents per unit area. These testers are good for quality control, but are not suited for investigations or failure analysis. These devices were designed to be used with RMA fluxes and may act erratically with PWBs fabricated with low residue fluxes [1]. In addition, the results from one piece of equipment are not equivalent to another. For example, an Ionograph result of 10 mg NaCl eq./in2 is not the same as 10 mg NaCl eq./in2 on a ZeroIon. The MILP- 551100 document does have a table of equivalency factors. Once again, this was developed using RMA fluxes and is not applicable to PCBs assembled using other flux systems.

An ion chromatograph is the most common tool for precision testing and process baseling. This system can quantify and identify specific ionic species that are present on an electronic device. The most common test method is the IPC TM-650 2.3.28. The device is placed into an ionically-clean bag (e.g. Kapak), and is immersed in an extract solution of 75 percent isopropyl alcohol (IPA) to 25 percent deionized water. It is important to be sure that the sampling procedures do not introduce any ionic contaminates. The EMPF laboratories use HPLC grade IPA and 17 MW or better deionized water to insure an ionically-clean extract solution. The PCB is extracted in the solution for one hour at 80oC, the resulting extract is then injected into the ion chromatograph (IC). The IC then sepa rates and detects each individual ion for which it was calibrated. For an overview of ion chromatography see http://pubs.acs.org/journals/chromatography/chap3.html.

The EMPF uses a Dionex DX-500 system, which is typically used to measure fluoride (F-), chloride (Cl-), bromide (Br-), nitrate (NO 3 -), nitrite (NO 2 -), phosphate (PO 4 2 -), sulfate (SO 4 2 -), and weak organic acids (WOAs). The results of the IC, in terms of mg/L concentration, of each of these species can then be used to calculate the concentration per unit area of each of the individual species, knowing the volume of extract used, and the surface area tested. Using this information, a trained scientist or engineer can usually deduce the source of the ionic contamination. For example, a customer using a low residue flux observed intermittent failures. IC testing detected elevated amounts of chloride (22 mg/in 2 ) on assemblies. Testing of the bare boards used for these assemblies also showed high levels of chloride (27 mg/in 2 ). From this, it was determined that the fusing fluids used during the HASL process to fabricate the boards remained on the boards, carrying through to the final assemblies.

Surface insulation resistance (SIR) testing is an environmental test that measures the effect of ionic contamination. A test board with a comb pattern is processed and placed at elevated temperature and humidity (often 85°C and 85 percent relative humidity). The presence of ionic contamination lowers the insulation resistance of the material between conductors. The measured resistance is often compared with that of test boards that have not been exposed to flux residues. The IPC has published a Surface Insulation Resistance Handbook (IPC-9201) which gives an overview SIR test methods, materials, equipment, and interpretation of the results.

Water drop testing is a way of observing qualitative surface insulation resistance without having to purchase expensive equipment. A voltage is applied across a "Y" pattern, or across any pads or traces that are in the area of interest. A drop of water is placed on the board so that it bridges the two conductors. The water drop is viewed under an optical microscope, and the time it takes for a short to form between the conductors is measured. The greater the ionic cleanliness, the longer it will take for a dendrite to bridge the conductors.

There have been many debates in the literature regarding the correlation between all these ionic test methods. A recent paper presented by Michael Weekes of Phoenix International at the SMTA International Meeting in Chicago [2] is an example of a study correlating the various ionic testing methods. For testing non-ionic species, the most common method is Fourier-Transform Infrared (FT-IR) spectroscopy. FT-IR is a spectroscopic method where the absorption and transmission of infrared light is used to determine the structure of an organic species. Most polymers and organic species react in a very specific way to infrared light. Using a FT-IR spectrometer the response of the target can be measured and compared against a database on the PC or on the web for identification of the target species. The EMPF laboratories have used this technique to identify a material present on a sensor, then measure other suspected species against the unknown. FT-IR allowed for the identification of the unknown residue and a recovery plan was then implemented.

The EMPF laboratories are well equipped to help with cleanliness testing, and using the demonstration factor floor adjacent to the laboratories, we can test a variety of cleaning methods, saponifier chemistries, and paste chemistries in order to aid in determining the optimum process for maximum