A review of best-in-class controls for maintaining factory worker health.

The National Institute for Occupational Safety and Health (NIOSH) received a confidential employee request for a Health Hazard Evaluation (HHE) at an electronics manufacturer specializing in printed circuit board fabrication, assembly, and testing for different end-user applications (sidebar). At the time of our evaluation, approximately 2,300 employees worked at the facility in either the medical section plant or the defense and aerospace section. Employees were primarily concerned about exposure to solder paste and fumes, as well as dust and noise. They reported coughs, burning eyes, nosebleeds, voice loss, headaches, sinus infections, bronchitis, and respiratory problems.

At the time of our evaluation, the medical section had four wave soldering lines and eight surface mount lines. Wave solder lines 1, 2, and 3 used Pb-free solder (96.5% tin), and wave solder line 4 used solder composed of 63% tin and 37% lead. The defense and aerospace section had five wave solder lines and six surface mount lines. It also had ruggedization, conformal coating and bonding operations, where the finished PCBs are fitted with additional structural supports and hand-brushed or sprayed with an acrylic copolymer to provide increased environmental and mechanical protection. Spraying during conformal coating was conducted in a ventilated, open-face, bench-top spray booth. Employees wore safety glasses and air-purifying, elastomeric, half facepiece respirators with organic vapor cartridges while spraying. At least once per shift, wave solder operators clean solder dross by using a ladle to remove dross floating on top of the molten solder (Figure 1). Residual molten solder inadvertently collected during this operation is separated with a sieve, and the remaining dross disposed of in a drum sealed with a lid. Employees are required to wear heat-resistant gloves over disposable nitrile gloves, a face shield, and an apron when performing dross cleaning. In addition to dross cleaning, employees periodically clean and maintain the wave solder machines. Finally, the facility has an auto insertion (AI) operation, which involves machine insertion of components onto a PCB by punching through it.

Assessment. During our first visit to the facility, we conducted confidential medical interviews with 40 employees, observed work processes and practices, collected general area air samples for volatile organic compounds (VOCs), and collected surface wipe samples for lead and tin. We also reviewed air sampling records, the respiratory protection program, injury and illness records, and material safety data sheets. After sharing our findings with employee and management representatives in an interim letter, we returned to the facility to collect personal breathing zone and general area air samples for lead and specific VOCs. We also measured noise levels at the AI stations, evaluated room acoustics near the air rotation units (ARUs), collected hand wipe samples to assess lead contamination on skin, and evaluated the effectiveness of the local exhaust hoods for the wave solder and surface mount machines (Figure 2).

Occupational exposure limits. NIOSH investigators use both mandatory (legally enforceable) and recommended occupational exposure limits (OELs) for chemical, physical, and biological agents as a guide for making recommendations. OELs have been developed by Federal agencies and safety and health organizations to prevent the occurrence of adverse health effects from workplace exposures. They suggest levels of exposure to which most employees may be exposed up to 10 hr. per day, 40 hr. per week for a working lifetime without experiencing adverse health effects. However, a small percentage of employees may still experience health effects due to personal susceptibility, a preexisting medical condition, and/or hypersensitivity (allergy). In addition, some hazardous substances may act in combination with other workplace exposures, the environment, or with medications or personal habits of the employee to produce health effects, even if occupational exposures are controlled at the level set by the exposure limit. Also, some substances can be absorbed by direct contact with the skin and mucous membranes in addition to being inhaled, which contributes to the individual’s overall exposure.

Results and Discussion

One of six personal breathing zone air samples for lead exceeded the Occupational Safety and Health Administration action limit (OSHA AL) of 30 micrograms per cubic meter (µg/m3) and was close to the OSHA permissible exposure limit of 50 µg/m3. This employee, a wave solder operator in the DAS, was cleaning the wave solder machines without wearing a respirator. The employee was exposed to an airborne lead concentration of 49 µg/m3. However, with this one exception, wave solder operators had lead exposures well below the OSHA AL.

Our surface sampling showed the presence of lead and tin on work surfaces in both sections of the plant. Currently, there are no OELs for surface metal contamination in occupational settings. We also sampled tables in two of the break rooms and found detectable levels of lead and tin in one of the break rooms. This suggests that workplace contamination is being tracked into the break rooms by employees’ footwear, clothing or hands, and that these areas should be kept cleaner.

Additionally, despite hand washing prior to sample collection, three of seven hand wipe samples tested positive for lead.

Full-shift noise exposures for the AI operators in the medical and defense and aerospace section were well below the NIOSH recommended exposure limit. Because telephone communication is not required in the production areas, and communication between employees is minimal, the louder noise levels experienced by employees with work stations near the ARUs were within the criteria specified by the balanced noise criteria (NCB) 55-70 curves. The NCB curves are a set of noise criteria for occupied interior spaces, devised to limit noise to levels at which speech can be reasonably understood. Refer to “Resources and Links” for more information on the NCB curves, as well as Appendix B of the HHE report (referenced in this section).

Our ventilation evaluation revealed that several local exhaust hoods were not effectively capturing process emissions. These included three hoods in the medical section and two hoods in the defense and aerospace section. This could have been due to local exhaust ventilation systems being imbalanced or improperly maintained. Air sampling results for specific VOCs indicated that employee exposures were well below all applicable OELs.

Although both the medical and defense and aerospace section share the main workspace and have similar tasks and equipment, the health concerns originated exclusively from employees in the medical section. Of the 40 medical section employees we interviewed, 23 did not report any work-related symptoms. The most commonly reported symptoms were upper respiratory, including runny nose, cough, and sinusitis; fatigue (frequently related to overtime work), and voice loss. However, these symptoms are also common in the general population, and we could not attribute them to the exposures documented.

Last, we found inconsistencies between the facility’s written respiratory protection program and employee practice. The written respiratory protection program required respirators to be worn when cleaning wave solder machines. However, it did not identify the appropriate type of respirator that should be worn for this task, and we did not observe employees wearing respirators when performing this activity. In addition, employees were voluntarily wearing respirators during spraying in the conformal coating area.

Recommendations

We made a number of recommendations to the facility to improve employee health and safety. Many of these recommendations may also apply to readers’ facilities:

• Use engineering controls such as portable exhaust hoods when removing solder dross and cleaning wave solder machines. Periodically evaluate the local exhaust hoods to ensure they are performing per manufacturer’s specifications. Evaluate their effectiveness at least every three months, and within five days of a production, process, or control change.
• Comply with all requirements of the OSHA lead standard (29 CFR 1910.1025). This includes instituting an industrial hygiene monitoring program to assess lead exposure for employees working with leaded solder, or for employees having the greatest potential for lead exposure. Additionally, institute a maintenance program for evaluating enclosed local exhaust hoods used to control lead exposure.
• Improve general housekeeping practices to ensure break rooms and work surfaces are clean. Use commercial cleaning wipes that are made for removing lead and other heavy metals every shift to keep work surfaces such as control consoles clean. Use high-efficiency particulate air vacuum cleaners in work areas to minimize dust in the air.
• Practice good personal hygiene. Thoroughly wash hands before eating, before and after using the restroom, and before leaving work. Eating, drinking, smoking, and applying cosmetics in work areas should be prohibited.
• Ensure that the respiratory protection program meets all applicable requirements of the OSHA Respiratory Standard (29 CFR 1910.134). The written respiratory protection program should identify job tasks requiring a respirator to be worn, the type of respirator needed, and the locations where respirator use is required. Evaluate respiratory hazards for job tasks currently requiring respiratory protection to ensure protection is adequate. The program should also note any job tasks for which employees voluntarily wear respirators.
• Encourage employees to report work-related health concerns. Consult a physician trained in occupational medicine (see “Resources and Links”).
• Train employees about potential exposures and how to protect themselves from hazards, and promptly communicate any health and safety changes to employees.

Sidebar:

The Health Hazard Evaluation Program

Based on a federal law, NIOSH conducts Health Hazard Evaluations (HHE) to investigate possible workplace health hazards. Employees, employers or union representatives can ask our comprehensive team of experts to investigate their health and safety concerns by requesting an HHE. Our team contacts the requestor and discusses the problems and how to solve them. This may result in sending the requestor information, referring the requestor to a more appropriate agency, or making a site visit, which may include environmental sampling and medical testing. If we make a site visit, we prepare a report of our investigation that includes recommendations specific to the problems found, as well as general guidance for following good occupational health practices. HHE reports are available on the Internet (http://www.cdc.gov/niosh/hhe/).

Resources and Links

1. NIOSH HHE program information, cdc.gov/niosh/hhe/HHEprogram.html.
2. Link to this HHE report: cdc.gov/niosh/hhe/reports/pdfs/2007-0201-3086.pdf.
3. Code of Federal Regulations (CFR),  29 CFR 1910.95, US Government Printing Office, Office of the Federal Register.
4. L.L. Beranek, “Criteria for Noise and Vibration in Communities, Buildings, and Vehicles,” Noise and Vibration Control, Rev. ed. Cambridge, MA: Institute of Noise Control Engineering; 1988:554-623.
5. L.L. Beranek, “Balanced Noise Criterion (NCB) Curves,” J Acoust Soc Am., 1989; 86(2):650-664.
6. Acoustical Society of America, ANSI S12.2-1995, “Criteria for Evaluating Room Noise,” 1995.
7.  Association of Occupational and Environmental Clinics (aoec.org).
8. American College of Occupational and Environmental Medicine (acoem.org).

Srinivas Durgam was an industrial hygienist at NIOSH (cdc.gov/niosh) during the time of this evaluation; srinivasdurgam@hotmail.com. Chandran Achutan, Ph.D., is assistant professor at the University of Nebraska Medical Center (unmc.edu); cachutan@unmc.edu. Carlos Aristeguieta, M.D., was a NIOSH medical officer during the time of this evaluation; arcwor@gmail.com. Maureen Niemeier is a freelance technical writer; mtnrtn@fuse.net.

Testing parts for gaps reveals which parts are electrically intact but prone to future delamination.

In an ideal world, plastic-encapsulated microcircuits (PEMs) would leave the component manufacturer in pristine condition. BGAs, QFNs, PQFPs, TOs and other package types would have no electrical or structural anomalies that could lead to electrical failures. The components would arrive for assembly in this condition, and would pass through handling, reflow and testing with no alteration of their ideal state. They would give years of flawless service.

In the real world, things do not always work out like this. A component may pick up a gap-type structural anomaly – delaminations, voids and cracks are the common ones – somewhere between the molding process and post-assembly shipment. Many of these gap-type anomalies make no difference in the proper functioning or lifespan of the component. Some, though, depending on size and location within the package, can expand until they break a connection. Even without expanding, gaps are the locations where moisture tends to collect, whether it has percolated through the mold compound or entered by a crack along a lead finger. The combination of moisture and contaminants, which can arrive by the same routes, can promote corrosion that may lead to electrical failure.

The least dangerous gap-type PEM anomaly is probably a small void (air bubble) in the mold compound not near to or in contact with any other surface such as a wire or the die. Generally such an anomaly is not considered a defect. Far more dangerous are defects such as delamination between the mold compound and die face. Normal thermal cycling, with repeated coefficient of thermal expansion stresses, can cause the delamination to expand across the face of the die until it shears off a wire bond.

There are numerous other gap-type anomalies, and many pose a greater or lesser threat to electrical performance. The important questions during assembly are:

• What percentage of the components in a given lot has internal structural anomalies?
• How likely are these anomalies to cause eventual electrical failure?

Sonoscan’s applications laboratory, which images lots of components for assemblers, routinely uncovers the answers to both questions. The assembler that will use the components writes the rules that will distinguish harmless anomalies from dangerous defects. Acoustic micro imaging systems image and analyze internal gap-type features, which reflect ultrasound at higher amplitude than other internal features such as the die or lead fingers.

PEMs in which a gap-type feature has already broken a connection will, of course, fail electrical tests. The purpose of acoustic screening is to identify and remove those PEMs that are electrically intact, but that have gap-type defects likely to cause field failures. For commercial products, defects of a certain size may be acceptable if the risk of failure is low. In military, aerospace and medical products, the degree of risk permitted ranges from very low to zero.

The outcome of acoustic screening varies greatly from one lot of components to another. It is not uncommon for a lot of components to have up to 5% rejects. The accept/reject criteria are defined by the user of the part and are applied to internal anomalies visible in the acoustic images.

Acoustic images reveal all sorts of anomalous situations. The ultrasound pulsed into a component is reflected from material interfaces at many depths, but the echoes can be selected by their arrival time to limit the acoustic image to a desired depth. The acoustic image of the BGA in Figure 1 captured the depth from just below the top surface of the mold compound to just above the die, and included the tops of the wires (at center). The red and yellow features are voids in the mold compound. Such isolated voids are generally considered harmless. But this BGA has a great many voids, running down to sizes so small that they get lost in the tiny white features created by the distribution of particles in the mold compound.

But by using a new technique to image the bulk of the mold compound in many sequential slices, the Sonoscan lab found the real danger: The largest void, at top, extends vertically from the substrate upward almost to the top surface. Moisture and contaminants that percolate through its thin wall have a free ride to the center of the package.

Figure 2 is the acoustic image of the die surface in a BGA that would almost certainly pass electrical tests, but that is loaded with potential disaster. The red areas at each corner of the die face are small delaminations of the mold compound from the die. At this early stage, the delaminations are unlikely to have yet broken a wire bond on the die. The normal thermal cycling of service, however, is what makes die face delaminations so perilous; they may expand laterally until they find and break a wire bond.

A PEM that is being tested in a laboratory that determines Moisture Sensitivity Level (MSL) is initially dried to remove existing moisture, then moisturized to a specific level. It next goes through reflow three times, although it is not soldered to the board, since it will later be sectioned physically. The three passes through reflow simulate the experience of a component on a board that is reflowed twice and is then reworked. After the third reflow, the PEM undergoes a basic electrical test to find opens. It is then screened acoustically to look for gap-type defects, and then physically sectioned to see the defects optically. The acoustic image is the guide for selecting the best location for physical sectioning. Some defects, though dangerous, cannot be seen after physical sectioning because they are too thin. Delaminations as thin as 1 µm reflect virtually all of the ultrasound and are imaged strongly, but such a thin gap may be invisible optically.

MSL test personnel often encounter a PEM whose acoustic image shows multiple problems – voids, delaminations, even cracks – but which passed the electrical tests. Such a PEM may be loaded with structural anomalies that will cause electrical failure under the stress of service, but none of those anomalies happens to have caused a problem yet. In a production environment, such a “train wreck” can sail through basic electrical testing and go right into production.

Figure 3 is an acoustic image of a high-risk PEM. It has two types of defects. First, the die paddle is delaminated (red) from the mold compound around the periphery of the die. This is a moderately serious condition; the delamination could expand and move under the die and the die attach material. But in two locations on the lead fingers is another dangerous condition: delaminations on adjacent lead fingers. The wall of mold compound between the lead fingers is very thin, and can easily be dissolved by corrosion, causing a short between the lead fingers.

The same result – a short between adjacent lead fingers – can be triggered by a different mechanism often encountered in Sonoscan’s laboratories. In some PEMs, the tape (white in Figure 2) looks red or yellow in the acoustic image because it is delaminated from the mold compound. Although frequently considered unimportant, corrosion along such a delamination can also put two adjacent lead fingers in electrical contact.

The locations and types of gap-type defects are largely determined by the package design. Delamination of the mold compound from a lead finger, for example, may not even be considered a defect if a package has long lead fingers on which there are a few small delaminations. They are unlikely ever to grow sufficiently to like the outside atmosphere and the die or the wire bonds. But a package design having a very small die and very long wires connected to very short lead fingers is a different story. The distance from the outside atmosphere to the outer wire bonds is very short, and can be bridged even by a small delamination. The very long wires are also vulnerable to wire sweep during molding.

One of the defects MSL test personnel expect to see in some PEMs is the popcorn crack. Popcorn cracks form when moisture within the package turns to steam during reflow and expands in volume by about 1600 times. Popcorn cracks are not as frequent in production as they were a decade or two ago, but they still occur. They usually originate at the die attach level and travel downward to the bottom surface of the package. Very few travel to the top surface, and even fewer go sideways. A popcorn crack exiting the bottom of the package is invisible during production, but over the life of the components provides a convenient path for moisture and contaminants to reach the die.

Acoustic screening uncovers all sorts of things. Die face delaminations have a reputation for being lethal, but there are a very few specific packages in which, users
have found, small die face delaminations practically never cause electrical failure, probably because of a lucky combination of design and materials. Acoustic screening lets the user check the initial state of these delaminations. Another user might find that a lot of familiar PEMs suddenly shows numerous, large delaminations in a large percentage of components. These may be counterfeit PEMs, starting on their second lifetime of service. Or the defects may simply result from a change in mold compound or fabrication processes.

The most dramatic outcome of acoustic screening is the avoidance of mass field failures. In less dramatic scenarios, the weeding out of problem components gives a boost to overall product reliability.

Tom Adams is a consultant at Sonoscan Inc. (sonoscan.com); tom100adams@comcast.net.