Corrosion Testing Procedures Adapt to Rising Air Pollution

Boards waiting for Corrosion Testing in Pollution Chamber

Intel develops new testing protocols after seeing uptick in corrosion levels on server boards and connectors used in Asia-Pacific region.

After discovering a “surprising” increase of corrosion in systems primarily from the Asia-Pacific region, Intel is now testing systems for damage potentially caused by air pollution, sulfur in particular. Concerns arose when systems returned by customers for quality and reliability analysis came back with an abnormal amount of corrosion on connectors and boards.

Anil Kurella Intel Technology and Manufacturing Group

Anil Kurella, failure analysis engineer, Intel Technology and Manufacturing Group, prepares boards for corrosion testing in a pollution chamber.

“We are starting to see corrosion on systems (returned to Intel) that are higher than what we would have expected,” said Tom Marieb, a vice president in Intel’s Technology and Manufacturing Group. “They were disproportionately from Asia-Pacific and it gave us pause. The only other times we had seen this level was from known industrial usage segments like inside a factory, not data centers that are supposed to be controlled, sealed-off environments with air conditioning.”

The Asia-Pacific region — China and India especially — is known to produce higher rates of corrosion due to a hot and humid environment. However, what was observed on server boards and connectors had trended toward the abnormal starting roughly two years ago. “We noticed an uptick and it wasn’t just from one site,” Marieb said, adding that the Intel parts “were fine.” Still, that didn’t lessen the need to investigate.

“We need to make sure our parts are reliable wherever they are being used and that Intel components are robust against potential issues,” said Marieb, who also serves as director of Technology Development and Manufacturing Quality and Reliability. “To do that we need to understand where our parts are being used, what the environment is and determine the physics of failure rate, meaning how do humidity, temperature and contaminant level impact the speed of corrosion?”

Existing tests and industry standards based on lower pollutant levels than what was discovered were no longer meeting the needs according to Marieb and testing procedures needed to adapt. “We saw that the standards of the ’80s were not applicable in the worsening environments,” he said. “We realized that we had no real science behind understanding how our parts behave in environments that have high-pollutant content along with humidity and heat. We normally have a million tests to look at aspects of our parts, but this is new and unique.”

Because previous testing protocols did not account for air pollution, the corrosion rate was significantly under-accelerated, according to Marieb. The added testing scheme required Intel to design and build a new type of mixed flow gas chamber at a cost of $300,000. The customized machinery, sitting at an Intel facility in Hillsboro, Ore., is described by Marieb as resembling “a huge oven with many wires hanging off of it.” The chamber, in operation since the beginning of the year, gives engineers a higher level of control when introducing contaminants that mimic what is in polluted air that may be causing an increased rate of corrosion.

Boards being placed to pollution chamber to test for corrosion

Boards are placed in the pollution chamber, which can mimic different air conditions to gauge rates of corrosion.

“In most cases it’s sulfur,” said Marieb, citing International Electrotechnical Commission figures. “Hydrogen sulfide and sulfur oxide are the worst for corrosion.”

Determining the corrosion rate and the factors contributing to it will help Marieb and his team answer the question, “How do we make sure our parts will survive?”

“It’s a relatively difficult problem,” he said. “It’s not just sticking things in the chamber and watching parts corrode. It could also be the voltage that’s affecting the corrosion rate or something else. Marieb said his team is within a year of having some answers to what he calls “an industry-wide concern.” When these findings are made, they can take shape in a number of ways.

“We can try to create a new standard through standards bodies like the Electronic Industries Association,” he said. “Another way is to go to the specific people and companies involved and share our data. We can also publish our findings, our science, and hopefully make a change in the industry.”

Corrosion Testing Procedures Adapt to Rising Air Pollution

Boards waiting for Corrosion Testing in Pollution Chamber

Intel develops new testing protocols after seeing uptick in corrosion levels on server boards and connectors used in Asia-Pacific region.

After discovering a “surprising” increase of corrosion in systems primarily from the Asia-Pacific region, Intel is now testing systems for damage potentially caused by air pollution, sulfur in particular. Concerns arose when systems returned by customers for quality and reliability analysis came back with an abnormal amount of corrosion on connectors and boards.

Anil Kurella Intel Technology and Manufacturing Group

Anil Kurella, failure analysis engineer, Intel Technology and Manufacturing Group, prepares boards for corrosion testing in a pollution chamber.

“We are starting to see corrosion on systems (returned to Intel) that are higher than what we would have expected,” said Tom Marieb, a vice president in Intel’s Technology and Manufacturing Group. “They were disproportionately from Asia-Pacific and it gave us pause. The only other times we had seen this level was from known industrial usage segments like inside a factory, not data centers that are supposed to be controlled, sealed-off environments with air conditioning.”

The Asia-Pacific region — China and India especially — is known to produce higher rates of corrosion due to a hot and humid environment. However, what was observed on server boards and connectors had trended toward the abnormal starting roughly two years ago. “We noticed an uptick and it wasn’t just from one site,” Marieb said, adding that the Intel parts “were fine.” Still, that didn’t lessen the need to investigate.

“We need to make sure our parts are reliable wherever they are being used and that Intel components are robust against potential issues,” said Marieb, who also serves as director of Technology Development and Manufacturing Quality and Reliability. “To do that we need to understand where our parts are being used, what the environment is and determine the physics of failure rate, meaning how do humidity, temperature and contaminant level impact the speed of corrosion?”

Existing tests and industry standards based on lower pollutant levels than what was discovered were no longer meeting the needs according to Marieb and testing procedures needed to adapt. “We saw that the standards of the ’80s were not applicable in the worsening environments,” he said. “We realized that we had no real science behind understanding how our parts behave in environments that have high-pollutant content along with humidity and heat. We normally have a million tests to look at aspects of our parts, but this is new and unique.”

Because previous testing protocols did not account for air pollution, the corrosion rate was significantly under-accelerated, according to Marieb. The added testing scheme required Intel to design and build a new type of mixed flow gas chamber at a cost of $300,000. The customized machinery, sitting at an Intel facility in Hillsboro, Ore., is described by Marieb as resembling “a huge oven with many wires hanging off of it.” The chamber, in operation since the beginning of the year, gives engineers a higher level of control when introducing contaminants that mimic what is in polluted air that may be causing an increased rate of corrosion.

Boards being placed to pollution chamber to test for corrosion

Boards are placed in the pollution chamber, which can mimic different air conditions to gauge rates of corrosion.

“In most cases it’s sulfur,” said Marieb, citing International Electrotechnical Commission figures. “Hydrogen sulfide and sulfur oxide are the worst for corrosion.”

Determining the corrosion rate and the factors contributing to it will help Marieb and his team answer the question, “How do we make sure our parts will survive?”

“It’s a relatively difficult problem,” he said. “It’s not just sticking things in the chamber and watching parts corrode. It could also be the voltage that’s affecting the corrosion rate or something else. Marieb said his team is within a year of having some answers to what he calls “an industry-wide concern.” When these findings are made, they can take shape in a number of ways.

“We can try to create a new standard through standards bodies like the Electronic Industries Association,” he said. “Another way is to go to the specific people and companies involved and share our data. We can also publish our findings, our science, and hopefully make a change in the industry.”