August_EDFA_Digital

edfas.org 55 ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 24 NO . 3 GUEST COLUMN the ability to control the known and evident, as well as insight into possible/potential unknowns, and finally the presence or absence of management support. All semiconductor tools, whether intended for the laboratory or the wafer fabrication facility (fab) have a published set of environmental operating specifications. They cover topics like temperature and humidity, electric power require- ments and power quality, electromagnetic fields, floor vibration, acoustic disturbances, air movement, gas and vacuum services, chilled water heat exchange, lighting, ESD (electrostatic discharge) prevention, andmore. Some are more comprehensive or strict than others. Adhere to the specs and you stand a great chance that your tool will function at its best. Great care goes into the design of amodern wafer fab. And yet manufacturing tools typically have to function in a pretty hostile space. Ballroom layouts tend to be noisy with a lot of mechanical clamor, may feel “windy” due to the high rate of air exchange, and are quite equipment- dense. To counter this, a portion of the tool cost often includes full environmental enclosures to shield them from their neighbors and the room itself. Tools sit on isolation pedestals, while pumps and other facilities are isolated on the floor below. Lab tools are typically not as robust and thus rely on a much higher level of external environmental control to survive. Tool skins/shells, if they exist at all, tend to bemore cosmetic than functional. This dictates that lab tool space be designed to an elevated standard and may need to be highly customized to meet the specific needs of the tool contained therein. Unfortunately, allocation of lab areas is too often an afterthought and rarely a “designed from the ground up” portion of the overall facility. Lab spacemay be carved out of “office” space if management likes you. Or it may be assigned fromreal estate farmore hostile and undesirable Y ou canprobably relate to the following: You’ve spent countless hours over the past two years surveying the field of new instruments, working with vendors over specs and traveling to do initial demos. Solidly con- vinced you can’t meet the requirements of the next-gen process nodewith your old tool, you heatedly arguedwith management to justify the need, and took considerable abuse from the Capital Appropriations committee. You knowthedrill. Labs are expensiveoverhead, andadequate funding to do your job is a rare luxury. But you prevailed and designed/supervised modification of a space for the tool to the best of your knowledge and ability, and with the limited funds you were given. The tool was built and you completed a highly successful factory acceptance test. The vendor has spent the past few weeks assembling and rerunning the factory acceptance test in your lab. But then comes the bad news: something is seriously wrong. The exemplary preinstall results cannot be reproduced in your facility. The exact issue isn’t important. The factory is consulted, experts flown in, and critical parts are exchanged. Then comes the truly bad news. It’s not a tool build issue. A third-party site survey expert is brought in to help explain it to your management. There is an external environmental factor (or perhaps a combination of low- level factors) in your lab that’s impacting performance. It’s caused by something that wasn’t noted or understood or caughtmonths beforewhen the contractswere signed. Or maybe it was noted in the initial survey report, but your actions to counter it didn’t properly address the problem, or worse, the lab rearrangement process amplified it! No tool tuning by the vendor is going to make it go away. Now what? There are probably quite a few factors that contributed to the above, but due to editorial length constraints we’ll have to keep it light. For a more in-depth analysis check out “Building a Better Lab Space” to be presented at ISTFA in November. The main issues come down to knowledge and experience, team support and resident experts, PLAN WELL IN ADVANCE TO ENSURE YOUR LAB MEETS HIGH-RESOLUTION TOOL REQUIREMENTS Steven Herschbein, Independent Consultant steven.herschbein@gmail.com Herschbein