Fermi National Laboratory

Volume 23  |  Friday, December 1, 2000  |  Number 20
In This Issue  |  FermiNews Main Page

SiDet Adds Precision

by Kurt Riesselmann

Sabin Aponte wire bonds silicon microstrips and integrated circuits They wear special clothing. They work with special machines. And they have special skills.

Employees at Fermilab's Silicon Detector facility must have the "right touch." Dropping a screw or making a wrong move can destroy a day's work--or even more. Operating expensive precision machines and handling delicate silicon wafers are a large part of their daily work. And on top of it all, they need to finish their projects on time. With Collider Run II just around the corner, experimenters rely on the dedicated expertise of the SiDet technicians.

"We were concerned whether we would be able to meet the tight schedule," said Jeff Spalding, project manager of the CDF silicon detector. "But the technicians at SiDet really exceeded our expectations." To his surprise, the pace of production was set by the delivery of components from the vendors.

Silicon detectors will be at the core of the upgraded CDF and DZero experiments, which will start operating in March 2001. Arranged in layers, the silicon devices record the tracks of particles escaping from high-energy proton-antiproton collisions. Tiny electronic chips amplify the signals created inside the silicon and transmit them to a computing unit that reconstructs the tracks.

Combining these results with data obtained from other devices inside the 50-foot-high CDF and DZero detectors, physicists are able to identify the types of particles produced and study their interactions. Using their improved detectors, physicists will obtain very precise information on the characteristics of the top quark, the heaviest elementary particle ever observed. Recent results also indicate that they have a chance of finding the Higgs boson, the key to explaining the origin of mass in our entire universe.

Producing silicon detectors is a science of its own. Physicists and engineers spent years designing and building small prototypes of the new detectors. To achieve their physics research goals, they used computer simulations to optimize the shape and location of the silicon wafers. Interestingly, the two collaborations favored different configurations.

The CDF collaboration decided to base its entire silicon detector on rectangular silicon wafers that are assembled to form an open-ended barrel with five layers of silicon. A second construction, called the Intermediate Silicon Layer, will surround the barrel from the outside, separating it from the non-silicon detector components of the CDF experiment. In addition, the CDF scientists will insert a separate silicon assembly, called Layer 00, into the center of the 3-foot-long barrel. It will rest directly on the pipe that provides the vacuum environment for the Tevatron particle beams.

Sharon Austin inspecting F-disk assembly The DZero collaboration opted for both rectangular and wedge-shaped wafers. Similar to the CDF silicon detector, the rectangular pieces form the layers of a compact barrel. The wedge-shaped wafers are assembled into disks that look like pizzas with a shiny surface. These disks, large enough to cover the barrel openings, will sit perpendicular to the beam. A hole at their centers allows scientists to slide the disks right over the beam pipe.

The CDF and DZero scientists worked closely with four manufacturers to obtain silicon wafers that met the stringent experimental requirements. To determine the location of particle tracks to high precision, the silicon on the wafers is divided into microstrips, each less than one twentieth of a millimeter wide. By alternating the orientation of the strips and stacking them in layers, physicists obtain a precise three-dimensional map of the paths of all charged particles that cross the construction.

"To reduce the amount of material, the collaborations designed wafers with silicon strips on both sides," said Lenny Spiegel, associate leader of the SiDet facility. "The double-sided approach is more efficient, but handling is more difficult." Future projects, already underway at the SiDet facility, will return to using single-sided wafers.

Technicians, dressed in clean-room outfits, carefully inspect each silicon panel as they arrive from the manufacturer. To check the quality of the silicon microstrips, which are invisible to the naked eye, they use precision tools to measure the conductivity of all strips.

Designing the mechanical structure to support the wafers proved to be a crucial task. Physicists wanted to minimize the amount of material close to the collision area, since escaping particles would bounce off it. Yet the support structure needed to be stable enough to carry the weight and maintain the precise alignment of the wafers. These seemingly contradictory requirements resulted in the choice of thin carbon fiber rails as supports for the silicon wafers.

"It took a lot of prototyping," recalled Mary Morfin, who was involved in the rail production. To minimize the stress on the silicon wafers, which will rest on the rails, Morfin had to consistently produce rails that were "close to perfectly plane."

Gluing several silicon wafers next to each other on a single rail, technicians produced so-called ladders.

Andrew Foland (left) and Bert Gonzalez work on SVXII silicon wafers. "Aligning the silicon wafers to one another is a critical phase of the assembly process," said Bert Gonzalez, the lead technician for the construction of the CDF silicon barrel.

Gonzalez and his colleagues used precision measurement machines to determine the location of wafers to better than several thousandths of a millimeter.

"We got the Zeiss precision machines, which cost $2.7 million in the 1970s, for a fraction of the original cost," said Greg Sellberg, who was responsible for the production of parts of the DZero silicon detector. "Looking for new equipment, I found them on a government surplus list."

Technicians normally use coordinate-measuring machines to check the alignment of a construction after its assembly. Sellberg and his colleagues retrofitted and maintained the machines so that the actual assembly of ladders could take place on them.

To modify the machines, physicists had to ask the Zeiss company to share the blueprints, a top industrial secret. Half an hour after DZero scientists called the company's CEO, the requisite fax arrived at Fermilab.

Once the ladders were assembled, technicians needed to electrically connect the silicon wafers to electronic chips by a process called bonding. Peering through microscopes, Tammy Hawke and her coworkers operated machines that fused aluminum wires, one-thousandth of an inch in diameter, to chips and individual silicon strips, producing a total of three million bonds.

"It takes about six months of training to learn the bonding process," said Ron Lipton, a co-leader of the DZero silicon detector efforts. "It takes the right personality, patience and hands."

Cristian Gingu, inspects the bulkheads for the SVXII Detector The final assembly stage consisted of mounting the ladders inside a frame made of beryllium, a metal that is lighter but more stable than carbon. Attaching the ladders without breaking or scratching the silicon wafers required a steady hand. Only a few technicians had the skills to do the job.

"I had to use tweezers to place each screw, securing it with a nut," said Gonzalez, who installed all 180 ladders for the CDF silicon barrel. "I just couldn't afford to drop a screw inside the barrel since it could have done severe damage." Physicists expected that it would take half a day to mount and align a single ladder in the barrel. Gonzalez greatly exceeded all expectations and, by the end of the project, was installing up to 12 ladders a day.

Last week, SiDet workers put the finishing touches on the silicon detectors. As the devices get moved to the collision halls inside the Tevatron, a ten-year project comes to an end.

"It has been a great team effort," said Spalding. "The technicians and everyone else have done a fantastic job."

last modified 12/1/2000   email Fermilab