Creating Unique Research Facilities to Pursue the Newest Scientific Exploration

The LUX Project: Design Thinking a Mile Underground
  • Retrofitting a Gold Mine

    The mine’s walls of raw rock have been sprayed with shotcrete to keep them from crumbling. The walls and floors are all painted with epoxy to mitigate any naturally occurring radon.

    Photo courtesy of LEO A DALY

Most A&E teams will never have to plan the descent of a highly sensitive, one-of-a-kind particle accelerator a mile down a wet, dark, crooked shaft to an astrophysics research facility built in a decommissioned gold mine. Or collaborate on a strategy to acquire and store the equivalent of 20 percent of a year’s production of xenon gas without making a massive one-time purchase that could trigger a drastic spike in market prices. Or order equipment from around the globe that has to be transported by ship, rail, or truck because of the exposure to radiation in flight. The professionals who faced these challenges will probably not encounter them again on future projects. However, as scientific discovery continues to push the frontiers of the unknown, the need to create unique research environments is likely to become more frequent.

Based on the experience of the architects at LEO A DALY on the Large Underground Xenon and ZonEd Proportional Scintillation in Liquid Noble Gases (LUX) project in Lead, S.D., a different approach to the standard linear project delivery process will be indispensable. 

The methodology that has proven instrumental in organizing and managing the thousands of moving parts of the LUX project is the model of design thinking originated by IDEO and the Stanford school of design. 

Everything Is a Research Item       

LUX is a large underground experiment using liquified xenon in the search for weakly interacting massive particles, or WIMPs, which are theoretical particles thought to make up dark matter. The project site is the Sanford Underground Research Facility (SURF), in the Black Hills of South Dakota, in what was once North America’s largest gold mine, Homestake, which sprawls over 200 miles of underground caverns, at depths up to 7,000 feet. Selected for its low-radiation environment, LUX is housed an area 4,850 feet below ground, where the ambient temperature hovers around 95 degrees F.

The science behind the experiment had been worked on for roughly a decade prior to the architect’s engagement in 2015. The firm had previously completed a nuclear science project—Compact Accelerator System for Performing Astrophysical Research (CASPAR)—in the decommissioned mine, and that activity needed to continue uninterrupted during LUX construction.

While complex, CASPAR had followed the standard design timeline (represented graphically as an inverted triangle), starting at the narrow bottom tip with schematic design, to confirm work scope and budget; broadening to the longer phase of design development; and capped by an extensive period developing construction documents.

The uncertainties and unknowns required for LUX equipment and construction flipped the triangle.

“When we started LUX, we found that we were spending all of our time on schematic design,” says Steven Andersen, senior architect at LEO A DALY. “We had two defined goals, but we didn’t have a defined work scope. The path to get there was completely unknown. We had budget numbers, but we couldn’t actually set a budget. Estimators can’t really estimate the cost of a facility when they have never seen one before and did not know how the work would be done.”

Realizing that the conventional delivery method was inadequate to manage a project where “everything turned into a research item,” the design team stepped outside the A&E industry to borrow from the IDEO and Stanford design thinking approach. This model is expressed graphically as a circle divided into three phases: understand, explore, and materialize. Each of the three phases is subdivided into two parts, creating a total of six steps.

“We thought this would be much more adaptable because of the fluidity of these types of projects,” says Andersen. “Every piece becomes its own little design section and requires a design charrette—an intense design session—to get there.”


This first phase of design thinking is broken down into two steps, empathize and define. Andersen admits that “empathize,” taken directly from the Stanford model, may be an unusual word choice in the construction world, where terms like “partnering” and “shared goals” are much more common to describe client relationships.

The difference, he says, is that “partnering and shared goals mean that you and your stakeholders understand the goals when you start.”

For projects like LUX, where the science was being developed and engineered concurrent with the design process, it takes a significantly higher level of commitment—and investment—on the architect side.

“The designers needed to dive deep into the stakeholders’ world and be more in touch with their concerns, obstacles, motivations, and stresses,” he explains. “The stakeholders need a team willing to do research, iterate, and prototype, until solutions seem like they will work. You have to be willing to invest in the science and go really deep to help the stakeholders move the project along.” 

In the define step, it’s critical to establish a complete roadmap of the science workflows, in much the same way that designers lay out programs and work with adjacency diagrams. This detailed look forward allows the missing pieces to be identified at the outset instead of having them crop up along the way.

For example, well into the project, the team discovered that the radon detector had to sit on a base of ultra-pure titanium. Locating a source of the rare material and getting it on a schedule at the right moment was no mean feat.

“It would have been much easier if we had had an entire roadmap of all the different science workflows,” comments Andersen. “Then we would have been able to identify the unknowns and we would have managed that process better and more quickly.”

Another important element to define is the communications strategy. With more than 250 scientists from 38 institutions on the LUX team, it was essential to establish a consistent meeting format, with only one expert at a time, and adhere to a uniform style of minutes. A single project manager all the way through was also required to organize and structure the massive undertaking.


The two steps of the explore phase are ideate and iterate.

“Generate a range of crazy, creative ideas,” the design thinking model advises.

Many ostensibly simple tasks turned out to be anything but, confronting designers with a series of logistical challenges that demanded exploration from multiple perspectives—“behind, around, sideways, and in between,” literally and figuratively. Then the out-of-the-box ideas can be subject to further refinement. 

The issues surrounding a new cleanroom at surface level illustrate the need for fresh thinking. All the components of the detector had to be repackaged in the cleanroom’s radiation-free environment to protect them from contamination in transit down the mine. The designers discovered, however, that the already-procured radon-reduction system couldn’t produce enough radon-free air to charge the cleanroom.

“That was a major challenge for a while, until we went back to the performance goals of the project and said, ‘let’s rethink this whole process.’”

Further exploring the goals, the team arrived at a solution: By initially running the radon-reduction system for a full eight hours, it would go through enough cycles to produce the actual airflow and particle count acceptable for the work to start.

An extraordinary amount of time and iteration went into planning the installation of the cleanroom’s radon detector. The grated floor in one area had to be removed to lower the equipment in the designated pit. What was expected to be a straightforward task rapidly became complex, as considerations about height, depth, and the order of panel removal emerged, each taking intense examination.

“We are still amazed at how long this took,” says Andersen.

Similarly, an off-center hoistway beam presented several obstacles that the team had to sort through to determine how to pick up the detector and turn it to the optimal position for descent.

“This was one of those simple things that took lots of iteration to get right.” 


The final phase, materialize, begins with testing and culminates in implementation. However, because LUX was a unique project with so many custom solutions, the design team did not wait until the final phase to start testing individual elements, like how to pick up and flip something of similar weight to the detector to figure out exactly how to lower it down the shaft.

The non-linear nature of design thinking affords the flexibility to accommodate the fluidity of the developing science.  

“If you have an unknown, test it, at any stage of the project possible, even in what would normally be schematic design,” says Andersen. “This will yield more unknowns that will have to be defined later, but it’s better to get there sooner than later.”

Making mock-ups is another aspect of testing. For Lux, spatial mock-ups were done on site. Science mock-ups were done at the labs where they were being developed.

“Rapid, easy mock-ups become really effective for these one-off projects,” he notes. “I would highly recommend doing them early in the project to identify unknowns or issues.”

With summer 2020 the approximate date for equipment completion, the full-on implementation step looms ahead. Project design was completed in late 2017, and at this point all the infrastructure issues have been resolved. Systems are working as designed, and science equipment installation and assembly will start soon.

“It will take a lot of manpower to make these things happen,” says Andersen. “Assembly will be intense, and I’m sure that some professors and graduate students will be living in the cleanroom for a while.”

Changing Team Members But Not Budget

Underscoring the critical importance of the design team’s commitment and empathic attitude, Andersen also singles out early identification of the science workflow as key to effective project leadership. Additionally, he points out that team membership and rules changed and developed over LUX’s lengthy timeline. As the project progressed and the science evolved, the need for people with different skillsets emerged, especially for those who could smoothly shift their thinking from big concept to detail.

“At every phase or milestone date, we would reevaluate whether the team was performing at the level then needed, or if we should make adjustments, because the science is just so demanding on these types of projects.”

The budget was not nearly as malleable. Using as many off-the-shelf systems as possible, the team was constantly revisiting ways to get things done without adding to the cost—all while facing the limitations of working a mile underground. Hindrances ranged from not being able to retrieve items forgotten at the surface to the prohibition of welding and various chemicals.

“The critical piece in the budget was coming to an understanding with the stakeholders of a solution good enough for the project to achieve the desired performance. In some ways the budget issue always came back to truly understanding what the science infrastructure needed to achieve in order not to inhibit the science. Our solutions were simple, but getting there was always a process of defining, iterating, and prototyping until an affordable answer was found,” concludes Andersen.

By Nicole Zaro Stahl

Project Data
Completion Date: 
December, 2018
Published 12-12-2018