© 2012 Jonathan Ochshorn
From the Critique of Milstein Hall introduction: Milstein Hall at Cornell University, designed by Rem Koolhaas and OMA, is an interesting building, in some ways an amazing building, and, by virtually any conceivable objective criterion, a disaster. That something amazing can simultaneously be a disaster is hardly a paradox. In fact, disasters are often amazing, and our amazement often increases proportionally with the range and scope of the disaster.
I will not be criticizing the visual appearance of this building, or making judgments about its subjective, aesthetic merit. I personally find the building interesting, and its underlying formal rationale provocative and compelling. But I am not particularly qualified to render such judgments, and other authorities or connoisseurs of architectural taste may well disagree. What follows, instead, is an objective critique of Milstein Hall, looking at the building in some detail from a series of different points of view, none of which are driven by aesthetic considerations.
Cornell lists the "sustainable design initiatives" it has taken in the design and construction of Milstein Hall1 and summarizes these initiatives with the image reproduced in Figure 1. These initiatives are grouped into eight categories, with my comments following each specious claim.
Comments: Cornell's cogeneration and lake-source cooling initiatives, however commendable, are not design initiatives of Milstein Hall. Chilled beams are relatively efficient, but hardly radical. Insulated walls and glazing to reduce building air loss? This both makes no sense and is inaccurate. It makes no sense because "air loss" (infiltration) is reduced by designing and installing a continuous air barrier system for the building, not by providing "insulation." Milstein Hall actually performs extremely poorly on both counts (air barriers and insulation). First, the building has a relatively ineffective air barrier system. Especially at seismic building separation joints along the entire perimeter between Milstein and Rand/Sibley Halls, not only has air barrier continuity not been established, but insulation has not been installed with seismic joints detailed to accommodate movement. Second, rather than being a role model for building insulation, Milstein Hall is actually a case-study in inefficient thermal form and thermal bridging.
Purely from an energy-efficiency and insulation standpoint, the most logical geometry for a building is one that minimizes surface area. Milstein Hall does exactly the opposite, by creating a large, extended floor plate that is then elevated above the ground, exposing not only its roof but also its underside to the exterior. Below-grade spaces also extend well beyond the boundaries of the upper floor plates, so that they too are needlessly exposed to the exterior. The entire wall area of the building, excluding below-grade spaces, is glazed. Of course, insulated glazing is better than uninsulated glazing, but this misses the point: all glazing, unless designed as part of a passive-solar system, is less thermally efficient than an insulated wall. Milstein's undifferentiated glazing (all vertical surfaces, whether facing north, south, east, or west, are glazed) has not been designed in this way and so only contributes to gratuitous heat gain or heat loss. The contribution of the glass to daylighting is certainly real, but in no way compensates for the increased energy usage for heating and cooling. Given an already tenuous thermal-design strategy consisting primarily of undifferentiated glazing for all four facades, the building is then detailed with substantial thermal bridges, which create additional express pathways for heat loss. For example, the unusually large steel columns that penetrate the studio floor from the exterior act as conduits for both heat gain during the cooling season and heat loss during the heating season, causing an increase in heat loss through the floors of approximately 10%.2 Stone veneer panels are fastened to the structure with horizontal steel angles that, like the steel columns, also penetrate the insulation layer, thereby greatly reducing its effectiveness (see Figure 2).
Employing a vegetated roof does not reduce solar heat gain nor does it reduce building air loss, as claimed by Cornell, if such a system is compared to a well-detailed but otherwise ordinary insulated roof with a reflective surface. As Joseph Lstiburek astutely points out: "Grass and dirt are not energy efficient. Work with me here. Which saves more energy—2 inches of dirt or 2 inches of insulation? Which saves more energy—grass or a white colored membrane? Which is more expensive and does not save energy—grass and dirt or insulation and a white colored membrane?"3
Comments: These claims mirror some of the LEED credits in "sustainable sites" and "materials & resources." Milstein Hall's location next to existing bus lines made it impossible not to tie into a pubic transport network—this "initiative" has nothing to do with the design of Milstein Hall. As described above, Milstein Hall—using the minimum bike storage standards of the LEED guidelines instead of actually responding to the needs of bike users—does an extremely poor job of accommodating them. As to the "sustainability initiative" accomplished by accommodating "pedestrian access," one is at a loss to imagine what this could possibly mean. Is it that the building has a door at ground level, thereby permitting pedestrians to enter? Or that Cornell's existing system of walks and paths is not separated from the entrance to Milstein Hall by some sort of moat or electronic barrier?
Comments: Daylighting, in the form of continuous perimeter glazing and skylights, can only be considered a sustainable (i.e., energy-saving) design feature if it reduces the need for electric lighting. On Milstein Hall's large studio floor, electric lighting is never turned off, so that no energy saving can be attributed to its daylighting sources. In fact, both the day and night lighting conditions have been criticized by users of the space.4 Energy-efficient light fixtures are, of course, better than, say, incandescent fixtures, but using energy-efficient fixtures inefficiently—as is being done in Milstein Hall—should not be characterized as "sustainable."
Comments: As described elsewhere, Milstein Hall uses steel not just inefficiently, but extraordinarily inefficiently, with more steel per square foot of floor area than in some 100-story buildings. That this steel has recycled content does not make such an incredibly inefficient design sustainable, especially when the basis of this recycled steel—junked cars—is the disposable culture of planned obsolescence. Milstein's concrete recycles fly ash as part of its cementitious content; whether recycled "concrete aggregate" is also used is unlikely, but possible. The claim that Milstein Hall's finishes reduce material use is puzzling, since one can always imagine a design that has either more, or less, material content in its finished surfaces. Milstein Hall, for example, has concrete floors, and does not have carpet or tile on these floor surfaces. The concrete surface seems perfectly adequate for its intended use. If not using an additional material that is not necessary is counted as "sustainable," then the bar for sustainable design has been set pretty low.
Comments: Rather than renovating Sibley and Rand Halls, and providing a small amount of added program space to accommodate increased program needs,5 the design for Milstein Hall created 50,000 square feet of new space, very little of which is actually needed. This has become clear now that the new building is occupied: virtually the entire 2nd and 3rd floors of E. Sibley Hall (excluding about 33% of the 2nd floor, still being used for faculty offices and other program needs) along with the large space and mezzanine under Sibley's dome—in other words, the entire space formerly housing the Fine Arts Library—is currently vacant. This 20,000 sq.ft. of potentially useful space, suitable for studios or offices (in fact, these now-vacant spaces in E. Sibley Hall formerly housed much of the architecture studios before the Fine Arts Library expanded into those spaces during the 1970s), is more or less the same size as the 25,500 sq.ft. studio floor in Milstein Hall. All of this raises the question of exactly what urgent programming imperatives caused the University to spend in excess of $50 million—much of it improperly taken from a departmental endowment meant for program support rather than construction—for a major new building when a more modest allocation of funds to renovate (not just to "reuse") existing buildings and perhaps add a small addition to those buildings would have accomplished more at less expense. In other words, the claim that Milstein Hall somehow allowed for the "reuse" of Rand and Sibley Halls is disingenuous: Sibley and Rand Halls remain in terrible shape—they both are completely inadequate in terms of energy and IEQ—because no funds from the Milstein Hall project were permitted to be used to improve them, other than what was minimally necessary to comply with building code provisions triggered by Milstein Hall's construction.
It is also sometimes claimed that Milstein Hall's design "saved" Rand Hall from demolition.6 This, too, is spurious. Rand Hall was slated for demolition when Milstein Hall was the subject of a design competition in 2000. In the same way that the University made the (bad) decision to demolish Rand in 2000, it then reversed the decision at a later date. What "saved" Rand Hall was not the design of Milstein Hall itself, but rather a change in the initial plan to demolish the building. If the University had not unilaterally made the bad decision to demolish Rand Hall in 2000, the building would never have needed to be "saved." In any case, it was the University's decision, not the design of Milstein Hall, which "saved" the building.
It is claimed by Cornell that Milstein is a flexible building. That Milstein Hall is the opposite of a flexible building will be discussed elsewhere.
Comments: All three of these claims are at least partly incorrect. First, Milstein Hall's vegetated roof may or may not be useful in filtering storm water. Some studies have measured increased amounts of nitrogen and phosphorus in green-roof runoff compared with conventional roof runoff during heavy rainfall.7 Second, Milstein Hall meets neither the quantity nor quality stormwater standards for LEED credit. Instead, virtually all storm water falling on the vegetated roof during heavy rainfall is directed through the building and into the storm sewer system, rather than being controlled or improved on site. Third, Milstein Hall's green roof has no native plants. The sedums planted on the roof are adapted plants, not native species.8 Using adapted, non-invasive, plants is not bad. It just isn't accurate to call them native. It is also more than a bit hypocritical of Cornell to boast about eliminating pesticide usage on this small vegetated roof, while simultaneously employing pesticides (the broadleaf herbicide SpeedZone9) over large parts of its grounds, including the arts quad adjacent to Milstein Hall (Figure 3).
Comments: This is simply reaching for the low-hanging fruit. Not using irrigation in Ithaca, NY is hardly a sustainable accomplishment, as it rains here quite a bit.
Comments: There is nothing radically sustainable about chilled beams and radiant slabs. They provide no individual comfort controls, so that individual variations in the experience of comfort cannot be accommodated. Daylighting, entering through wall-to-ceiling glazing and skylights, has already been described as unnecessary (since the electric lights are always on) and often counter-productive (causing both glare and unwanted illumination). Milstein Hall does not consistently eliminate products with high VOC content. While it gains a LEED point for using a small amount of "Green Label Plus" carpet in the auditorium, it still uses composite wood products indoors that do not satisfy the LEED criteria for indoor air quality. Milstein Hall provides outside air, as do all buildings, both old and new. This is a requirement of building and mechanical codes, not a sustainable design initiative.
As to Milstein Hall's alleged visual and direct connections to natural areas, one simply needs to walk through the 2nd-floor studio to form a more accurate impression: to the east is a parking lot, admittedly with some trees visible on the edge of Fall Creek gorge; to the north is the asphalt roof of the Foundry, which blocks any view of Fall Creek; to the west is Rand Hall; and to the south is E. Sibley Hall, along with a view towards other campus buildings. The floor plate is so deep that most work stations are located far from Milstein's glazed edges, and have even less of a chance to connect with nature. There are certainly no direct connections to natural areas from Milstein Hall, which is separated from Fall Creek (the only plausible "natural area" in the vicinity) by University Avenue and the Foundry. In fact, what Milstein Hall did was to eliminate numerous windows and outdoor views from Rand and Sibley Halls.<< previous | next >>
1 "Milstein Hall and Sustainability," AAP/ Architecture Art Planning, online here (originally accessed 7 August 2011; no longer available. Relinked 7/7/20).
2 Such thermal bridging (defined as highly conductive material interrupting the continuity of thermal insulation) is not inconsequential. Even without a sophisticated thermal analysis, one can make a rough estimate of the energy penalty by comparing the heat loss with studio floor column penetrations to the heat loss through an insulated floor without column penetrations.
As can be seen in Figures 4 and 5, the 14 exterior columns, each with a cross-sectional area of 178 square inches (based on an assumed W14x605 column section, as shown in Milstein Hall's structural drawings), create a total uninsulated area penetrating the studio floor of 14 × 178 = 2492 square inches or 17.3 square feet. These large column sizes do not literally penetrate the floor, as they are welded into girders or into the bottom chords of story-height trusses with variable cross-sectional dimensions. Assuming that the actual area of steel penetrating the insulation layer is 2/3 that of the large columns, we can assume a per-penetration area of 2/3 × 17.3 = 11.5 square feet. The insulated 2nd-floor area (total area minus the portion of the floor plate over insulated space) is approximately 25,500 - 5,685 = 19,815 square feet. Subtracting the column area, the exterior insulated floor area is 19,815 - 11.5 = 19,804 square feet. The heat loss due to these two elements is found by multiplying their areas by their U-values and by an assumed temperature differential between outdoors and indoors of, say, 70 degrees F. Assuming an R-value for the floor of 40 and for the steel of about 0.22 (based on an R-value for steel of 0.003 per inch and assuming an average curved trajectory length from outside to inside of 72 in.), the U-values for the floor and steel are, respectively, 1/40 = 0.025 and 1/0.22 = 4.545. Using these values, the heat loss for the two elements is as follows:
Floor: 0.025 × 19,804 × 70 = 34,657 BTU/hr.
Columns: 4.545 × 11.3 × 70 = 3,595 BTU/hr.
The total heat loss through the floor, found by adding these two components, is 34,657 + 3,595 = 38,252 BTU/hr.
Without these columns acting as thermal bridges, the heat loss through the floor would be 0.025 × 19,815 × 70 = 34,676 BTU/hr. The difference in total heat loss caused by the thermal bridging of the columns is 38,252—34,676 = 3,576 BTU/hr.
Remarkably, even though the column thermal bridges constitute only 11.5 square feet out of a total exterior floor area of 19,815 square feet, or just 0.06% of the floor area, their high conductivity results in an increased heat loss through the floor of 3,576 × 100 / 34,676 = 10.3%. Equally remarkable, all of this heat loss through the floor—the entire 38,252 BTU/hr, equivalent to the heat loss generated by a 24,000 sq.ft. well-insulated house—could have been avoided by designing Milstein Hall without floor areas exposed to the exterior. House heat loss assumptions are based on the Home Heat Loss Calculator found online here (accessed 7 Dec. 2011) assuming 2 floors, each 30 ft. x 40 ft. with 9-ft. floor-to-floor height, R-38 ceilings, R-24 walls, 500 sq.ft. of R-3.2 low-e insulated glass windows, and so on. Such a house—and therefore also the energy wasted solely by the exposed-floor of Milstein Hall—also contributes 12,216 lb of greenhouse gases (CO2) to the atmosphere each year, according to the calculator.
It is Milstein Hall's formal design concept that is fundamentally inefficient with respect to energy use; the thermally-penetrating columns take an already inefficient idea and only make it worse.
Additional gratuitous thermal bridging occurs over the below-grade spaces to the west of the first and second floors that support a plaza next to on-grade parking. Metal bollards penetrate through layers of concrete and rigid insulation, providing a classic thermal bridge to the spaces below.
3 Joseph Lstiburek, "BSI-007: Prioritizing Green—It's the Energy Stupid" online here (accessed 6 Dec. 2011).
4 The following email written to the department chair by a graduate architecture student September 1, 2011—shortly after Milstein Hall was occupied—was reproduced and displayed at the "OMA/Progress" exhibition at the Barbican Gallery in London:
"As you know, I am interested in the lighting design in Milstein, both as a user of the building and as a lighting designer. The arch. department may not be aware that the building has already become a teaching tool: students are witnessing a lighting system (that affects us day and night) that some believe was an [sic] overlooked from a sustainable design perspective. In our Environmental Systems II class, a third year undergrad shared their observation that we have moved into a supposedly sustainable building yet the lights are constantly on, even when there is adequate daylight delivered to the space via skylights during the daytime.
"I have measured the illuminance at my desk and the daylight level is around 250 fc and the night reading is 55 fc. The night-time level is excessive for a space where the students are primarily using computers. The human eye is adapted to deal with natural light and its dynamic nature, so the daylight level does not concern me,. People will put up with a lot of light as long as there is not uncomfortable glare. However, shadowless, even lighting at night to an excessive level can cause eye strain, especially when one is looking at a computer screen. The IES (Illuminating Engineering Society) currently recommends a range of 15-25 for office spaces with a separately controlled task light for user comfort.
"Sorry to seem like such a pest on this issue but I thought you should know that I am not the only one that is aware of the lighting and some of the BArch students seem to be getting cynical about the dept's stance on sustainability (wasting energy = wasting money)."
5 I have written about other more rational options to the OMA Milstein Hall proposal in "Thoughts on Milstein Hall," first posted February 11, 2011, here (accessed 7 Dec. 2011):
"Milstein Hall does virtually nothing to address the conditions in Sibley and Rand Hall that were specifically criticized in previous accreditation reports. Accessibility is finally being addressed in Sibley independently of the Milstein Hall project. An elevator and accessible bathrooms are proposed for Rand Hall as part of the Milstein Hall project, but such interventions could happen with or without a new building.
"Space, while not initially the most important NAAB concern, has certainly become a bigger issue recently, with the elimination of departmental space due to the expansion of administrative offices, computer rooms, the elimination of trailers, the addition of elevators, the growth of professional and other graduate programs, etc. However, additional program space, to the extent that it is needed, can be more rationally obtained by moving the Fine Arts Library out of E. Sibley Hall and from under the Dome; by changing the Dome space back into a major auditorium accessible from the Arts Quad; and by building a straight-forward, cost-effective, and sustainable addition to house a new library and perhaps some additional program space in a location that does not immediately trigger serious building code issues."
6 See, for example, "Arch Profs Ardently Support Building Milstein" in the Cornell Daily Sun, Feb. 11, 2009 (accessed 7 Dec. 2011; no longer available): "This is a building whose design leaves intact Rand Hall, whereas all previous schemes that have been developed for this project have proposed tearing down this perfectly functional building before then rebuilding its square footage—surely in a true accounting of ecological impact this fact alone would be worth several additional levels of certification if the Leadership in Energy and Environmental Design (LEED) objectives were more broadly considered."
7 Brett Long, Shirley E. Clark, et. al., "Green Roofs: Optimizing the Water Quality of Rooftop Runoff" (PDF accessed 7 Dec. 2011): "Another study conducted in Estonia investigated the water quality of a lightweight aggregate and humus green roof runoff compared a bituminous membrane roof found that during light to moderate rainfall events the concentrations of COD, BOD, total nitrogen, and total phosphorus were greater in the bituminous roof. However during heavy rainfalls greater amounts of nitrogen and phosphorus washed from the green roof (Teemusk, 2007)."
8 Information on Milstein Hall's sedums was provided by Marguerite Wells of MotherPlants, a nursery in upstate New York specializing in growing plants for green roofs (including Milstein Hall's roof).
9 For information on SpeedZone, manufactured by PBI/Gordon Corporation, see the MSDS here (PDF originally accessed 8 Dec. 2011 but not available at original link; new MSDS PDF found 7/7/20). According to the Cornell Grounds Department's "Mission and Scope of Services" online here (originally accessed 8 Dec. 2011, not available at original link; newly linked 7/7/20): "Weed controls (herbicides) are kept to an absolute minimum and are applied on a limited basis. Many lawns will have varying populations of broad leaf and grass weed species present." According to Kevin McGraw, Landscape Manager at Cornell (phone conversation 8 Dec. 2011), herbicide application may change in Spring 2012, utilizing Battleship Herbicide III, manufactured for the Helena Chemical Company. Its MSDS can be found here (PDF originally accessed 8 Dec. 2013 but not available at original link; new MSDS PDF found 7/7/20). [Update Dec. 2014] Cornell's herbicide-du-jour is now Triamine (see PDF), applied by TruGreen as shown in Figure 6.
First posted 18 July 2012. Last updated: 11 December 2014