When hijacked airliners crashed into the tall Towers of the World Trade Center, in New York City [on 11 September 2001], each injected a burning cloud of aviation fuel throughout the 6 levels (WTC 2) to 8 levels (WTC 1) in the impact zone. The burning fuel ignited the office furnishings: desks, chairs, shelving, carpeting, work-space partitions, wall and ceiling panels; as well as paper and plastic of various kinds.
How did these fires progress? How much heat could they produce? Was this heat enough to seriously weaken the steel framework? How did this heat affect the metal in the rubble piles in the weeks and months after the collapse? This report is motivated by these questions, and it will draw ideas from thermal physics and chemistry. My previous report on the collapses of the WTC Towers described the role of mechanical forces (1).
Summary of National Institute of Technology and Standards (NIST) Report
Basic facts about the WTC fires of 9/11/01 are abstracted by the numerical quantities tabulated here.
Table 1, Time and Energy of WTC Fires
ITEM WTC 1 WTC 2
impact time (a.m.) 8:46:30 9:02:59
collapse (a.m.) 10:28:22 9:58:59
time difference 1:41:52 0:56:00
impact zone levels 92-99 78-83
levels in upper block 11 27
heat rate (40 minutes) 2 GW 1 GW
total heat energy 8000 GJ 3000 GJ
Tower 1 stood for one hour and forty-two minutes after being struck between levels 92 and 99 by an airplane; the block above the impact zone had 11 levels. During the first 40 minutes of this time, fires raged with an average heat release rate of 2 GW (GW = giga watts = 10^9 watts), and the total heat energy released during the interval between airplane impact and building collapse was 8000 GJ (GJ = giga-joules = 10^9 joules).
A joule is a unit of energy; a watt is a unit of power; and one watt equals an energy delivery rate of one joule per second.
Tower 2 stood for fifty-six minutes after being struck between levels 78 and 83, isolating an upper block of 27 levels. The fires burned at a rate near 1 GW for forty minutes, diminishing later; and a total of 3000 GJ of heat energy was released by the time of collapse.
WTC 2 received half as much thermal energy during the first 40 minutes after impact, had just over twice the upper block mass, and fell within half the time than was observed for WTC 1. It would seem that WTC 1 stood longer despite receiving more thermal energy because its upper block was less massive.
The data in Table 1 are taken from the executive summary of the fire safety investigation by NIST (2).
The NIST work combined materials and heat transfer lab experiments, full-scale tests (wouldn’t you like to burn up office cubicles?), and computer simulations to arrive at the history and spatial distribution of the burning. From this, the thermal histories of all the metal supports in the impact zone were calculated (NIST is very thorough), which in turn were used as inputs to the calculations of stress history for each support. Parts of the structure that were damaged or missing because of the airplane collision were accounted for, as was the introduction of combustible mass by the airplane.
Steel loses strength with heat. For the types of steel used in the WTC Towers (plain carbon, and vanadium steels) the trend is as follows, relative to 100% strength at habitable temperatures.
Table 2, Fractional Strength of Steel at Temperature
Temperature, degrees C Fractional Strength, %
I use C for Centigrade, F for Fahrenheit, and do not use the degree symbol in this report.
The fires heated the atmosphere in the impact zone (a mixture of gases and smoke) to temperatures as high as 1100 C (2000 F). However, there was a wide variation of gas temperature with location and over time because of the migration of the fires toward new sources of fuel, a complicated and irregular interior geometry, and changes of ventilation over time (e.g., more windows breaking). Early after the impact, a floor might have some areas at habitable temperatures, and other areas as hot as the burning jet fuel, 1100 C. Later on, after the structure had absorbed heat, the gas temperature would vary over a narrower range, approximately 200 C to 700 C away from centers of active burning.
As can be seen from Table 2, steel loses half its strength when heated to about 570 C (1060 F), and nearly all once past 700 C (1300 F). Thus, the structure of the impact zone, with a temperature that varies between 200 C and 700 C near the time of collapse, will only have between 20% to 86% of its original strength at any location.
The steel frames of the WTC Towers were coated with “sprayed fire resistant materials” (SFRMs, or simply “thermal insulation”). A key finding of the NIST Investigation was that the thermal insulation coatings were applied unevenly — even missing in spots — during the construction of the buildings, and — fatally — that parts of the coatings were knocked off by the jolt of the airplane collisions.
Spraying the lumpy gummy insulation mixture evenly onto a web of structural steel, assuming it all dries properly and none is banged off while work proceeds at a gigantic construction site over the course of several years, is an unrealistic expectation. Perhaps this will change, as a “lesson learned” from the disaster. The fatal element in the WTC Towers story is that enough of the thermal insulation was banged off the steel frames by the airplane jolts to allow parts of frames to heat up to 700 C. I estimate the jolts at 136 times the force of gravity at WTC 1, and 204 at WTC 2.
The pivotal conclusion of the NIST fire safety investigation is perhaps best shown on page 32, in Chapter 3 of Volume 5G of the Final Report (NIST NCSTAR 1-5G WTC Investigation), which includes a graph from which I extracted the data in Table 2, and states the following two paragraphs. (The NIST authors use the phrase “critical temperature” for any value above about 570 C, when steel is below half strength.)
“As the insulation thickness decreases from 1 1/8 in. to 1/2 in., the columns heat up quicker when subjected to a constant radiative flux. At 1/2 in. the column takes approximately 7,250 s (2 hours) to reach a critical temperature of 700 C with a gas temperature of 1,100 C. If the column is completely bare (no fireproofing) then its temperature increases very rapidly, and the critical temperature is reached within 350 s. For a bare column, the time to reach a critical temperature of 700 C ranges between 350 to 2,000 s.
“It is noted that the time to reach critical temperature for bare columns is less than the one hour period during which the buildings withstood intense fires. Core columns that have their fireproofing intact cannot reach a critical temperature of 600 C during the 1 or 1 1/2 hour period. (Note that WTC 1 collapsed in approximately 1 1/2 hour, while WTC 2 collapsed in approximately 1 hour). This implies that if the core columns played a role in the final collapse, some fireproofing damage would be required to result in thermal degradation of its strength.” (3)
Airplane impact sheared columns along one face and at the building’s core. Within minutes, the upper block had transferred a portion of its weight from central columns in the impact zone, across a lateral support at the building crown called the “hat truss,” and down onto the three intact outer faces. Over the course of the next 56 minutes (WTC 2) and 102 minutes (WTC 1) the fires in the impact zone would weaken the remaining central columns, and this steadily increased the downward force exerted on the intact faces. The heat-weakened frames of the floors sagged, and this bowed the exterior columns inward at the levels of the impact zone. Because of the asymmetry of the damage, one of the three intact faces took up much of the mounting load. Eventually, it buckled inward and the upper block fell. (1)
Now, let’s explore heat further.
How Big Were These Fires?
I will approximate the size of a level (1 story) in each of the WTC Towers as a volume of 16,080 m^3 with an area of 4020 m^2 and a height of 4 m (4). Table 3 shows several ways of describing the total thermal energy released by the fires.
Table 3, Magnitude of Thermal Energy in Equivalent Weight of TNT
ITEM WTC 1 WTC 2
energy (Q) 8000 GJ 3000 GJ
# levels 8 6
tons of TNT 1912 717
tons/level 239 120
lb/level 478,000 239,000
kg/m^2 (impact floors) 54 27
lb/ft^2 (impact floors) 11 6
The fires in WTC 1 released an energy equal to that of an explosion of 1.9 kilotons of TNT; the energy equivalent for WTC 2 is 717 tons. Obviously, an explosion occurs in a fraction of a second while the fires lasted an hour or more, so the rates of energy release were vastly different. Even so, this comparison may sharpen the realization that these fires could weaken the framework of the buildings significantly.
How Hot Did The Buildings Become?
Let us pretend that the framework of the building is made of “ironcrete,” a fictitious mixture of 72% iron and 28% concrete. This framework takes up 5.4% of the volume of the building, the other 94.6% being air. We assume that everything else in the building is combustible or an inert material, and the combined mass and volume of these are insignificant compared to the mass and volume of ironcrete. I arrived at these numbers by estimating volumes and cross sectional areas of metal and concrete in walls and floors in the WTC Towers.
The space between floors is under 4 meters; and the floors include a layer of concrete about 1/10 meter thick. The building’s horizontal cross-section was a 63.4 meter square. Thus, the gap between floors was nearly 1/10 of the distance from the center of the building to its periphery. Heat radiated by fires was more likely to become trapped between floors, and stored within the concrete floor pans, than it was to radiate through the windows or be carried out through broken windows by the flow of heated air. We can estimate a temperature of the framework, assuming that all the heat became stored in it.
The amount of heat that can be stored in a given amount of matter is a property specific to each material, and is called heat capacity. The ironcrete mixture would have a volumetric heat capacity of Cv = 2.8*10^6 joules/(Centigrade*m^3); (* = multiply). In the real buildings, the large area of the concrete pads would absorb the heat from the fires and hold it, since concrete conducts heat very poorly. The effect is to bath the metal frame with heat as if it were in an oven or kiln. Ironcrete is my homogenization of materials to simplify this numerical example.
The quantity of heat energy Q absorbed within a volume V of material with a volumetric heat capacity Cv, whose temperature is raised by an amount dT (for “delta-T,” a temperature difference) is Q = Cv*V*dT. We can solve for dT. Here, V = (870 m^3)*(# levels); also dT(1) corresponds to WTC 1, and dT(2) corresponds to WTC 2.
dT(1) = (8 x 10^12)/[(2.8 x 10^6)*(870)*8] = 410 C,
dT(2) = (3 x 10^12)/[(2.8 x 10^6)*(870)*6] = 205 C.
Our simple model gives a reasonable estimate of an average frame temperature in the impact zone. The key parameter is Q (for each building). NIST spent considerable effort to arrive at the Q values shown in Table 3 (3). Our model gives a dT comparable to the NIST results because both calculations deposit the same energy into about the same amount of matter. Obviously, the NIST work accounts for all the details, which is necessary to arrive at temperatures and stresses that are specific to every location over the course of time. Our equation of heat balance Q = Cv*V*dT is an example of the conservation of energy, a fundamental principle of physics.
Well, Can The Heat Weaken The Steel Enough?
On this, one either believes or one doesn’t believe. Our simple example shows that the fires could heat the frames into the temperature range NIST calculates. It seems entirely reasonable that steel in areas of active and frequent burning would experience greater heating than the averages estimated here, so hotspots of 600 C to 700 C seem completely believable. Also, the data for WTC Towers steel strength at elevated temperatures is not in dispute. I believe NIST; answer: yes.
Let us follow time through a sequence of thermal events.
The airplanes hurtling into the buildings with speeds of at least 200 m/s (450 mph) fragmented into exploding torrents of burning fuel, aluminum and plastic. Sparks generated from the airframe by metal fracture and impact friction ignited the mixture of fuel vapor and air. This explosion blew out windows and billowed burning fuel vapor and spray throughout the floors of the impact zone, and along the stairwells and elevator shafts at the center of the building; burning liquid fuel poured down the central shafts. Burning vapor, bulk liquid and droplets ignited most of what they splattered upon. The intense infrared radiation given off by the 1100 C (2000 F) flames quickly ignited nearby combustibles, such as paper and vinyl folders. Within a fraction of a second, the high pressure of the detonation wave had passed, and a rush of fresh air was sucked in through window openings and the impact gash, sliding along the tops of the floors toward the centers of intense burning.
Hot exhaust gases: carbon monoxide (CO), carbon dioxide (CO2), water vapor (H2O), soot (carbon particles), unburned hydrocarbons (combinations with C and H), oxides of nitrogen (NOx), and particles of pulverized solids vented up stairwells and elevator shafts, and formed thick hot layers underneath floors, heating them while slowly edging toward the openings along the building faces. Within minutes, the aviation fuel was largely burned off, and the oxygen in the impact zone depleted.
Fires raged throughout the impact zone in an irregular pattern dictated by the interplay of the blast wave with the distribution of matter. Some areas had intense heating (1100 C), while others might still be habitable (20 C). The pace of burning was regulated by the area available for venting the hot exhaust gases, and the area available for the entry of fresh air. Smoke was cleared from the impact gash by air entering as the cycle of flow was established. The fires were now fueled by the contents of the buildings.
Geometrically, the cement floors had large areas and were closely spaced. They intercepted most of the infrared radiation emitted in the voids between them, and they absorbed heat (by conduction) from the slowly moving (“ventilation limited”) layer of hot gases underneath each of them. Concrete conducts heat poorly, but can hold a great deal of it. The metal reinforcing bars within concrete, as well as the metal plate underneath the concrete pad of each WTC Towers floor structure, would tend to even out the temperature distribution gradually.
This process of “preheating the oven” would slowly raise the average temperature in the impact zone while narrowing the range of extremes in temperature. Within half an hour, heat had penetrated to the interior of the concrete, and the temperature everywhere in the impact zone was between 200 C and 700 C, away from sites of active burning.
Thermal Decomposition — “Cracking”
Fire moved through the impact zone by finding new sources of fuel, and burning at a rate limited by the ventilation, which changed over time.
Heat within the impact zone “cracks” plastic into a sequence of decreasingly volatile hydrocarbons, similar to the way heat separates out an array of hydrocarbon fuels in the refining of crude oil. As plastic absorbs heat and begins to decompose, it emits hydrocarbon vapors. These may flare if oxygen is available and their ignition temperatures are reached. Also, plumes of mixed hydrocarbon vapor and oxygen may detonate. So, a random series of small explosions might occur during the course of a large fire.
Plastics not designed for use in high temperature may resemble soft oily tar when heated to 400 C. The oil in turn might release vapors of ethane, ethylene, benzene and methane (there are many hydrocarbons) as the temperature climbs further. All these products might begin to burn as the cracking progresses, because oxygen is present and sources of ignition (hotspots, burning embers, infrared radiation) are nearby. Soot is the solid end result of the sequential volatilization and burning of hydrocarbons from plastic. Well over 90% of the thermal energy released in the WTC Towers came from burning the normal contents of the impact zones.
Aluminum alloys melt at temperatures between 475 C and 640 C, and molten aluminum was observed pouring out of WTC 2 (5). Most of the aluminum in the impact zone was from the fragmented airframe; but many office machines and furniture items can have aluminum parts, as can moldings, fixtures, tubing and window frames. The temperatures in the WTC Towers fires were too low to vaporize aluminum; however, the forces of impact and explosion could have broken some of the aluminum into small granules and powder. Chemical reactions with hydrocarbon or water vapors might have occurred on the surfaces of freshly granulated hot aluminum.
The most likely product of aluminum burning is aluminum oxide (Al2O3, “alumina”). Because of the tight chemical bonding between the two aluminum atoms and three oxygen atoms in alumina, the compound is very stable and quite heat resistant, melting at 2054 C and boiling at about 3000 C. The affinity of aluminum for oxygen is such that with enough heat it can “burn” to alumina when combined with water, releasing hydrogen gas from the water,
2*Al + 3*H2O + heat -> Al2O3 + 3*H2.
Water is introduced into the impact zone through the severed plumbing at the building core, moisture from the outside air, and it is “cracked” out of the gypsum wall panels and to a lesser extent from concrete (the last two are both hydrated solids). Water poured on an aluminum fire can be “fuel to the flame.”
When a mixture of aluminum powder and iron oxide powder is ignited, it burns to iron and aluminum oxide,
Al + Fe2O3 + ignition -> Al2O3 + Fe.
This is thermite. The reaction produces a temperature that can melt steel (above 1500 C, 2800 F). The rate of burning is governed by the pace of heat diffusion from the hot reaction zone into the unheated powder mixture. Granules must absorb sufficient heat to arrive at the ignition temperature of the process. The ignition temperature of a quiescent powder of aluminum is 585 C. The ignition temperatures of a variety of dusts were found to be between 315 C and 900 C, by scientists developing solid rocket motors. Burning thermite is not an accelerating chain reaction (“explosion”), it is a “sparkler.” My favorite reference to thermite is in the early 1950s motion picture, “The Thing.”
Did patches of thermite form naturally, by chance, in the WTC Towers fires? Could there really have been small bits of melted steel in the debris as a result? Could there have been “thermite residues” on pieces of steel dug out of the debris months later? Maybe, but none of this leads to a conspiracy. If the post-mortem “thermite signature” suggested that a mass of thermite comparable to the quantities shown in Table 3 was involved, then further investigation would be reasonable. The first task of such an investigation would be to produce a “chemical kinetics” model of the oxidation of the fragmented aluminum airframe, in some degree of contact to the steel framing, in the hot atmosphere of hydrocarbon fires in the impact zone. Once Nature had been eliminated as a suspect, one could proceed to consider Human Malevolence.
Nature is endlessly creative. The deeper we explore, the more questions we come to realize.
Steel columns along a building face, heated to between 200 C and 700 C, were increasingly compressed and twisted into a sharpening bend. With increasing load and decreasing strength over the course of an hour or more, the material became unable to rebound elastically, had the load been released. The steel entered the range of plastic deformation, it could still be stretched through a bend, but like taffy it would take on a permanent set. Eventually, it snapped.
Months later, when this section of steel would be dug out of the rubble pile, would the breaks have the fluid look of a drawn out taffy, or perhaps “melted” steel now frozen in time? Or, would these be clean breaks, as edge glass fragments; or perhaps rough, granular breaks as through concrete?
The basements of the WTC Towers included car parks. After the buildings collapsed, it is possible that gasoline fires broke out, adding to the heat of the rubble. We can imagine many of the effects already described, to have occurred in hot pockets within the rubble pile. Water percolating down from that sprayed by the Fire Department might carry air down also, and act as an oxidizing agent.
The tight packing of the debris from the building, and the randomization of its materials would produce a haphazard and porous form of ironcrete aggregate: chunks of steel mixed with broken and pulverized concrete, with dust-, moisture-, and fume-filled gaps. Like a pyramid of barbecue briquettes, the high heat capacity and low thermal conductivity of the rubble pile would efficiently retain its heat.
Did small hunks of steel melt in rubble hot spots that had just the right mix of chemicals and heat? Probably unlikely, but certainly possible.
Pulverized concrete would include that from the impact zone, which may have had part of its water driven off by the heat. If so, such dust would be a desiccating substance (as is Portland cement prior to use; concrete is mixed sand, cement and water). Part of the chronic breathing disorders experienced by many people exposed to the atmosphere at the World Trade Center during and after 9/11/01 may be due to the inhalation of desiccating dust, now lodged in lung tissue.
Did the lingering hydrocarbon vapors and fumes from burning dissolve in water and create acid pools? Did the calcium-, silicon-, aluminum-, and magnesium-oxides of pulverized concrete form salts in pools of water? Did the sulfate from the gypsum wall panels also acidify standing water? Did acids work on metal surfaces over months, to alter their appearance?
In the enormity of each rubble pile, with its massive quantity of stored heat, many effects were possible in small quantities, given time to incubate. It is even possible that in some little puddle buried deep in the rubble, warmed for months in an oven-like enclosure of concrete rocks, bathed in an atmosphere of methane, carbon monoxide, carbon dioxide, and perhaps a touch of oxygen, that DNA was formed.
 MANUEL GARCIA, Jr., “The Physics of 9/11,” Nov. 28, 2006, [search in the Counterpunch archives of November, 2006 for this report and its two companions; one on the mechanics of building collapse, and the other an early and not-too-inaccurate speculative analysis of the fire-induced collapse of WTC 7.]
 “Executive Summary, Reconstruction of the Fires in the World Trade Center Towers,” NIST NCSTAR 1-5, , (28 September 2006). NIST = National Institute of Standards and Technology, NCSTAR = National Construction Safety Team Advisory Committee. https://www.nist.gov/topics/disaster-failure-studies/world-trade-center-disaster-study
 “Fire Structure Interface and Thermal Response of the World Trade Center Towers,” NIST NCSTAR1-5G, (draft supporting technical report G), http://wtc.nist.gov/pubs/NISTNCSTAR1-5GDraft.pdf, (28 September 2006), Chapter 3, page 32 (page 74 of 334 of the electronic PDF file).
 1 m = 3.28 ft; 1 m^2 = 10.8 ft^2; 1 m^3 = 35.3 ft^3; 1 ft = 0.31 m; 1 ft^2 = 0.93 m^2; 1 ft^3 = 0.28 m^3.
 “National Institute of Standards and Technology (NIST) Federal Building and Fire Safety Investigation of the World Trade Center Disaster, Answers to Frequently Asked Questions,” (11 September 2006). https://www.nist.gov/topics/disaster-failure-studies/world-trade-center-disaster-study
This article originally appeared as:
The Thermodynamics of 9/11
28 November 2006