Where Masonry, Glass and Water Meet  
The Interactions and Ramifications of Concrete Runoff on Glass 

by Paul F. Düffer 

The consequences of rainwater that flows over a concrete surface and comes to rest on glass have long been the center of apprehension among those in the glass industry. The main concern has been the inability to restore the affected glass to its original, pristine condition by means of conventional cleaning techniques. Over the years, a number of articles and technical service bulletins have been published by both the glass and concrete industries in an effort to account for the observed damage and offer suggestions regarding prevention. However, to date, a comprehensive explanation of the processes and reactions at the root of the problem remains elusive. As a consequence, effective techniques for prevention and restoration have been slow to develop.

Conventional Wisdom
Over time, an explanation of the root cause for stubborn concrete runoff deposits on glass (such as those shown in Figure 1) has evolved. This conventional wisdom has developed into two distinct schools of thought, as gleaned from the marketing and technical service literature of the glass and concrete industries. 

The first of these relies upon the fact that commercial flat glass is susceptible to attack by alkaline solutions whenever pH levels are on the order of 9.0 or greater. ^{1, 2} 

Under very specific conditions, permanent staining or “surface corrosion” may ensue.³ A typical example of this permanent staining effect, or classic glass surface corrosion, is shown in Figure 2. Another equally important element of this school of thought is the fact that newly cured concrete contains alkaline materials, particularly CaOH, that can exhibit pH levels in excess of 11.0 when immersed in water. Proponents of this point of view contend that, under the influence of rainfall, alkaline leachate material can exude from concrete and cascade randomly down a building facade. Eventually, some of this material comes to rest upon glass and soon leads to permanent surface damage in the form of corrosive etching. Over the years, this explanation for runoff damage has gained wide acceptance in the flat glass industry. The other principal explanation for runoff damage maintains that glass can undergo corrosive surface damage even in what is actually a dynamic and otherwise benign environment, one in which no surface corrosion is anticipated. More precisely, it is argued that dust and debris residing on a glass surface can entrap moisture for sufficiently long periods of time so that glass surface corrosion can occur without the added influence of alkaline runoff from concrete.⁴ 

Increases in temperatures significantly accelerate the surface corrosion process. As far as the generation of alkaline runoff is concerned, this view concedes that these materials can indeed originate from masonry and concrete surfaces. However, it is believed that acidic air pollutants that include oxides of sulfur, nitrogen and carbon quickly lower the pH of the alkaline leachates to render them harmless to glass.⁵ Developed more recently than the “alkaline-leachate” rationale, this “moisture-entrapment” theory finds the majority of its support within the portland cement and concrete industries. 

Unfortunately, the trade literature that supports the alkaline-leachate theory does not cite experimental evidence that supports what otherwise appears to be an acceptable explanation for the commonly observed runoff damage. There are no studies reported in which alkaline solutions have been used to reproduce glass surface features that are similar in appearance to those characteristic of runoff damage. As a consequence of the glaring lack of experimental data, this same literature does not clearly identify the time and temperature conditions required for producing a genuine alkaline etch on the exterior surface of architectural or residential glazing installations. Nor are there any reports of actual pH measurements being made on runoff solutions isolated from concrete building facades. Hence, it remains questionable as to whether or not alkaline runoff solutions can naturally reside on the exterior surface of glazing systems for periods sufficiently long to cause chemical erosion or etching. Furthermore, it has not been firmly established that concrete runoff is unequivocally alkaline.

Experiments conducted in our laboratory have demonstrated repeatedly that if a dilute solution of calcium hydroxide (pH ~ 11.5) is placed on a glass surface at 73 degrees Fahrenheit in a controlled environment to retard evaporation, no damage is observed after 20 hours of exposure. A similar exposure at 140 degrees Fahrenheit produced no discernable damage, even when viewed with a 1,000-watt inspection lamp. Closer scrutiny with the aid of an optical microscope confirms there is no evidence of chemical erosion or etching. If such a process had been aggressively at work, one would expect to observe a significant modification in surface topography. An illustration of the distinctive microstructure of chemically etched glass is presented in Figure 3 where the non-glare surface of a video display faceplate is shown after etching with a 70-percent hydrofluoric acid solution.

Obviously the conditions maintained in the above experiments are much more extreme than those normally encountered by glazing systems in the real world. For example, it is highly unlikely one would observe water droplets residing unabated on a glass surface after cessation of a rainstorm. In addition, even if glass did remain wet for several days in succession due to ongoing rainfall, the surface water droplets residing on the glass would be continuously diluted and replenished. Moreover, consider the fact that not all rainfall that contacts concrete has time to permeate the surface micro-pores to extract potentially alkaline materials. In fact, much of the impinging water flows harmlessly away. It is only the last vestige of droplets adhering to glass at the end of a rainstorm that could possibly present a threat to surface quality via alkaline etching, if indeed the solution pH is 9.0 or greater and the residence time extends well beyond 24 hours. Consequently, it is highly doubtful that these droplets could exist intact over the extended periods required for severe alkaline etching to develop. Once evaporation occurs, the threat of chemical etching is ended. Although laboratory studies have demonstrated that glass can succumb to alkaline induced surface damage, it does so under very specific conditions that do not typically prevail in the dynamic environments encountered by architectural and residential glazing systems.⁶

A Legitimate Question
As mentioned earlier, a legitimate question exists as to whether or not runoff from concrete is truly alkaline. For example, during the course of the current study, rainwater with an initial pH of 4.0 was cascaded over 50 feet of continuous concrete façade at a high-rise facility in Pittsburgh. The leachate material collected exhibited a solution pH of 7.5. In similar experiments conducted at another location, the collected leachate had a measured pH of 8.5 after a cascade of approximately 30 feet. Likewise, laboratory experiments in which rainwater was carefully directed over a concrete surface to simulate runoff conditions failed to produce leachates with pH values greater than 8.5. In one instance, 35 iterations of runoff recycling was undertaken with a final pH of 7.9 being observed. After repeated pH measurements on a number of concrete leachate solutions, this investigation failed to isolate an alkaline solution that would ever pose even a remote threat to glass surface quality.

The alkaline-leachate position comes up short in two key areas when attempting to account for the concrete runoff phenomenon. First of all, alkaline etching of commercial flat glass requires extreme conditions that do not typically exist in the real-world environment of residential and commercial glazing systems. Secondly, rainfall that cascades over concrete surfaces in the open air is not necessarily alkaline in nature. As a result, it is most unlikely that alkaline etching of glass occurs at the exterior surfaces of structures.

The Moisture-Entrapment School
On the other side, the moisture-entrapment school maintains that fine dirt and debris can trap moisture for periods sufficiently long to induce glass surface corrosion. Here again, experimental evidence is not found in the associated trade literature or even academic studies. The salient feature of this position is that glass is alleged to undergo surface corrosion in dynamic environments. However, this contention stands in opposition to several previously referenced treatments of the subject in which uniquely stagnant environments were identified as being critical in initiating the process.

Genuine glass surface corrosion requires a moist, stagnant environment in which water remains in prolonged and undisturbed contact with glass. After days or even weeks at ambient temperatures, these circumstances can lead to permanent surface damage. Such environments include the tiny spaces between stacked glass sheets and the inside of insulating glass units (IGUs) where the seals have failed and moisture has made ingress. In addition, there have been several isolated incidents where glass in large interior spaces has succumbed to surface corrosion. 

However, the glass in these instances was exposed to environments saturated with water vapor where condensation did not readily evaporate from the affected surfaces. One situation involved a heath spa in Japan, and the other was an apartment unit under construction in the United States where water pipes had burst and remained unattended for several weeks during summer’s heat. It is noteworthy that no confirmed incidents have been identified in this investigation where classic glass surface corrosion has occurred in a truly dynamic environment. This includes the exterior surfaces of both residential and commercial glazing systems.

Temperature is another key factor to consider when discussing glass surface corrosion. At ambient temperatures, the time frame required to produce visible surface corrosion can extend to weeks, if not months. Over a 5 to 7 day period, temperatures in the vicinity of 135 to140 degrees Fahrenheit are required, along with a moist, stagnant environment, in order to produce corrosive damage visible to the naked eye under normal lighting conditions. On the other hand, a dynamic environment would permit rapid evaporation of the moisture at these temperatures. In view of the conditions required for visible surface corrosion to occur over a relatively short time frame (5 days), it is difficult to envision a scenario of this sort prevailing in what is a naturally mercurial environment. Indeed, if the moisture-entrapment contentions were accurate, there would be many observable examples of spontaneous corrosion on the exterior surfaces of glazing systems. 

However, such observations have not been documented. What is commonly observed is surface corrosion that has occurred on the interior surfaces of IGUs in which the seals have failed. Moisture subsequently makes ingress into the affected units and eventually induces permanent surface damage. Additional evidence in support of abandoning both the moisture entrapment and the alkaline-leachate theories can be gleaned by comparing the microstructure characteristic of glass surface corrosion with that typical for concrete runoff damage. 

For example, Figure 4 presents a photomicrograph of glass corrosion magnified 200 times that was obtained with a scanning electron microscope. It is readily observed that this microstructure stands in stark contrast to that typical for concrete runoff as shown in Figure 5 at a similar magnification. The corroded glass exhibits a surface characterized by distinct cracks and fissures. 

On the other hand, the runoff microstructure does not resemble that of surface corrosion, nor is there evidence of any surface erosion or etching similar to that depicted in Figure 3. Quite the contrary, the runoff topography visible in the photomicrograph suggests that a deposit actually persists on the glass surface. Confirmation of this hypothesis was achieved by means of surface contact profilometry that has shown concrete runoff to consist of tenacious entities that actually reside in relief on the surface of glass. There are no depressions or fissures present and no evidence of surface corrosion or alkaline etching.

A New Perspective: Silica and Soluble Silicates
Experimental evidence for the alkaline etching of glass by concrete leachates from building facades has not been substantiated. Furthermore, there is a lack of evidence supporting the contention that glass surface corrosion can occur in dynamic environments due to moisture entrapment by dust and debris. 

As a consequence, the question remains unanswered as to what accounts for the permanent surface damage frequently observed on glass installed in juxtaposition to concrete building materials. Perhaps the most overlooked aspect of the water, glass and concrete discussion has been the fact that glass and portland cement—a key ingredient of concrete—are actually fairly close relatives from a compositional point of view. Even their thermal history in the melting tank and rotary kiln are similar. Table 1 shows typical compositions for a sample of portland cement and a commercial float glass product.

It is easy to see that portland cement contains a significant amount of silica bearing material. The principal chemical compounds in which the aforementioned constituents are found in portland cement have been identified to be tricalcium silicate dicalcium silicate, tricalcium aluminate and tetra calcium alumino ferrite.⁷ 

Professor Henk de Waal first suggested that dissolved silica in concrete leachates accounts for the difficulty in attempting to clean or refurbish glass affected by runoff.⁸ 

However, the silica deposits were described as causing a “chemical etch” and not identified to be a surface deposit. On the other hand, Dr. Ralph Iler pointed out in his work on the chemistry of silica that dilute solutions of silica and silicates can deposit insoluble, amorphous residues on solid surfaces. More precisely, upon evaporation to dryness such solutions reach a point at which super-saturation occurs.9 This, in turn, is accompanied by the formation of polymerized silica and polysilicate species that generate amorphous, water-insoluble precipitates upon the substrate under consideration. In the case of oxide surfaces, chemical bonding to the substrate can occur. When rainwater makes contact with a concrete surface, small amounts of silicate material dissolve. Silica concentrations of 4 to 8 parts per million are not uncommon. Of special interest here, however, is the fact that commercial flat glass is characterized by an amorphous network of silicon and oxygen atoms that can participate in the reaction described by Iler. 

In particular, the last droplets of silica-bearing leachate that come to rest on glass at the end of a rainstorm subsequently evaporate. Although the initial solution may be quite dilute, a point is reached where super-saturation occurs and the polymerization of silica and silicates ensue. The result is a tenacious, water-insoluble deposit that is chemically bonded to the glass and resists conventional cleaning agents.10 In fact, these materials are so difficult to dissolve that hydrofluoric acid is required for cleansing in laboratory experiments. X-ray diffraction studies undertaken in our laboratory have confirmed deposits such as those shown in Figure 5 exhibit an amorphous structure. Energy dispersive X-ray analysis (EDX) has revealed a composition that is rich in silicon oxide. Basically, the polymerized silicates form “glass-on-glass” defects that require specialized methods for removal on a commercial basis. 

The reaction between glass and evaporating silica solutions is significantly more rapid than either surface corrosion or alkaline etching. In experiments conducted as part of this study, it was observed that silicate solutions formed insoluble deposits as quickly as evaporation to dryness occurred at an ambient temperature of approximately 73o Fahrenheit. In addition, tenacious silica deposits also were observed to form in refrigerated environments at 43o Fahrenheit. In a comparative experiment that set out to intentionally induce glass surface corrosion, a test pack of flat glass was found to exhibit no visible surface corrosion after 75 days of exposure at 43o F.

Potential Remedies 
One prevention solution has been presented by Freedman et al of the Precast Concrete Institute (PCI) in Chicago.11 The design recommendations offered by these authors represent the best current technology for long-term protection against glass surface damage in structures where glass and concrete are used in proximity to one another. Unfortunately, not all buildings utilizing concrete and glass have been designed with the PCI recommendations in mind. Other methods for prevention have developed around the use of organo silicon compounds as surface treatments. 

Both silane- and silicone-based systems have been used, but in each instance there is a limited functional lifespan for an effective surface treatment. Most often, attention is focused on the situation after the surface damage has occurred. Therefore, cleansing or restoration procedures must be implemented before any surface treatments are applied. However, extreme care must be exercised in using chemical agents capable of dissolving the silicate residues since they also will dissolve glass and damage metal framing materials. 

The use of polishing compounds must also be carried out with diligence to prevent scratching the glass surface. There is clearly a need for technological developments in the area of concrete sealers and additives. For example, some sealers evaluated in this work were effective in reducing the amount of dissolved silica in simulated runoff solutions by more than 50 percent when compared with untreated concrete. Certain additives to concrete batch materials can also have a positive effect in reducing silica leachates. However, there remains a distinct need for improvements on all of these fronts in order to effectively reduce, if not eliminate, silica and silicate leaching from concrete.

Conclusions 
Conventional wisdom has failed to adequately explain the key factors that account for the significant difficulties encountered when attempting to remove concrete runoff damage from glass. The alkaline-etching theory fails to clarify the primary mechanism at work due to the fact that conditions necessary for alkaline etching of glass do not prevail within the environments where glass is actually installed in residential and commercial glazing. Furthermore, runoff from concrete is not necessarily alkaline, a fact that tends to contradict the commonsense view. Speculation that moisture-entrapment by fine dust and debris induces classic surface corrosion also fails on the grounds that the conditions required for this process to become influential do not exist in the real world of flat glass installations. 

The polymerization of silicate and silica materials upon the evaporation of concrete leachate solutions presents a new perspective that accounts for the tenacious deposits on glass that remain undaunted by normal cleansing procedures. As a result of the glass-on-glass entities that form, extreme care must be exercised in attempts at restoration using either chemical dissolution or mechanical polishing. Future improvement and potential elimination of the runoff problem will require technological advancements in concrete sealers and batch additives and the judicious use of these materials. Likewise, improved surface treatments for glass can make a significant contribution. 

In the meantime, it is worth giving due consideration to the building design recommendations offered by the Pre-Cast Concrete Institute as a means of reducing the incidence of runoff problems on glass. 

USG
© Copyright 2008 Key Communications Inc. All rights reserved.
No reproduction of any type without expressed written permission.