This lecture was recorded during a symposium held May 13, 2006 at the Victoria Mansion in Portland Maine

The Evaluation of Alkoxysilane Consolidants on Portland Brownstone

Dr. George Wheeler, Director of Conservation, Columbia University Historic Preservation Program

Good morning. Thank you for being invited to speak here, to Robert and Ivan, and to have the opportunity to build on some of the things that George said and laid a good foundation for. I’d like to talk about alkoxysilane consolidants in general. I’d like to talk a little bit about the inherent properties of these materials. We know quite a bit about the inherent properties a Portland brownstone’s and then put that specifically in the context of things that go on in Victoria Mansion. We can use that as a model for any kind of conservation project in which you have an accretion material, and inherent material of the building, and in the exposure that is exposed to.

I’d like to acknowledge my co-authors here, Craig Oleszewski as a graduate student who is now being hired by building conservation associates, and Carola Garcia Manzano who is at the archaeology department at Columbia. Just a quick little background. There are no new ideas. Tetraethoxysilane was first synthesized in 1846, and in 1861 suggested as a stone consolidant. We’ve been merely fooling around for the last 120 years or so trying to get these things to actually work and understand the properties. They were made into a commercial product as early as the 20s by AP Laurie. Certainly Wacker, as we all know, patented the formulation that is similar, is not the same, as we used today.

Wacker OH

Wacker OH

That is the record in progress of work in using these materials. Moving into the chemistry, basically we are starting with a mobile liquid based on something like tetraethoxysilane and small ligaments of it, and hydrolyzing with water and producing silica gel. Which is essentially the same as those little packets of material you see when you buy electronic kind of equipment. Making a silicate polymer, again, in the form of a gel. What are some of the basic things we like about alkoxysilane? One is, as George mentioned in that stone history, is they are quite low in viscosity and surface tension. They are very mobile liquids. We will also see in a few minutes that that’s something we don’t like about them as well.

One of the reasons we put so much attention in these materials is their UV stability. Silicon oxygen silicon bonds do not break down in solar UV. They are not oxidized. They are a relatively stable material as we understand by all the silicate minerals that are in the Earth’s crust. Just a little bit about this property of surface tension. Water, as you know, has a very high surface tension. It’s something we might call a narcissistic molecule. It really loves itself and holds on to it. Things like alkoxysilane do not like themselves very much. They are self loathing materials, and they lay down in surface tension very well. They really do run, if you will. These are some things we like about alkoxysilane when trying to apply them to a course material.

What don’t we like about alkoxysilane consolidants ? The list is longer. Low viscosity and surface tension. I’ll show you a picture of that in a moment. Relatively long curing time. This makes jobs for contractors very difficult logistically to make sure the stone is dry beforehand. Give it enough time not exposed to rainwater for it to cure and to develop the good properties it might have. During that long curing time, they are also water repellent, even though the general itself in the end is not water repellent, during the curing time they are. You can’t carry out any activity nearby that might involve water, like packing and pointing. Things of that nature. Again, they are inherently immiscible with water. If you have water in the stone and your material is moving in it will develop a reaction at the interface that will precipitate the gel in a place that you may not wanted to be.

Color changes can persist for years or longer.

Color changes can persist for years or longer.

Color changes were mentioned just a few moments ago. There can be significant color changes that are rather persistent with alkoxysilane. Because of the water issue, it’s often difficult to clean after it’s consolidated building materials. The gels, themselves can’t bridge very large gaps. Has been some discussion among scientists about how big this number is. But 50 microns is very small and this has important consequences for a place like restoring walls. As I said, low viscosity is a blessing and a curse. It goes on and it goes into the stone really well and it also goes everywhere else. If you are trying to, perhaps as you mentioned, restore an isolated area, you have to isolate that area from the rest of material because it will run.

Color changes can persist for years. This is a stone from the lab that has been exposed, now, for three years and it still retains difference in color and notice also the water repellency that is retained for at least several months when it’s not exposed in and out during. Again, you can’t go back and do water related treatment nearby this material. Alkoxysilane will screw up the curing of mortar if you treat them on areas with newly installed mortar as a packing material. On the other hand, if you bring mortar in contact with alkoxysilane, that will screw up its curing. Water immisciblity. There is a big lump of liquid alkoxysilane Wacker OH in the bottom of this container of water. It does not mix with water.

The gels from alkoxysilanes are not good adhesives and they are not good gap bridging materials. We look at one of the main manifestations of deterioration in Victoria Mansion, the scaling, flaking and spalling. You can’t expect this alkoxysilane gel to bridge those gaps and did hear those things back together we’re talking about something on the order of 50 microns is an ability to bridge the gap. Very very tiny space. Here’s an example of a limestone which has rather large pores. These are probably on the order of 300 microns. This is the size of the biggest piece of gel you are going to get in there. It cannot bridge those gaps caused by the delamination, the flaking, the scaling, the spalling that you see often with Portland brownstone.

Water immiscibility is inherent to alkoxysilanes.

Water immiscibility is inherent to alkoxysilanes.

Want to look at some particular considerations for Portland brownstone here at Victoria Mansion. Again, emphasizing the point that it’s not going to cure this manifestation of deterioration. It’s not going to mend the current condition. It might be of the slowdown future deterioration. Just to give you a sense, this is a thin section of Portland brownstone from Victoria Mansion which has been thin sectioned. The blue is the size of the pore spaces and here is a bar that’s 300 microns. Notice there is a lot of these spaces are on the order of 50 microns or less so that in the case of Portland brownstone, on the microscale, we can hope to introduce the liquid and bridge some of the gaps from green to green, but not from large structures like flaking, scaling and spalling.

What I want to focus on today some new work we’ve done with mechanical testing of Portland brownstone in which we’ve made slices of the stone, and the slices are approximately 50 millimeters across and 2.5 millimeters thick. Then subjected them to a biaxial pressure flex. Which was, again, developed by some of our German colleagues. Where you place a sample on the larger ring and then impinge in the smaller ring onto the sample and look at its deformation and ultimately failure. One of the reasons we like to use this test, we get a lot of samples out of one piece of stone. In addition, if you are looking at evaluating treatments that have been done in the field, you can take a core drill through the external wall and look at the depth of penetration and the change in the property with that depth over time.

I’m not advocating drilling hundreds of two inch diameter holes in Victoria Mansion. This is a very good way of evaluating my condition. I wanted to give a little bit of, or actually, quite a bit of [inaudible 00:09:01] stress-strain data based on these tests. On the left-hand side, we have the applied load and on the right-hand side of the strain or displacement of that second head. The actual plot is the black line and I’ve taken some liberty here to draw in a blue line which did not conform to any of the particular protocols for measuring thickness or elastic modulus but I didn’t want it getting in the way of the black line. This is a way of looking at how thick the sample is. The more steep the graph, the more thickness of sample.

Just comparing the Longmeadow with Portland, one notices with the stress-strain curve that Portland is actually quite a bit stronger than the Longmeadow and in fact is inherently somewhat stiffer. This is Longmeadow and Portland untreated. We had done anything to this yet. We’re just looking at how the stone performs in relation to the stress. These are, in this case the discs are all cut parallel to the bedding plane. These are all 10 samples each with the exception of one sample at the end. One of the things we want to look at is what happens when we treat these samples in the lab. And I said that from before, how are we treating, these are, as I said, ideal conditions.

We are learning how good we could possibly be the with these treatments versus what may happen in the field. Notice, if we go back a little bit, notice how much more steep the graph gets. These materials are responding much much stiffer. Maybe not such a good thing as we taught in relation to what George was saying just a few moments ago. We have to things occurring. With Longmeadow it’s getting relatively much stronger. It’s not as strong as Portland but relatively much stronger. 250 percent versus 80 and a 550 percent increase in stiffness. We might be concerned about treating Longmeadow and increasing it stiffness that much. Portland also a relatively high increase in stiffness. Not, again, as high as Longmeadow that an 80 percent increase in the load failure.

We are getting some additional strength by this treatment and we want to see if it’s useful for us or not. It’s worth looking at stones when they are wet, as well. We measure the strength Portland, untreated Portland when it’s dry which is the same graph you saw untreated before. And then when it’s wet. Notice that the modulus does change somewhat. The stiffness changes somewhat. The slope is a little less. In relative terms, it’s not that much weaker. It’s only about a 30 percent loss in strength when it’s wet. That is for Portland, many other stones are much weaker when they are wet. We want to also look at when we have treated Parliament, what happens. These are both treated samples, now.

Again, this is how good it is when it is dry and treated. What if you saturate the sample with water and then break it after it’s been treated. Notice, we do have a decrease in strength. It’s similar to the decrease in strength that occurs just from untreated samples going from dry to wet. The other thing that you noticed is that this strength here of the white material is actually higher than the untreated stone while it’s dry. Again the treatment is helping a little bit in terms of the strengthening of stone in conditions that it might be exposed to here in Victoria Mansion.

One of the things we believe occurs when you treat sandstone with alkoxysilane, because of the similarity, chemical similarity, between the filling in the stone, that means the stone is made up of silicate, quartz and feldspar and clay, the silicate minerals, and the gel, is that you actually get some bonding that occurs between the gel and the stone substrate. One of the things that we see when we treat limestone versus sandstone, the sandstone gets much higher strength increases and higher strength because of this bonding effect. One of the things that will probably occur when you wet treated stone is that it will start breaking some of those bonds between the minerals and gels. That may, in fact, be part of this difference in the strength that occurs here.

Section of Brownstone from Victoria Mansion

Section of Brownstone from Victoria Mansion

What happens, then, if you take this which treated stone and let it dry again? Do those things reform and do you regain the strength that you had before now that it’s dry, after it’s been wet? This is the stone that treated saturated and then dried. If we go back one, notice this is where the treated is wet, then as you go forward, it is dry again. We’ve essentially regained all the strength that we had before. In this single occurrence, this is one cycle now, we don’t meaningfully lose all of the value of the treatment through that one cycle. Perhaps the bonds that we believe form, I believe there is some data that these bonds do form, are reformed after wedding and then drying again. It’s typical of the condition of Victoria mansion where it is wet and then dry again.

One of the things we like to do, and I would like to use George’s apparatus but we’ve done the wet drying cycles as well to examine both what happens to the stone and what happens to the treated stone. When we look at untreated stone that’s been cycled, this is only 20 cycle so this is not a long period of time, but what we notice with untreated stone on cycles is that it doesn’t change its properties very much on that timescale. We don’t think it will be worthwhile doing many more cycles than this. Here’s the treated stone with no cycles. After only about 20 cycles, and this is an average of 10 samples, we see a real change in the strength of the treated material. The untreated stone, in fact, doesn’t change very much in 20 cycles.

The treated stone does begin to change on the order of 10 or 15 percent. What’s that? That’s telling us that our treatment is starting to lose effectiveness in a relatively short period of time. We should extend the cycles more to see where the bottom line is on this but, again, our treatment is starting to lose some of its effectiveness in a very short period of time as measured by the strength. And of course, wetting and drying is an important condition here at Victorian mansion, and of course in other places as well. Another condition that we see is freeze thaw cycles and here we have samples with no freeze thaw cycle and then we use a cycling to -10 degrees Celsius and then back to room temperature on the order of about four hours at a time.

Notice that it actually appears after 20 freeze thaw cycles, the stone is getting a little bit stronger. It stiffness is fairly similar. Again, in 20 cycles, is doing a lot to the stone itself. Maybe it’s improving it a little bit and depositing some little grants of material between the brands that are causing it to be a little stronger or a little stiffer. Now look at the treated Portland stone, again with no cycles. And then after 20 cycles, we’ve already begun to see a loss of something on the order of 15 or 12 percent of the strength. These are real differences averaging out over 10 samples will similar to the wetting and drying we begin to lose some of the effectiveness of the treatment after only 20 freeze thaw cycles.

Two conditions that are prominent here at Victoria mansion, wetting and drying, freezing and drying. The stone is not changing a whole lot in 20 cycles we are getting appreciable differences in the consolidation material. Again, we are losing effectiveness of the consolidant. Again, with the treated material the slope of this graph is much much higher so the material is much stiffer. If we go back to the hybrid dwelling problem, we may be creating a condition that’s less favorable to that part of the deterioration from a Portland or sandstone here. One of the things that’s done, and I just want to shift for a few minutes to what we do in the field.

That is one of the common phenomenon that occurs in the field is that we get access consolidant on the surface as we are applying it and the manufacturer [inaudible 00:18:58] suggest you apply acetone on any treated surface to get rid of that glazing affect. See you don’t get access consolidant on the surface. One of the things I was concerned about is if you do that to you cause a problem in the actual strength improved by the treatment. Conditions for anesthetic effect, would it influence the mechanical properties. Here is the treated material, again, with no acetone. When you apply the acetone it appears, in fact it’s getting a little stronger. Maybe it’s redistributing some of the consolidant and better between the grains. Is not causing a diminution of the strength of material by applying that acetone to reduce the glazing affect.

For a field condition, that’s a good thing for us to know. As George mentioned, asked if this material gets is also important in terms of how it deteriorates. George also mentioned the company Remmers. They produce a couple consolidation formulations that have elastomeric materials and then in an attempt to make both the gel that forms and the treated stone less stiff. We thought maybe if we did a comparison of these products to what the typical Conserve OH or Wacker OH which is just an ethyl silicate and dilorate catalyst. Here is a graph of our typical Conserve OH and here is the same material with an elastomeric component. Notice that the slope of the graph is almost exactly the same.

Slightly less stiff but this elastomeric material isn’t really giving us a significant difference in that stiffness. Whatever the formulations, in fact we know what it is, is a silicon resin attached to some of the alkoxysilane polymers. It doesn’t manifest itself being treating Portland sandstone. Maybe other stones it would have more of a significant effect. It does not have a significant effect on Portland sandstone. Again, we just are adding some of these linear segments to give a little bit of flexibility to the gel and hopefully transfer some of the flexibility to the stone. Again, very little of that is transferred to the treated material. We’re not gaining a whole lot in this property change that these treatment materials are designed to do.

Victoria Mansion

Victoria Mansion

The thought being do we want to start importing some of these materials from Germany because they have magically better properties than what we already have here. For Portland sandstone, probably not. Another product, which is an elastomeric material from Remmers as well, has a little bit higher molecular weight, slightly higher viscosity, not significantly higher. 500 STE, comparing that to OH, actually this one is a little stiffer than our OH. We’re not gaining anything from the elastomeric properties. It’s higher molecular weight does produce an appreciably higher strength increase. We are on the order of 225 rather than on the order of 185. The higher molecular weight material does generally produce higher strength and we do see that here in the 500 STE versus the typical OH.

It’s not coming out to be less stiff so I think that’s also an important thing for us to note do we want to import this material for Portland brownstone, probably not. Having said that, we can probably give you this material from things we can buy off the shelf, that way you don’t have to import it from Germany, they don’t have any patents here. I think that’s an important thing for us to note as well. Getting mixed up with my hands here. This is a very important set of graphs. I’d like to go through this carefully. This is following the protocol that the Germans developed in the early 80s in Munich, where they’ve actually gone out in the field and treated the building as they would in a typical application by a conservator, core drilling the building from its facing and slicing through in depth and looking at the different strength as they move in depth.

We did this with a core and followed the typical conservator treatment protocol of one minute in five minutes out, one minute and five minutes out for nine cycles, which is fairly typical for consolidating a stone in the field. Then we sliced this thing up in depth and looked at how the strength changed in depth. This thing is actually a fairly disturbing graph. Here is the graph for the first slice. We have, actually, a little better strength improvement in this case and we have for most of the samples we included earlier. They were down at about 100, 185. We will call this good for the moment. We had a high strength increase. We go into the next bed and we are down here below the treatment that we did in the lab.

We’re down to about 165. You go to the third disc, we are down here at 130, and the fourth fifth and sixth we are down here to untreated materials. What does this mean? Remember, these disks are 2.5 millimeters in thickness. Within 2.5 millimeters our strength has dropped from 230 to 170. What we’re saying is we’re not getting enough depth of penetration. Down here at five millimeters in we’re already approaching the level of untreated stone. We have to be incredibly careful and good about how we do our field treatments if we want to make sure we had cadets of penetration. What might that mean, practically? You probably have to dilute these materials with solvents, as has been done in Europe.

You do that, you get better depth of penetration and perhaps more even strength profile. Now, we are out of DOC compliance and we have-just take a couple of minutes to summarize and give you some photos in some of these summary points. Alkoxysilanes do cause significant change in color and many stones, particularly dark colors. We do see that in Portland brownstone. In Portland, the color change is much less that it is in Longmeadow and it does abate over time. And I want to underline abate and doesn’t necessarily go away. It’s usually amounts to about a 10 or five percent difference in coloration. That’s important consideration. Timeframe, that can be on the order of years to get down to that 10 or five percent point.

We need to know that before you run out and treat that and see how long it might take for it to go away and whether that’s accessible to you are not. It may not be. The other hand, you might use this material to treat the lighter colored new stones that are installed. Second point, this is important for us to understand. I think we are oversold on some of these materials over the years. Alkoxysilanes’s cannot mend one of the main manifestations of deterioration at Victoria Mansion, scaling flaking and spalling. The gaps are too big. It’s not a good adhesive, it’s not a good Bridging material. You can’t ask it to do that, it will not do it. Grout injections and adhesives are the more typical types you use and conservation and mending those conditions.

It will make the stone so much stronger and perhaps resist these somewhat longer but not that long. Again, we can’t ask these materials to do things that they are not designed to do. Try to be honest about them and what of the good and bad things about these materials. Alkoxysilanes to give noticeable increases in strength and stiffness to Portland brownstone. This may be a good or bad thing. Again, with respect to high-grade swelling and making it much more stiff may not be a good thing to do. We have to be careful about what we think we are gaining. The strength increase well, again, slow down some of the effects. Slow down some of the effects of freeze thaw and wetting and drying.

It will not make them go away. Though three-star and wet dry cycles reduce the performance of alkoxysilane consolidants . The consolidants themselves are affected by freeze thaw and wet dry in a way that reduces their performance. That means they should be considered a maintenance material and not restoration material. By restoration material I think that’s something that you can apply and it holds onto most of the performance for 25 years or longer. You’re not going to get that out of these consultants on Portland sandstone. It’s going to be more like painting your house that it is building a new house. These are maintenance materials, they are not restoration materials. We are talking about applications on the order of five and 10 years.

Knowing the logistical issues in relation to applying his consolidants , making sure the building is dry, making sure it’s protected from rain water, not being able to carry out other activities while these things are curing, it’s a real return to their use as a maintenance material. You have to understand, again, what these properties are before you jump ahead and consider using it. On the bright side, Victoria Mansion in the isolated sections can be treated but, certainly, as an overall treatment it would be very difficult to carry out, like painting a house, by doing it everywhere. Just to emphasize, again, your loss a performance of the treatment through these wet dry cycles as well as the freeze thaw cycles.

The treatment is being affected in a short period of time, 20 cycles. The treatment is being much more affected than the stone is. I haven’t shown this here because I wanted to emphasize this new data. One of the things we do know about alkoxysilanes with respect to granular disintegration on stones like Portland sandstone, this is a very useful treatment. It will reduce granular disintegration for longer periods of time that will hold onto the strength increases. That is a condition that is at the Victoria Mansion. There are areas in Ivan’s talk where we are not going to try to replace things as new material that there is some erosion. In those localized conditions of granular disintegration, we will get a pretty good improvement in the performance that would be an area where you may consider using alkoxysilanes.

Again, rather than an overall … Does not cause a diminution in the performance of the consolidant. If you do get this glazing affect and you want to wash it away with acetone or methyl ethyl ketone, you’re not going to be stripping out all of the consolidation material from the surface. It will retain whatever good properties you got by that initial treatment. Continue doing that in the field to protect from the glazing affect. This is really important. As a care must be taken in field applications to ensure sufficient depth of penetration of the consolidant. This graph, as I said before, is very disturbing. To, within 2.5 millimeters, lose 40 percent of the effectiveness of the treatment and virtually all of the effectiveness of the treatment down to five millimeters.

The traditional wisdom with these consolidants is a minimum of one centimeter, 10 millimeters for depth of penetration. On the order of five millimeters, we’ve lost all of the effectiveness of treatment. This protocol that is suggested by Prosoco and [inaudible 00:31:55] literature is not enough for Victoria Mansion. You have to let the staff sit there longer and penetrate better and do some of this testing to understand. Do some of this testing to find out what is that protocol look like. I should emphasize this. All these tests were done on only mildly weathered material. A lot of it, actually, from Victoria Mansion. The material that was removed in 88. More severely eroded material will take in the liquid faster.

You need to understand the conditions on the building and how it influences the uptake of the consolidant and how it affects the strength depth profile. It is a very important condition, particularly for this building. The elastomeric modified alkoxysilanes don’t significantly reduce the stiffness. He stated advantage of elastomeric material, as formulated by Remmers doesn’t really play out on Portland. Some of the data for some of the German sandstones that have been tested over the years are better than this. Again, we are talking about Portland here. It doesn’t really do that much for you. We don’t need to run out and try to import this. However, products containing elastomeric or these more highly polymerized ethyl silicate materials may conform to the new VOC regulations.

Higher molecular weight materials leave more stuff behind Celeste evaporates. There may be an advantage for both the elastomerics, because remember the elastomeric material is basically a solid that will never go away. It will increase the elastomeric component and increase the molecular weight a little bit, we may conform to VOC. We may have increased the molecular weight are we may have increase the viscosity and reduce the depth of penetration. That’s a playoff that we need to look at. The new regulations are in place. My colleague Norman Weiss and I have been discussing testing VOC on these and other materials. Most of the material that comes off of ethyl silicate is ethanol. The VOC on a martini is actually much higher than it is on the effort silicate material.

We will see if we can add a little martini, see how well that does. I do believe we can reduce the VOC on some of these materials enough to comply. If the regulations do work their way up to Maine, we still need to be able to supply you with a solution.

National Center for Preservation Technology and Training
645 University Parkway
Natchitoches, LA 71457

Email: ncptt[at]nps.gov
Phone: (318) 356-7444
Fax: (318) 356-9119