Clemson University TigerPrints
All Theses Theses
12-2012 Characterization and restoration of historic Rosendale cement mortars for the purpose of restoration Stephanie Hart Clemson University, [email protected]
Follow this and additional works at: https://tigerprints.clemson.edu/all_theses Part of the Materials Science and Engineering Commons
Recommended Citation Hart, Stephanie, "Characterization and restoration of historic Rosendale cement mortars for the purpose of restoration" (2012). All Theses. 1549. https://tigerprints.clemson.edu/all_theses/1549
This Thesis is brought to you for free and open access by the Theses at TigerPrints. It has been accepted for inclusion in All Theses by an authorized administrator of TigerPrints. For more information, please contact [email protected].
CHARACTERIZATION AND RESTORATION OF HISTORIC ROSENDALE CEMENT MORTARS FOR THE PURPOSE OF RESTORATION
A Thesis Presented to the Graduate School of Clemson University
In Partial Fulfillment of the Requirements for the Degree Master of Science Materials Science and Engineering
by Stephanie Anne Hart December 2012
Accepted by: Dr. Denis Brosnan, Committee Chair Dr. John Sanders Dr. Kathleen Richardson
ABSTRACT
Mortar was a very common building material in today’s historic sites. Before Portland
cement was manufactured at a global level, Rosendale cement was commonly used in
these mortars. Over time, these mortars in historic sites have begun to break down and wear away. With Rosendale cement in production again, measures can be taken to restore and repair the historic mortars. However, little testing has been done to establish durability of modern Rosendale cement mortars. This presentation highlights the common mix techniques used at the time, and undergoes experiments to establish general properties and predict future durability. Six different mortar mixes were tested with varying cement content and using various lime additions. Properties observed include compressive strength, absorption, porosity, permeability, and bond strength. Ion chromatography was used on seawater-soaked samples to determine how the Rosendale cement mortar would react with the seawater. Relationships between these properties were also addressed. It was found that cement content played a large role in compressive strength, while lime content had an effect on bond strength. Ion chromatography was used on seawater-soaked samples to determine how the Rosendale cement mortar would react with the seawater. Magnesium sulfates, and chloride were taken up into the mortars, indicating that Rosendale would be venerable to salt attack.
ii DEDICATION
I would like to thank my family, close friends, and my love Graham. Without all of you,
I would have surely given up or gone crazy. Your belief and support in me means
everything.
iii ACKNOWLEDGMENTS
There were many people involved that made this thesis possible. First, I would like to thank my advisor, Dr. Brosnan for his guidance and direction. I would also like to thank
Dr. Sanders, for his help and advice with all my testing and Dr. Richardson for her
participation in my committee from afar.
None of this would have been possible without Clemson University for accepting me into
the program and the National Park Service for the funding of this project.
Finally, I would like to thank everyone at the National Brick Research Center: Dr. John
Sanders, Gary Parker, Michael Mason, Graham Shepherd, Parker Stroble, Mark Young, and Greg Bellotte; for providing the facilities, testing equipment, and their continued help and support throughout my experience.
iv TABLE OF CONTENTS
Page
TITLE PAGE ...... i
ABSTRACT ...... ii
DEDICATION ...... iii
ACKNOWLEDGMENTS ...... iv
LIST OF TABLES ...... vii
LIST OF FIGURES ...... viii
CHAPTER
I. INTRODUCTION AND MOTIVATIONS ...... 1 Introduction ...... 1 Motivations ...... 2
II. LITERATURE REVIEW ...... 4
Rosendale Cement ...... 4 Strength ...... 9 Salt Attack - Introduction...... 11 Sulfate Attack...... 10 Magnesium Effects ...... 15 Effect of Chlorine ...... 18
III. EXPERIMENTAL PROCEDURES ...... 23
Description of Materials and Mixes ...... 23 Fabrication of Mortars ...... 24 Compressive Strength Test ...... 27 Tensile Splitting ...... 29 Bond Wrench ...... 30 Pier Compression ...... 31 Water Vapor Transmission Test ...... 34 X-Ray Diffraction ...... 35
v Table of Contents (Continued)
Page Thermogravimetry and Differential Scanning Calorimetry ...... 35 Mercury Intrusion Porosimetry ...... 36 Initial Rate of Absorption ...... 36 X-Ray Flouresence...... 36 Salt Leaching Experiments ...... 37
IV. RESULTS AND DISCUSSION ...... 38
Rosendale Characterization ...... 38 X-Ray Diffraction Mix Characterization ...... 40 X-Ray Fluorescence Characterization ...... 41 Simultaneous Thermal Analysis ...... 43 Porosity ...... 46 Compressive Strength ...... 48 Water Vapor Transmission ...... 52 Initial Rate of Water Absorption ...... 54 Summary of Mechanical Testing ...... 56 Bond Strength ...... 57 Pier Compression ...... 58 Magnesium Sulfate Attack ...... 58 Chloride Effect ...... 66
V. CONCLUSIONS...... 69
APPENDIX A: RAW DATA ...... 71
REFERENCES ...... 81
vi LIST OF TABLES
Table Page
2.1 Chemical compositions of various natural Roman and American cements ...... 5
2.2 Phase composition of various natural Roman and American cements ...... 6
2.3 Typical composition of seawater ...... 11
3.1 Description of mixes used in experiments in weight percent ...... 24
4.1 Oxidized chemistry of raw materials ...... 39
4.2 Oxidized chemistry of mortar mixes ...... 42
4.3 Loss of ignition for the tested mixes ...... 42
4.4 MIP results for all tested mixes ...... 46
4.5 Summary of results sorted by mix ...... 56
4.6 Results of the bond strength test ...... 57
vii LIST OF FIGURES
Figure Page
2.1 Micrograph of clinkered Portland cement ...... 7
2.2 Graph showing the relationship of porosity with strength ...... 10
2.3 DTG of Portland cement paste hydrated for one week ...... 14
2.4 TG and DTG of a plain Portland cement mortar (a) before and (b) after 200-day exposure to 5% MgS solution ...... 17
2.5 Expansion of Portland cement mortars stored in seawater and groundwater solutions ...... 19
2.6 Strength reduction in Portland cement mortars ...... 20
3.1 Addition of water to dry mortar mix ...... 25
3.2 Placement of the wet mortar mix into the molds ...... 26
3.3 Fresh mortar cubes in the initial humidity chamber ...... 27
3.4 Compressive strength testing of the 2”x2” mortar cubes ...... 29
3.5 A brick couplet in the bond wrench apparatus after failure ...... 31
3.6 A brick pier following the pier compression test ...... 33
3.7 Water vapor transmission testing apparatus ...... 34
4.1 X-ray diffraction pattern of Mix C ...... 41
4.2 Mix C TG, DSC, water, and carbon dioxide analysis after 28 days curing ...... 44
4.3 Thermogravimetry of all tested mixes ...... 45
viii List of Figures (Continued)
Figure Page
4.4 Comparison of cement content with porosity with different mix types ...... 47
4.5 Comparison of different mix types’ cement contents with bulk density ...... 48
4.6 Compressive strength over 28 days of curing ...... 49
4.7 Comparison of cement content and compressive strength ...... 51
4.8 Comparison of porosity and strength in tested mixes ...... 50
4.9 Results of the water vapor transmission test for each mix ...... 53
4.10 Comparison of water vapor transmission and volume of large pores...... 54
4.11 Water absorption for the mortar mixes over a 24-hour period ...... 55
4.12 Sulfate ion concentration in seawater over a 180-day soak period ...... 59
4.13 TG, derivative TG and water analysis of an unsoaked Mix C cube ...... 60
4.14 Mix C TG, DTG, and water analysis after the 180-day seawater soak ...... 61
4.15 Comparison of Mix C unsoaked (green) and soaked (red) TG/DTG and water analysis ...... 62
4.16 Magnesium ion concentration in seawater over a 180-day soak period ...... 63
4.17 Concentration of magnesium ions varying lime putty content...... 64
ix List of Figures (Continued)
Figure Page
4.18 Concentration of calcium ions varying lime putty content...... 66
4.19 Chloride ion concentration in seawater over a 180-day soak period ...... 67
x CHAPTER 1
INTRODUCTION AND MOTIVATIONS
Introduction
Cement has long been an important material in the use of construction. Its first uses can
be dated back over 4000 years ago in Egyptian times [1]. Brick and mortar were also a
key component in building America. Following the Revolutionary War and the War of
1812, natural cement was primarily used in mortars to build buildings and forts. This use
of natural cement continued through the Industrial Revolution. In 1875 the first
American Portland cement was produced. By the turn of the century, Portland cement
dominated production.
During the time of natural cement, a majority of it came from cement rock mined in
Ulster County, New York, near a town called Rosendale. Nearly half of all natural
cement used was Rosendale [1].
Thinking back, the structures built with natural cement are over a century old. Some of
these structures are beginning to degrade, so interest in natural cement has resurfaced in
the interest of restoration. It is usually recommended that a repair mortar is of similar compositions and physical properties [2]. Because Rosendale cement was the most
popular at the time, it has again become available for these restoration projects. Because
of its popularity historically, this project will focus on using modern Rosendale cement
for fabrication of the mortars.
Over a mortar’s lifetime, various factors affect how the mortar will perform. Some of these factors are apparently evident, such as strength and porosity. Others take time to make themselves apparent, such as resistance to salt attack or freeze/thaw durability.
However, these factors could also be estimated using analysis techniques such as X-ray
fluorescence, X-ray diffraction, and permeability.
Understanding salt attack is essential for restorative mortars. One of the main issues
today in historical monuments is salt crystallization [3]. Many structures were built on or
near coastlines. Knowing how the seawater compromised the mortar can lead to a better understanding on how to make mortars to minimize this effect.
The goal of this project was to characterize modern Rosendale mortars and establish a
relationship between composition and durability. Being able to predict how the mortars
will react will lead to a better understanding of how to use Rosendale cement in
restoration for the future.
Motivations
The primary motivation for this project was to determine if modern Rosendale cement is
a reasonable material to use for restoration projects in structures that used Rosendale
2 cement at the time of construction. The goal was to improve the structure while not
causing any harm to the existing components. The modern Rosendale mortar must therefore have a high compressive strength, good durability from water and seawater ions, and not cause any future damage, such as calcium leaching out to damage bricks
[3]. Cement is the most expensive component of a mortar, so retaining good properties while minimizing the cement and cost was also important. This project looked at varying cement contents, plus some cement replacement with other lime-based binders. It was investigated to see how the binders would act and if they would be an option to use in restorations.
3 CHAPTER 2
LITERATURE REVIEW
Rosendale Cement
Natural cement was produced and used in construction from 1818 through 1970. Of this
time, over half of the 35 million tons produced originated from near the town of
Rosendale, NY [4]. Natural cement is made from the natural cement rock that is
excavated. These “cement rocks” are limestone rocks rich in clays [5]. More
specifically, Rosendale is produced from argillaceous sedimentary rocks with high
dolomitic content. The rock is then calcined by burning, and then ground into a powder
to form the cement.
Natural cement differs from lime cement in two noticeable ways. First, it does not slake,
or become liquid, when water is added. Also, it demonstrates hydraulic properties such
as setting under water [6]. Even though it exhibits hydraulic properties, it still differs
from the most popular hydraulic cement today – Portland cement. Portland cement is a
chemically controlled, synthetic mixture that has little variability in chemistry or
properties. Natural cement comes from burnt natural rock, which can show significant
variability between layers and locations. Variation in color and chemistry are common.
Table 2.1 shows chemistries of some Rosendale cements. A physical property that
differs greatly from Portland cement is the time it takes to set, or harden. Natural cement
sets much more quickly than Portland, resulting in a lower ultimate strength [6].
4
Table 2.1: Chemical compositions of various natural Roman and American cements [5]
Chemically, Rosendale cements can be very comparable to some Portland cements.
Table 2.1 shows the variation between numerous natural cements, including two different
Rosendale mixes. Even if two cements have identical chemistries, the difference in cement lies within the firing of the rock. Portland ceement is fired in excess of 1400oC, commonly called “clinkering.” The calcium silicate phase alite (C3S) is formed at around
o 1300 C from belite (C2S) and lime. This presence of alite is what causes the differences in strength and set time. Rosendale cement was produced when tthese temperatures could not be obtained, so no alite forms. Rosendale cement is; however, fired over the calcination temperature of around 800oC, which lends to the cement having hydraulic properties [7]. This is when the belite phase forms, which is the cement phase that hydrates to form the calcium-silicate-hydrate (C-S-H) gel phase, which gives the cement its strength.
5 Table 2.2 shows the phase compositions for the same cements observed in Table 2.1. As expected, the Rosendale mixes did not contain any alite. It is important to note that the
Rosendale mixes contained significant quantities of periclase, or crystalline MgO, compared to other natural cements. This can also be used as an identifying tool to distinguish Rosendale forensically from other natural cements. This is due to the preexisting limestone already present in the Rosendaale cement rocks. This does not cause much of a problem in America; however, since Rosenndale was one of the predominant natural cements used.
Table 2.2: Phase composition of various natural Roman and American cements [5]
Figure 2.1 shows a micrograph of a cement with alite present, indicating that the cement was clinkered and is, therefore; not a natural cement..
6
Figure 2.1: Micrograph of clinkered Portland cement [7]
Both lime paste and powdered lime hydrates were commonly added to Rosendale cement
[4]. Lime was a common additive because it added favorable workability for the mason, demonstrated some self-healing properties, and increased the bond strength between mortar and brick [8]. The joint strength and self-healing comes as a result of the chemical reaction:
Ca(OH)2 + H2O + CO2 CaCO3 + 2 H2O
This calcium carbonate can fill small cracks or voids in the mortar itself, or those within the adjoining brick.
7 Sand was also added as aggregate to complete the mortar mixture. Aggregates are used to give the mortar volume, as cement paste is expensive alone. Based on the application, binder-sand ratios varied from 1:1 to 1:2¾ [6].
With the reintroduction of Rosendale cement in 2004, restoration projects that were once overlooked can now be attainable [4]. Many structures in the United States contain
Rosendale cement mortars, including: the Brooklyn and Washington Bridges in New
York City, the Capitol in Washington, D.C, a multitude of historic forts on the eastern seaboard, and South Carolina cotton mills [9]. The use of other cements or materials for preservation or restoration would not be historically accurate, as well as posing potential problems for the existing materials. Using Rosendale similar to that in initial construction would ensure accuracy, as well as not harm the existing structure.
Distinguishing if one has Rosendale cement in their historical mortar is the first step to deciding to use Rosendale. Methods for analyzing mortar can be found in ASTM C
1324. Chemically a cement mortar can be distinguished from a lime mortar due to the higher silica and aluminum fractions [7]. However, as stated earlier, Portland and
Rosendale cements can be similar in composition. X-ray diffraction[5]or microscopy [7] can be used to determine the presence of alite, thus distinguishing a Portland cement mortar from Rosendale.
8 Strength
Hanley studied the workability of natural hydraulic lime mortars (NHL) and the effects on strength [10]. Three different NHL binders varying in intensity of hydraulic behavior were tested at three different water contents resulting in flows of 165, 195, and 195mm.
Flow is an important property to determine the workability of the mortar, typically a higher flow decreases stiffness and indicates improved workability; however, too high of a flow leads to a watery consistency that is also difficult to work with. Compressive and flexural strengths were calculated at 28-day and 56-day curing periods. The results concluded that stronger binders correlated with stronger mortars, so lime can increase the strength of mortars. Within each category of limes, the one described as “easiest to work with” over “dry” or “runny” also yielded the most desirable strength results. This visual workability test makes making a stronger mortar on-site easy for anyone, but must be kept consistent on a batch-to-batch basis. The difference between “good” and “runny” mortar can vary by less than one percent water content, so consistency is essential.
Papayianni also conducted a study of mortars and their relationships to strength. It is stated that strength is inversely related to porosity in cement mortars by the equation
1 [11]. Figure 2.2 shows that experimentally this also holds true. Lime- pozzolan mortars were compared with 1-porosity (percentage of bulk volume) with strength, and a direct correlation can be made.
9
Figure 2.2: Graph showing the relationship of porosity with strength [11]
Various ratios in the raw materials of mortar were tested to see the effect on porosity, and by extension, strength. Similar to Hanley’s study, it was found that the water to binder ratio had the greatest effect of all the mortars tested. AAdding more water resulted in higher percentage of porosity. The binder to aggregate (or sand) ratio was also observed.
As more sand was added versus binder, the porosity went down considerably, potentially due to the fact that pores tend to form around aggregaates like sannd. This proves that the dry mix has an influence on strength with similar effect to that of water content. Finally, porosity was observed at cure times varying 28 to 730 days. As tthe mortar hydrates, the hydration products fill in the existing pores. This too showed a decrease in porosity.
10 Salt Attack - Introduction
One of the main issues concerning mortar losing strength and degrading over time is due to salt attack. Natural cements were commonly used in coastal areas [9], so knowing how the mortar will behave in these conditions is critical.
Table 2.3: Typical composition of seawater [12]
Table 2.3 shows the different dissolved salts typically found in seawater. Most of them are sulfates and chlorides of magnesium and sodium, which all have some effect on the performance of the mortar due to exposure. How all the ions interact together to create a wear effect is complex, but one can look at the effects of each ion and draw conclusions about the total damage seawater can cause.
Sulfate Attack
The main contributing factor for salt attack is sulfate. Sulfate ions are found in seawater and can cause significant damage to the mortar. The sulfate ions can react with the
11 portlandite or any calcium aluminate hydrate and form gypsum and ettringite [13]. These
new phases are more voluminous than the previous ones, and the expansion can cause
cracking and spalling of the mortar.
There are several contributing factors that determine whether or not a cement or mortar
will have poor durability in the presence of sulfate. Cements containing free lime, like
Rosendale, are prone to disintegration in the presence of sulfate[14] Another factor is if there is any portlandite or calcium aluminate hydrate present to react when exposed to sulfate [12]. These limes and portlandites react with the sulfates and water and undergo expansive, damaging reactions. If these phases are not present in large enough quantities, then no reaction will occur. Portland cement contains little to no free lime, and the
Portland cement industry has developed cements with a low C3A content to minimize this
effect [14-17]. While this is effective today for chemically controlled Portland cements;
as stated earlier, Rosendale cement cannot be chemically controlled so easily. This leads
into investigation into other techniques that could possibly minimize the damage due to
sulfate attack.
Another way to minimize the sulfate attack is through the use of pozzolans in the mix[18]
Pozzolans are silica-rich additives that are used to strengthen mortars. They can be added
to Portland cement mixes, where the silica reacts with the portlandite to form C-S-H gel.
Again, this cannot be controlled in Rosendale cement, but if the cement happened to
contain some silica-rich clay, they could act as natural pozzolans.
12
The sulfate ions are present in seawater; so much of the effect of the sulfate ions is
dependent on how easily the water can penetrate the mortar. This is directly controlled
by the water absorption, permeability, and water vapor transmission of the mortar [19].
A good mortar should not be able to absorb much water so that not much sulfate can
react. Conversely, a mortar with a high permeability and water vapor transmission
allows for any water that does get into the mortar to evaporate quickly, limiting exposure time [20]. While permeability can be good for limiting exposure to sulfate, repeated
soakings and evaporations could have a detrimental effect since salts in the water would
crystallize in the pores. This is later discussed with the chlorine effects.
All of the aforementioned properties also directly relate to porosity, so many of the same
rules apply when making a more durable and stronger mortar. Decreasing porosity will reduce absorption and permeability, which overall is good for the mortar from a sulfate
durability and strength standpoint. Lopez-Arce states that porous materials with high
porosity, a large amount of small pores, and low strengths will be the most prone to salt
weathering [21]. Large volumes of small pores induce capillary suction, drawing in
water, which will lead to the damage because of the large amount of exposed surface
area.
Thermogravimetry is a useful tool to determine if a sulfate attack has occurred in a mortar [22]. It can be used to see at what temperatures the mortar experiences weight
13 loss. Correlating the weight losses to specific reactions at specified temperatures and comparing them to controls could lead to a better understanding to see if new phases have formed, or if existing phases increased.
Figure 2.3: DTG of Portland cement paste hydrated for one week [22]
Figure 2.3 shows the derivative thermogravimetry (DTG) of a Portland cement paste.
The peak labeled “P2” at 121°C corresponds to ettringite. Ettringite is a phase that forms as a result of sulfate and water exposure to portladite. It is an expansive reaction that can damage the mortars. Ettringite dehydrates in the range from 120°C to 150°C, resulting in a weight loss [23]. Since the peak size at that temperature corresponds to how much ettringite dehydrated, one can determine if more ettringite forms by looking at samples both exposed and not exposed to sulfate. If the sulfate exposed sample either has a larger
14 peak on the DTG in that range; corresponding to more weight loss due to ettringite, it can be determined that more ettringite has formed due to the sulfate exposure.
Magnesium Effects
Sulfate attack can be exaggerated or more detrimental if magnesium ions are also present,
which Table 1 shows is true for seawater. Sulfate ions are always present in water as part
of a dissolved salt, and magnesium is a common cation. The presence of magnesium can escalate sulfate attack problems like ettringite formation, as well as create new issues
related to magnesium alone. It has been suggested that the magnesium is the main cation
responsible for deterioration in seawater attack by attacking aluminate and portlandite
phases [24].
Ettringite is typically a hydration product of cement paste [22], so it can sometimes be
hard to determine if more ettringite has formed. Magnesium; however, is not commonly
found, so its presence can be more easily detected.
The presence of magnesium in the seawater can result in a few things. First, the
magnesium will react with calcium hydroxide present in the hydrated mortar to form
brucite - Mg(OH)2 [15,16]. This reaction is also expansive, which can be destructive to
the mortar.
15 Figure 2.4 shows how brucite appears on a TG/DTG graph. Goncalves tested multiple
Portland cement/metakaolin mortar mixes and all of the magnesium exposed samples showed a characteristic hump in the DTG curve between 300 and 400°C. This new peak, labeled by the Mg(OH)2 arrow in Figure 2.4, corresponds to the dehydration of the brucite phase. This was not present before exposure to sea water, so the conclusion was drawn that the brucite forms because of the magnesium exposure.
16
Figure 2.4: TG and DTG of a plain Portland cement mortar (a) before and (b) after 200-
day exposure to 5% MgS solution [16]
17
Another reaction that occurs is a magnesium-calcium ion exchange, with the magnesium
reacting with the hydrated calcium silicates (C-S-H) forming a M-S-H [15,16]. C-S-H
gel is an amorphous phase that gives the cement paste its strength [12]. M-S-H is non-
hydraulic and will cause softening and disintegration of the mortar.
After the calcium undergoes the ion exchange with the magnesium, the calcium would
then leach out into the water. This increased calcium can cause damage to the
surrounding brick structures. [3]
Effect of Chlorine
While magnesium and sulfates are present in seawater, chlorine is also. This chlorine
presence is what differentiates seawater from groundwater [17]. The effects that these ions have on mortars can be altered with the presence of chlorine.
Figure 2.5 shows an experiment where groundwater-soaked specimens expanded more than seawater-soaked specimens. The sulfate content of the water was the same, the only difference being that the seawater also had chloride ions present. Al-Amoudi reports that a chloride presence in water can slow the rate of sulfate attack compared to water with sulfate alone [25]. The results are seen in Figure 2.6.
18
Figure 2.5: Expansion of Portland cement mortars stored in seawater and groundwater
solutions [17]
19
Figure 2.6: Strength reduction in Portland cement mortars [25]
Both authors concluded that the presence of chloride improved the strength and expansion properties compared to simple sulfate attack. This is due to a couple of mechanisms: (i) the solubility of the calcium aluminate hydrate phases increases, which leads to formation of a non-expansive ettringite that would not be as damaging to the mortar; (ii) the chlorides react with the calcium aluminate hydrate phases first, forming
Friedel’s salt (calcium alumino chlorohydrate), decreasing the amount of ettringite that can form. So the presence of chloride in the seawater is generally beneficial in slowing sulfate attack.
20 Chloride ions also have an effect on the magnesium ions as well. If the cement is not
fully hydrated when exposed to the seawater, chloride will promote the formation of a porous C-S-H gel [15]. This resulting porosity makes it easier for the magnesium to
penetrate the mortar and react to form the M-S-H gel, decreasing the strength of the
mortar. If; however, the mortar is mostly hydrated, it has been observed that saltwater with chloride ions tend to form a thicker brucite layer on the surface of the mortar versus
a groundwater without the chloride [17]. Since this layer is on the surface, it is not
damaging to the mortar, and provides a barrier layer protecting the mortar from direct
exposure to the seawater, slowing further penetration.
While this formation of Friedel’s salt (calcium aluminum chloride) can slow the effects
of sulfate attack, it is not without problems itself. These salts, as with the ettringite, will
crystalize in the pores, and, over time, will eventually build up enough to cause damage
by exerting pressure on the pore walls and putting stresses on the mortar. This, again,
comes back to porosity. Crystallization pressure, or the pressure these salts are putting on
the pore walls, is inversely related with pore diameter [2]. Mortars with a large volume
of small pores will see more damage than those with larger pores. Mortars with small
pores will be subjected to much higher crystallization pressures with the capability of
destroying the mortar itself.
In addition to Friedel’s salt, sodium chloride and potassium chloride can also form along
the pore walls [26]. This can cause irreversible dilation of the mortar and will also
21 crystallize in layers on pore walls. Diffusion rates of the salts are temperature dependent, and over time and season changes the salts can penetrate deep into the mortars, causing damage throughout the material and structure [27].
One way to tell if the mortar has taken in chloride would be through the use of ion chromatography [21,28]. The soak water can be tested, and if ion concentrations go down over time, then it can be assumed that the mortar has taken in the chloride.
22 CHAPTER 3
EXPERIMENTAL PROCEDURES
Description of Materials and Mixes
Mortar cubes were fabricated as per ASTM C109 using the following materials:
Rosendale cement from Edison Coatings, standard grated sand (meeting ASTM C7781),
Virginia Lime Works lime putty, and Greymont dolomitic lime. Tap water was used for the water content.
Seven unique mixes were made and are shown in Table 3.1. There were two mixes without any lime, two with the lime putty, and two with the Greymont lime. Mix C was fabricated as per ASTM C10 but with a graded sand instead of the specified 20/30 sand to better resemble the materials that would be used in restoration.
1 ASTM refers to the American Society of Testing Materials
23 Table 3.1: Description of mixes used in experiments in weight percent
Mix Rosendale, Lime (Type) Graded Volume as- [wt%] Sand, as- proportions received, received dry [cement:lime:sand] dry [wt%] [wt%] C 50 0 50 1:0:0.65 1 21.1 0 78.9 1:0:2 ¼ 2 26.4 12.6 (As- 61.0 1:½:1½ (dry) received Lime Putty, wet) 3 19.85 19.05 (As- 61.10 1:1:2 (dry) received Lime Putty, wet) 4 28.25 6.37 65.37 1:½: 1½ (Greymont, in bag, dry) 5 22.07 9.96 67.96 1:1:2 (Greymont, in bag, dry)
Fabrication of Mortars
To fabricate the mortar cubes, first a 2000g batch of the dry ingredients (cement, sand,
lime) was mixed in a Hobart N50 stand mixer on level one. A full graduated cylinder
(1000mL) was tarred on a scale. Water was then slowly added while mixing on level one until it reached a wet, but clumpy consistency, as seen in Figure 3.1. The mixer was then turned off and the mortar was left to soak up the water for one minute. After one minute the mortar was mixed by hand to ensure that nothing had clumped to the bottom. Then the mixer was turned back on and water was slowly added until it reached a workable consistency. The water weight was recorded. The mixer was then turned to level two to
ensure that the mortar had been thoroughly mixed.
24
Figure 3.1: Addition of water to dry mortar mix
After the mixing, the mortar was tested on a Vicat penetrometer to ensure the right consistency was reached. It was ideal for the penetrometer displacement to fall between
2-4mm. If the displacement was less than 2mm, the mortar was returned to the mixing bowl and more water was added. If the displacement was more than 4mm, the mortar needed to be remade with less water.
25
Figure 3.2: Placement of the wet mortar mix into the molds
After the penetrometer test the mortar was moved back to the mixing bowl. Then the mortar was put into the cube molds. Figure 3.2 shows the mortar being put into the molds. The molds were 2” cube disposable plastic molds. After each cube is made, they were placed into a 100% humidity chamber as shown in Figure 3.3.
26
Figure 3.3: Fresh mortar cubes in the initial humidity chamber
After one week, the cubes were removed from the initial humidity chamber, taken out of
the molds and moved to a Lingl experimental drying oven set at 90% humidity and 100°F
to ensure consistent curing for all samples. They were left there another twenty-one days
to achieve a total curing time of twenty-eight days.
Compressive Strength Test (ASTM C109/C10)
The compressive strengths of the mortar cubes were tested at 7-day intervals throughout
the curing process. At each interval, three samples were randomly selected. The dimensions were measured and the cubes were loaded into the compression testing machine with the exposed mold face toward the user for consistency. They were tested
27 on a Satec testing machine with a 400K-load cell at a loading rate of 200 lbf/s. Figure 3.4 shows one of the mortar samples being tested.
Compressive strength was calculated using the following equation:
fm = P/A
Where fm is the compressive strength, P is the total maximum load, and A is the area of the loaded surface.
28
Figure 3.4: Compressive strength testing of 2”x2” mortar cubes
Tensile Splitting (ASTM D3967)
Three two-inch diameter cylinders of each mix were fabricated at the same time using the same method as the cubes. After 28 days of curing, the dimensions were measured and tested on the Satec. The splitting tensile strength is calculated using the formula:
2 /
29 Where is the splitting tensile strength, P is the maximum applied load, L is the
thickness of the sample and D is the diameter
Bond Wrench (ASTM C1072)
Two-brick piers were manufactured using Old Carolina pressed brick and mix 2. Twenty piers were fabricated. They were then covered with a moist paper towel and left to cure for 28 days. Once cured, a pier was loaded into the bond wrench apparatus. A wrench
was used to tighten the apparatus until the two bricks became disjointed as shown in
Figure 3.5. The force and from which brick it was removed from was recorded.
30
Figure 3.5: A brick couplet in the bond wrench apparatus after failure
Pier Compression
Four-brick piers were manufactured and cured in a similar fashion as the two-brick piers.
Eight samples each of mixes C and 2 were made. Once cured, they were capped using sulfur and left for 24 hours. Then the individual piers were measured, weighed and
31 underwent a compressive test on the Saetec as shown in Figure 3.6. The breaking load was recorded.
32
Figure 3.6: A brick pier following the pier compression test
33 Water Vapor Transmission Test (ASTM E96)
For this test, three samples of each as-cured mortar were randomly selected. They were
cut to a thickness of 20 mm with the mold-exposed face being the top. Dimensions and
weights were recorded.
Figure 3.7: Water vapor transmission testing apparatus
The mortar samples were then mounted into 4” x 4” square Ziploc container lids using
hot glue to make it air and watertight as shown in Figure 3.7. The containers were filled
with deionized water to a depth of ½” in the container and sealed with the cube-mounted
lid. The cube was not in direct contact with the water. The whole apparatus was then weighed and recorded along with the current temperature and humidity. In addition to
the samples, there was one dish that had no cube mounted for a control. The cube- mounted dishes were weighed every twenty-four hours until a trend could be established.
Water vapor transmission could then be calculated according to the formula:
34
Where WVT is the water vapor transmission, G is the weight change from the original
apparatus, T is time, and A is the cross-sectional area.
X-Ray Diffraction
X-ray diffraction was performed on both the raw materials and the mixed mortars. A dry,
finely ground sample was placed in a 1”x1” sample holder in a Scintag PAD-V X-Ray
Diffraction Unit with a copper radiation of 1.54 angstroms. The test was performed at a
2-theta range from 5° to 65° with a step of 0.2° and a dwell time of four seconds. The
resulting diffraction patterns were analyzed using Jade08 software.
Thermogravimetry and Differential Scanning Calorimetry
Thermogravimetric (TG) and differential scanning calorimetriy (DSC) measurements
were carried out simultaneously on a Netzsch STA 449C coupled to a Brucker Vector 22
Fourier transform infrared spectrometer (FTIR) for the evolved gas analysis (EGA). The
data was gathered at a heating rate of 10°C/min and simulated air (20% oxygen in
nitrogen) flowing at 100 mL/min. The samples were 50 mg (±2.5mg) in alumina
crucibles that were calibrated using a sapphire disk.
For the samples that did not have a DSC measurement run simultaneously, larger samples
were used to improve accuracy. Ground samples of 1000 mg (±2.5mg) were used in a
larger alumina crucible.
35
Mercury Intrusion Porosimetry
Mercury intrusion porosimetry (MIP) was performed on each of the mortar mixes using a
Quantachrome Pore Master 33. Samples of around one gram were used. MIP uses the
Washburn equation:
4γcosθ d P
Where d is the apparent pore diameter, γ is the surface tension of mercury, θ is the contact angle between the surface and mercury, and P is the intrusion pressure.
Initial Rate of Absorption
To begin this test, three samples of each cured mix were randomly selected and weights
were recorded. They were then placed in a pan raised with glass stirring rods. Deionized
water was then added until it reached a height of ¼” up from the base of the cubes. The weights were the recorded at 15 minutes, 30 minutes, one hour, two hours, four hours, six hours, and 24 hours. The results were plotted at weight percent water absorbed.
X-Ray Flourescence
A dry sample of each mix is placed into a boat and fired to 1000°C to ensure everything
had been oxidized. The weights before and after firing were recorded so that the loss on
ignition could be calculated by dividing the difference between the dry and fired weights
by the dry weight. The fired samples were then prepared to be fused. Dry, finely ground
36 sample powder (0.750g) was mixed with a flux (lithium borates, 6.00 g) and oxidizer
(ammonium nitrate, 1.0x g) in a platinum-gold alloy crucible. The crucible was then
placed in a Claisse M4 Fluxer and run so that the powder is formed into a glass disk.
Once completed, the disk was placed in a ThermoNoran QuanX EC, which was
calibrated to a copper standard and run to produce the x-ray fluorescence results.
Salt Leaching Experiments
Six as-cured samples of each mix were weighed and measured. They were placed in rigid plastic containers and 300g of a simulated seawater solution was added to each of the containers. After a week a milliliter of the leachate from each mix was diluted and tested using both ion chromatography and ICP. The samples were left in an ambient atmosphere for two weeks to equilibrate and then were weighed and measured. Samples were measured at 7, 28, 60, 90, and 180 day soak times.
37 CHAPTER 4
RESULTS AND DISCUSSION
Rosendale Characterization
Table 4.1 shows the oxidized chemistries of the Rosendale cement used in the mortar
mixes. The Rosendale was mostly comprised of calcium, followed by silica and
magnesia. This is consistent with what literature has studied in the past, but with higher
calcium content than previously reported.
38 Table 4.1: Oxidized chemistry of raw materials
Rosendale Major Constituents
Al2O3 % 4.65 SiO2 % 20.35 Na2O % <0.5 K2O % 1.04 MgO % 11.33 CaO % 56.57
TiO2 % 0.27 MnO % 0.30
Fe2O3 % 1.89 P2O5 % <0.05 S % 3.02 Sum of Major 99.41 Constituents % Minor Constituents Cl ppm 570 V ppm <150 Cr ppm 489 Ni ppm 368 Cu ppm 28 Zn ppm 24 As ppm <20 Rb ppm <30 Sr ppm 844 Zr ppm 34 Ba ppm <200 Pb ppm <50
39 X-ray diffraction showed that the Rosendale cement contained calcite, portlandite, quartz, belite, lime, and anhydrite. Again, this is fairly consistent with what has previously been seen in Rosendale. Free lime is common in mortars, although typically it combines with the belite at high temperatures to form alite. So seeing free lime and belite in a lower fired cement is not unexpected.
X-Ray Diffraction Mix Characterization
X-ray diffraction identified all the same phases present in all the mixes. The mixes showed after 28 days of curing that they were all reacting to form the same phases. The phases present included quartz, calcite, portlandite, belite, periclase, merwinite (a calcium magnesium silicate phase), bassanite (a hemihydrate of gypsum), and gehlenite (a calcium aluminosilicate phase). These can be seen in Figure 4.1 There could also be some brucite, but the major peak is difficult to detect since it is at a low angle. C-S-H gel is the phase that gives the mortars their strength, but since it is amorphous, it cannot be detected with XRD. The presence of hemihydrate and portlandite phases could lead to a potential sulfate attack. It could react with sulfates to produce ettringite. [12]
40 [Mix C.MDI] 4000
3000
2000 Intensity(Counts) 1000
0 04-008-7651> Quartz - SiO 2
04-012-8072> Calcite - Ca(CO 3)
04-007-5231> Portlandite - Ca(OH) 2
00-024-0034> Calcio-olivine - Ca 2SiO 4
04-007-8540> Larnite - Ca 2(SiO4)
04-004-7535> Periclase - MgO
00-035-0591> Merwinite - Ca 3Mg(SiO4)2
00-041-0224> Bassanite - CaSO 4·0.5H 2O
04-011-9811> Gehlenite - Ca 2Al 2.22Si 0.78O6.78(OH)0.22
10 20 30 40 50 60 Two-Theta (deg)
Figure 4.1: X-ray diffraction pattern of Mix C
X Ray Flourescence Characterization
Table 4.2 highlights the major constituents of the elements found in the mortar mixes.
The chemistry reflects the chemistry and percentages of the dry mix of raw materials.
The silica content directly correlates to the amount of sand in the dry mix, since the sand is the main contributor to the silica. The Rosendale cement and limes had high calcium percentages, so the calcium content relates to those as well.
41 Table 4.2: Oxidixed chemistry of mortar mixes
Major Constituents C 1 2 3 4 5
Al2O3 % 2.75 1.25 1.78 1.59 1.76 1.42 SiO2 % 63.92 85.10 74.62 75.14 74.40 75.81 Na2O % <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 K2O % 0.42 0.13 0.25 0.10 0.16 0.17 MgO % 4.84 1.83 3.51 3.21 4.68 4.99 CaO % 25.10 10.23 17.97 18.20 17.13 15.95
TiO2 % 0.12 0.05 0.07 0.08 0.06 0.06 MnO % 0.13 0.05 0.07 0.07 0.08 0.05
Fe2O3 % 0.85 0.36 0.51 0.43 0.51 0.42 P2O5 % <0.05 <0.05 <0.05 0.05 <0.05 <0.05 S % 1.32 0.56 0.76 0.66 0.76 0.66 Sum of Major Constituents % 99.45 99.57 99.55 99.53 99.53 99.55
Table 4.3: Loss on ignition for the tested mixes
Mix LOI [wt%] C 10.89 1 5.57 2 8.77 3 8.15 4 8.60 5 8.59
42 The loss on ignition was performed prior to the x-ray fluorescence tests. The main contributors to weight loss are dehydroxylation and carbonate decomposition of various phases. The results are presented in Table 4.3
The loss on ignition seems to have some correlation with cement content as well. This is difficult to determine, since there is burnout from the lime as well. But by comparing the pairs of mixes with the same materials, the ones with lower cement contents also have lower weight losses due to ignition.
Simultaneous Thermal Analysis
While loss on ignition is a good indicator of weight loss, simultaneous thermal analysis gives a better depiction of when these losses occur. Figure 4.2 shows the complete STA of Mix C.
43 H2O TG /% DSC /(mW/mg) CO2 [1] TG Mass Change: -0.27 % [1] DSC Peak: 758.3 °C [1] CO2 Peak: 759.5 °C exo 100 [1] TG Mass Change: -1.04 % Carbonate Decomposition (Dolomite) 0.40 [1] H2O Peak: 393.6 °C [1] TG Mass Change: -2.16 % 0.5 20
98 [1] H2O Peak: 457.3 °C 0.30
[1] TG Mass Change: -0.90 % [1] DSC Peak: 454.9 °C 0.4 15 Quartz Inversion [1] DSC Peak: 576.0 °C 0.20 96 [1] H2O Peak: 191.4 °C
[1] DSC Peak: 392.0 °C [1] 0.3 0.10 10
94 [1] DSC Peak: 182.8 °C
[1] TG Mass Change: -5.22 % 0.2 0 5
[1]
92 Calcium Carbonate Decomposition [1] CO2 Peak: 847.1 °C -0.10 [1] CO2 Peak: 399.2 °C 0.1 0
[1] TG Mass Change: -0.76 % 90 -0.20 [1] 0.0 200 400 600 800 1000 Temperature /°C
Figure 4.2: Mix C TG, DSC, water, and carbon dioxide analysis after 28 days curing
Figure 4.2 shows the steps of weight loss over a heating of the mortar. The water peak and weight loss at around 190°C is most likely a result of the ettringite going through a dehydroxylation. This is consistent with what was seen previously in literature[23].
Since there was a large amount of free magnesia in the cement, it is possible that brucite had formed, shown with the weight loss and water peak at 386°C. The weight loss and water peak at 446°C is consistent with that of portlandite decomposition, which is both present in the mortar and a by-product of the belite hydration reaction into C-S-H.
Because there was also free lime in the cement, a large portion of the weight loss is due to the carbonate decomposition of the calcium carbonate phase in the lime putty.
44 Figure 4.3 shows a comparison of all the Mixes’ thermogravimetric results. Similar to the loss on ignition, it is apparent that Mix 1, which had the least cement, lost the least weight; while Mix C, with the most cement, lost the most. All of the cement-lime mortars (Mixes 2-5) were very similar in the stages and the amounts of weight lost.
TG /%
100
98
[1] Mix C_22July11.dsu 96 TG [2] Mix 1_25July11.dsu TG [2] [3] Mix 2_26July11.dsu 94 TG [4] Mix 3_26July11.dsu TG
[5] Mix 4_28July11.dsu [5][4] 92 [6] TG [3] [6] Mix 5_28July11.dsu TG
90 [1]
200 400 600 800 1000 Temperature /°C
Figure 4.3: Thermogravimetry of all tested mixes
45 Porosity
Mercury intrusion porosimetry tests yielded results seen in table 4.4.
Table 4.4: MIP results for all tested mixes
Mix Property Unit C 1 2 3 4 5 Total Intrusion Volume ml/g 0.154 0.184 0.153 0.196 0.152 0.170 Median Pore Diameter µm 0.235 1.093 0.248 0.336 0.287 0.297 Bulk Density g/cc 1.84 1.74 1.86 1.77 1.85 1.75 Apparent Density g/cc 1.87 1.82 1.88 1.92 1.89 1.79 Porosity % 28.5 32.1 28.7 34.8 28.2 29.8 Pores >10 Microns ml/g 0.008500.0283 0.00650 0.0470 0.0113 0.0113 Pores 10-1 Microns ml/g 0.01520.0687 0.00677 0.00951 0.0121 0.00722 Pores <1 Microns ml/g 0.131 0.0877 0.141 0.140 0.129 0.152
Table 4.4 gives a great deal of information that will later relate to many other properties.
Mixes C, 2, and 4 all had lower porosity than their counterpart mixes with less cement.
Mix C had a lower porosity than Mix 1, both being cement-sand mortar mixes with no lime, Mix 1 having less cement. Mixes 2 and 3 were the mixes with lime putty; Mixes 4 and 5 with the Greymont powder lime. Like with C and 1, Mixes 3 and 5 had less cement than 2 and 4, respectively. This suggests that in similar mix designs, a mix with more cement would result in a lower porosity as seen in Figure 4.4.
46 Cement content and porosity comparison
36
34
32
30 Percent Porosity [%]
28
26 15 20 25 30 35 40 45 50 55 Weight percent cement [%]
No lime mortars (C & 1) Lime putty mortars (2 & 3) Greymont lime mortars (4 & 5)
Figure 4.4: Comparison of cement content with porosity with different mix types
Similar trends were seen in bulk density in figure 4.5. Mixes C, 2, and 4 all had densities in the 1.84-1.86 g/cc range, while 1, 3, and 5 had lower densities in the range of 1.74-1.77 g/cc. This is consistent with the porosity trends seen in Table 4.4, since bulk density includes pores.
47 Cement content vs. bulk density
1.88
1.86
1.84
1.82
1.80
1.78 Bulk density [g/cc]
1.76
1.74
1.72 15 20 25 30 35 40 45 50 55 Cement content [wt %]
No lime mortars (C & 1) Lime putty mortars (2 & 3) Greymont lime mortars (4 &5)
Figure 4.5: Comparison of different mix types’ cement contents with bulk density
Compressive Strength
One way to have a strong mortar that has over a long period of time is to have a stronger
mortar at the time of fabrication. Figure 4.6 shows how the different mortar mixes
performed over seven-day intervals. Mixes C and 1 have 7-day compressive strength
values at zero because they had not yet acquired enough stiffness and strength to be
demoulded. Mix C was consistently the strongest while Mix 1 was the weakest. This
would suggest that the amount of cementicious material in the mortar plays a large role in
the strength of the mortars. Also, the lime putty mixes (2 and 3) had higher strength
48 values than Mixes 4 and 5, respectively. This could mean that the lime putty has better hydraulic properties than the Greymont lime.
Strength vs. Curing Time
1000
800
600
400
200 Compressive Strength [psi]
0
5 1015202530 Cure Time [days]
Cure time vs C Cure time vs 1 Cure time vs 2 Cure time vs 3 Cure time vs 4 Cure time vs 5 Cure time vs 6 Cure time vs 7
Figure 4.6: Compressive strength over 28 days of curing
Figure 4.7 illustrates what effect cement content had on the 28-day compressive strength
of the mortars tested. There is a general positive correlation between the weight percent
of cement and the strength. As mentioned earlier, the addition of different limes could
also play a role in the strength, leading to some variability in the results. Also, the use of
49 graded sand yielded results lower than that specified in ASTM C10, showing that the coarser sand aggregates typically yield better results. The precision in the data falls within the range of 21.8%, which is reported as typical by ASTM C109.
50 Cement Content vs. Strength Comparison
900
800
700
600
500
400
300
28 Day Compressive Strength [psi] Strength Compressive Day 28 200
100 15 20 25 30 35 40 45 50 55 Weight Percent Cement [%]
Figure 4.7: Comparison of cement content and compressive strength
51 Strength vs. Porosity Comparison
900
800
700
600
500
400
300
28 Day Compressive Strength [psi] 200
100 26 28 30 32 34 36 Porosity [%]
Figure 4.8: Comparison of porosity and strength in tested mixes
Figure 4.5 shows how the tested mortars relate strength and porosity. Again, a general inverse correlation can be drawn, but the variability in the mixes’ designs and limes could be causing some variability in the data. This relationship is similar to that seen in other cement mixes in literature[11].
Water Vapor Transmission
Figure 4.9 illustrates how each mix performed in the water vapor transmission test. Mix
C performed the best, transmitting the least water, followed by Mixes 5 and 2. Mix 1 was shown to be the poorest performer in this test. The disconnect with Mix 1 and the others
52 could be due to Mix 1 having the largest volume of pores greater than ten microns, as shown in Figure 4.10. This evidence could lead to the conclusion that water vapor travels more easily in larger pores versus a more tortuous path of small pores. The graph seems to reach a maximum and then taper off, which suggests that there is a volume where the water vapor transmission is the worst, but having large pore volumes smaller would result in less water vapor transmission.
Water Vapor Transmission
8
6
4 WVT [g/s-m^2] WVT
2
0 C12345 Mix
WVT
Figure 4.9: Results of the water vapor transmission test for each mix
53 Water Vapor Transmission vs. Large Pore Volume
8
7
6
5
WVT [g/s-m^2] WVT 4
3
2 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Volume of Pores > 10 microns [ml/g]
Figure 4.10: Comparison of water vapor transmission and volume of large pores
Initial Rate of Water Absorption
Water absorption is an important property for seawater durability. Having a slower rate
of absorption, as well as having a lower total water uptake after 24 hours can limit
seawater damage.
Figure 4.11 shows that mixes 1, 4, and C had the highest initial rates of absorption,
respectively. Mixes 2, 5, and 3 yielded better results by having lower absorption rates.
This is seen by looking at the slopes of the lines in the early water uptakes, before the
samples start to become saturated and the weights level out.
54 20.00
18.00
16.00
14.00 C absorbed
12.00 1 10.00 water 2 of
% 8.00 3 6.00 4 Weight 5 4.00
2.00
0.00 0 4 8 12162024 Time [h]
Figure 4.11: Water absorption for the mortar mixes over a 24-hour period
It appears that the initial rate of absorption has an inverse relationship dependence on the
volume of pores less than a micron. Mixes 1, 4, and C also were the mixes with the
smallest volume of sub-micron pores with volumes of 0.8772, 0.1287, and 0.1309 ml/g,
respectively. Mixes 2 and 5 had the two slowest rates, and also had the two highest
volumes of sub-micron pores. Originally, one would have thought the capillary suction of sub-micron pores would lead to faster rates of water absorption, but the results from this test suggest the opposite.
55 One possible explanation could come from looking at the volume of pores between one
and ten microns. Mix 1 has an order of magnitude larger volume of pores in comparison
to Mixes 2 and 5. All the volumes between one and ten micron closely correlate with the
rates of absorption. It could be concluded from the results of this data that pores within that range are more responsible for the initial rate of water absorption than the sub- micron pores.
Summary of Mechanical Testing
Table 4.5 summarizes the results from various tests with the compositions of the mixes.
The green cells highlight the mix that performed the best at a given experiment, while the
yellow cells signify the mixes that were second and third in the desired result.
Table 4.5: Summary of results sorted by mix
28- Wt% day WVT water Lime Graded Porosity CS [g/s- absorbed Mix Cement (Type) Sand (%) [psi] m2] (24hr) C 50 0 50 28.46 768.37 3.2 18.07 1 21.1 0 78.9 32.14 197.01 6.08 16.19 12.6 (Lime 2 26.4 putty) 61 28.65 563.83 4.33 13.16 19.05 (Lime 3 19.85 putty) 61.1 34.81 471.92 5.05 16.04 6.37 4 28.25 (Greymont) 65.37 28.2 516.15 4.93 18.91 9.96 5 22.07 (Greymont) 67.96 29.83 381.46 3.81 16.02
56
Looking at Table 4.5, Mix C was the best performer in compressive strength and water
vapor transmission. It also had the second lowest porosity. Mix 2 was also a good
performer in the tests in table 4.5, falling in the top half of every category and the best in
water absorption.
Because these two mixes could be considered the best performers of the group, they were chosen to undergo further testing with the bond wrench test. Elimination of the other mixes was necessary to be economically feasible and ensure good sample sizes for the tested mixes.
Bond Strength
Two brick couplets were used in the bond strength tests. Molded brick was used to
simulate the type of bonds that would be seen in historical walls.
Table 4.6: Results of the bond strength test
Flexural tensile Mix strength [psi] St. dev. C 34 22.6 2 94 22.0
Table 4.6 illustrates the results of the bond strength test. There was quite a bit of
variation in this test due to the nature of the test itself as well as some slippage of the
apparatus. Nevertheless, it can be said with a 95% level of confidence that the two
57 average flexural tensile strengths are significantly different. As stated in literature, one
possible reason for this difference is because the additional lime increasing the bond strength.
Pier Compression
Mix 2 was used in the pier compression tests because of its good bond strength. The four brick piers were tested, and the average pier compressive strength was calculated to be
2412±176 psi. This test was considered a success since the brick failed before the mortar did, so the mortar can withstand a compressive stress in a wall to at least the threshold of the bricks’ strength.
Magnesium Sulfate Attack
Figure 4.12 shows the sulfate concentration remaining in the soak water over a 180-day
period. The general trend suggests that the sulfate concentrating in depleting in the water
over the soak time. This would imply that the mortar is taking in the sulfate ions and
reacting with the mortar.
58 Sulfur Ions
3000
2800
2600
2400
2200
2000
1800
1600 Soluble Ions [microgram/gram] Ions Soluble
1400
1200 0 20 40 60 80 100 120 140 160 180 200 Time [days]
Mix C Mix 1 Mix 2 Mix 3 Mix 4 Mix 5
Figure 4.12: Sulfate ion concentration in seawater over a 180-day soak period
Another way of determining if sulfate attack has occurred is by looking at thermogravimetry and water evolution results [22]. Figures 4.13-15 show a Mix C sample that was not soaked in the seawater solution, one sample after soaking in the seawater solution for 180 days, and a comparison of the two.
59 DTG /(%/min) TG /% H2O H2O Peak: 163.3 °C 100 [1] 0.0 TG Mass Change: -3.29 %
98
TG Mass Change: -1.75 % 15 -0.1 96
94
-0.2 H2O Peak: 417.3 °C 10
92
90 -0.3
5
88
-0.4 86 [1] MixC_unsoaked_17Oct12.dsu TG 0 H2O [1] 84 DTG -0.5
[1] 82 200 400 600 800 1000 Temperature /°C
Figure 4.13: TG, derivative TG, and water analysis of an unsoaked Mix C cube
One interesting side note about Figure 4.10 is that even though it is the same mix, the results look different than in Figure 4.2 which was collected over a year previously. The portlandite peak in the water analysis had vanished. It could be due to a pozzolonic reaction, or a slow hydration reaction, but the precise reason for this phenomenon is undetermined.
60 DTG /(%/min) TG /% H2O H2O Peak: 139.2 °C 100 0.00
[1] TG Mass Change: -3.84 %
98 -0.05 20
96 TG Mass Change: -2.04 % -0.10
94 15 -0.15
92 H2O Peak: 391.3 °C -0.20
10 90 -0.25
88 -0.30 5
-0.35 86 [1] MixC_180_17Oct12.dsu TG H2O DTG 84 0 -0.40
[1] [1]
200 400 600 800 1000 Temperature /°C
Figure 4.14: Mix C TG, DTG, and water analysis after 180-day seawater soak
61 DTG /(%/min) TG /% H2O
100 [1] 0.0 [2]
98 20 -0.1 96
94 15 -0.2
92
10 90 -0.3
88 [1] MixC_unsoaked_17Oct12.dsu TG 5 H2O -0.4 86 DTG [2] MixC_180_17Oct12.dsu TG H2O 84 DTG 0 [1][2] -0.5 [2] [1] 82 200 400 600 800 1000 Temperature /°C
Figure 4.15: Comparison of Mix C unsoaked (green) and soaked (red) TG/DTG and
water analysis
The first water peak indicates the presence of ettringite in both samples. The water peak between 100-200°C for the seawater soaked sample is larger than the control sample peak. This would imply that more ettringite formed in the seawater sample than a sample not exposed to seawater. The area under the curve in the derivative of the TG indicates the weight loss at any particular step. Since the downward peak in the 100-200°C range is larger in the seawater soaked sample, it adds more evidence to suggest that more ettringite formed due to the sulfate that was taken up by the mortar.
62 Magnesium ions had a similar result. Figure 4.16 shows that the magnesium concentrations in the seawater also went down with soak time. This also means that the mortar took in magnesium. This data resembles more of a decay-type trend than the sulfate. This could mean that the mortar and the seawater were approaching equilibrium.
Magnesium Ions
3500
3000
2500
2000
1500
1000
500 Soluble [microgram/gram] Ions
0
0 50 100 150 200 Soak Time [days]
Mix C Mix 1 Mix 2 Mix 3 Mix 4 Mix 5
Figure 4.16: Magnesium ion concentration in seawater over a 180-day soak period
63 Further evidence of magnesium uptake is seen in the thermogravimetric graphs in Figure
4.15. The water and DTG peaks around 400°C could be indicative of brucite formation.
The higher water peak and weight loss suggests that some brucite formed during the soaking period. The brucite could be forming on the surface of the mortar, but if it forms
beneath the surface it would undergo an expansive reaction that could be detrimental to existing brick and mortar in a restoration.
Magnesion Ions Varying Lime Putty Content
3500
3000
2500
2000
1500
1000
500 Soluble Ions [microgram/gram] Ions Soluble
0
0 20 40 60 80 100 120 140 160 180 200 Soak time [days]
Mix C (0 wt% lime putty) Mix 2 (12.6 wt% lime putty) Mix 3 (19.05 wt% lime putty)
Figure 4.17: Concentration of magnesium ions varying lime putty content
Figure 4.17 highlights the magnesium ion content with increasing additions of lime putty.
It appears that the more lime putty that is presents in the mix, the faster the magnesium
ion depletion. Because the lime putty is so high in calcium, this observation could be due
64 the increased calcium content. In addition to forming brucite, the magnesium could be
undergoing an ion exchange with the calcium in the C-S-H gel, which would result in a
softening and loss of strength.
Figure 4.18 shows the concentration of calcium varying lime putty content. While it is difficult to establish trends, the mix with the highest lime putty content generally released more calcium into the leachate than the other mixes with little to no lime putty. This could be a confirmation of the ion exchange with the magnesium. As magnesium ions
were being taken in by the mortar, some calcium ions were being released into the leachate once exchanged. The more calcium a mortar had could yield a greater potential for ion-exchange.
65 Calcium ion concentrations varying lime putty
3500
3000
2500
2000
Soluble Ions [microgram/gram] Ions Soluble 1500
1000 0 20 40 60 80 100 120 140 160 180 200 Soak time [days]
Mix C (0 wt% lime putty) Mix 2 (12.6 wt% lime putty) Mix 3 (19.05 wt% lime putty)
Figure 4.18: Concentration of calcium ions varying lime putty content
Chloride Effect
Similar to the magnesium ion trend, the chloride ion concentration in the seawater
underwent a decay-type regression, as seen in Figure 4.19. This could also mean the mortar and seawater reaching equilibrium. Since the salts believed to have formed are
not hydrates or carbonates, it would be difficult to determine through thermogravimetry
or evolved gas analysis the existence of these salts.
66 Chloride Ions
40000
38000
36000
34000
32000
30000
28000 Soluble ions [microgram/gram]
26000
24000 0 20 40 60 80 100 120 140 160 180 200 Soak time [days]
Mix C Mix 1 Mix 2 Mix 3 Mix 4 Mix 5
Figure 4.19: Chloride ion concentration in seawater over a 180-day soak period
67 All of the ion uptakes (sulfate, magnesium, chloride) were difficult to distinguish between the various mixes. All of the mixes took in the ions, so any of these mixes could be venerable to seawater attack.
68 CHAPTER 5
CONCLUSIONS
The Rosendale cement tested yielded results in the physical properties tests similar to that found in literature for other cements and mortars. It was successfully shown that the strength increases with cement content in a given mix, and bond strength between mortar and brick improves with the addition of lime. Further, it was shown that many properties
were in some way dependent on porosity. Strength and porosity were shown to be
inversely related, as seen in literature. Water vapor transmission also showed some
relationship with the unit volumes of large pores; and initial rate of absorption was closely related to the volume of pores between one and ten microns.
The results from the seawater tests proved to be very interesting. The mortars took up sulfates, and thermogravimetry also showed that ettringite was forming. Similar results were seen with magnesium - the mortars taking up the ions and thermogravimetry indicating brucite formation. The magnesium could also be undergoing an ion exchange with the calcium, causing further damage. Chlorides were also taken up, which could be beneficial or also harmful. Studies would have to be done comparing sulfate exposure with and without chloride to see if there is a positive or negative effect. Also repeatedly exposing the mortars to new seawater environments would better simulate reality and their behavior could be observed to see if any damage or strength loss presents itself.
69 This could also show more pronounced differences with different mixes. Testing with
brick couplets would be good to see the effects on the bricks as well.
Other testing that would also be beneficial would include running any of the tests done in this experiment with a variety of different Rosendale batches to see how consistently they perform.
Overall, this project was a success. Physical properties were tested, and the results coincided with either what had been seen in literature, or could be explained through other physical relationships. The mortars took up common seawater ions that can cause damage, exposing potential problems in the future. Knowing that Rosendale cement is not impervious to seawater issues is the first step to successfully understanding it and
being able to successfully use it in historical restoration applications.
70 APPENDIX A
RAW DATA
Table A.1: 7-day compressive strength
Mix Number Weight [g] D1 [in] D2 [in] Peak Load [lb] Pressure [psi] 1 250.4 1.9815 1.9765 2018 515.2645528 C 2 251.6 1.977 1.992 1840 467.2204244 3 252.2 1.9685 1.988 1969 503.1458755 1 200 2 2 0 0 1 2 200 2 2 0 0 3 200 2 2 0 0 1 250.1 1.9865 1.97 405 103.4904386 2 2 248.9 1.9905 1.97 419 106.8527281 3 243.5 1.995 1.9785 338 85.63232696 1 200 2 2 0 0 3 2 200 2 2 0 0 3 200 2 2 0 0 1 253.3 1.9755 1.9895 486 123.6560279 4 2 252.5 1.985 1.987 404 102.4290128 3 256.3 1.991 1.9995 448 112.5344119 1 255.6 1.99 1.986 452 114.3684181 5 2 254.4 1.9815 2.0035 436 109.8254688 3 252.7 1.9845 1.999 490 123.5185494
71 Table A.2: 14-day compressive strength
Mix Number Weight [g] D1 [in] D2 [in] Peak Load [lb] Pressure [psi] 1 231.2 1.9945 1.992 3022 760.6258577 C 2 232.4 2.012 1.999 2802 696.6704028 3 232.2 2.0165 1.9995 2936 728.1760931 1 228.5 1.9535 1.9975 602 154.2752603 1 2 229.1 1.923 1.9985 483 125.6792829 3 231 1.9685 1.9955 562 143.0701934 1 222.4 1.9715 1.992 1444 367.6893728 2 2 223.6 1.982 2.008 1470 369.3600975 3 220.8 1.9875 1.995 1349 340.2216232 1 219 1.9725 1.9705 1085 279.1491354 3 2 216.4 1.9875 1.9785 1034 262.9525258 3 219.2 1.9825 1.99 1047 265.387467 1 221.5 1.975 2.0035 1356 342.6914292 4 2 227.1 1.9925 1.993 1830 460.8350053 3 225.4 2.0005 2.005 2026 505.1106301 1 222.2 1.985 2.013 1625 406.676502 5 2 221.9 1.9775 2.0005 1774 448.4340356 3 222.2 1.975 1.982 1667 425.8580388
72 Table A.3: 21-day compressive strength
Mix Number Weight [g] D1 [in] D2 [in] Peak Load [lb] Pressure [psi] 1 229.4 1.9845 1.999 3126 787.9979294 C 2 229.1 1.9905 2.005 3448 863.9541563 3 227.6 1.9655 2.029 3210 804.9148451 1 217.4 1.9235 2.014 673 173.7254348 1 2 218.2 1.9205 2.011 871 225.5234844 3 217.8 1.929 1.9995 880 228.1544984 1 218.3 2.0005 2.0065 1286 320.3784152 2 2 220.7 1.989 1.993 1482 373.8575209 3 218.7 1.973 1.999 1806 457.9076157 1 213.9 1.983 1.99 1493 378.3415312 3 2 211.3 1.9845 1.988 1211 306.9563767 3 212.5 2.0155 1.9895 1449 361.3612996 1 221.5 1.973 1.9905 1889 480.9973578 4 2 222.3 1.9735 1.9925 1585 403.0823722 3 222.6 1.9795 1.998 1374 347.4047425 1 221.1 1.972 1.976 1928 494.7811876 5 2 222.7 1.984 1.978 2048 521.8695978 3 221.4 1.987 1.9915 1804 455.8882043
73 Table A.4: 28-day compressive strength
Mix Number Weight [g] D1 [in] D2 [in] Peak Load [lb] Pressure [psi] 1 227.5 1.977 1.986 2880 733.5109041 C 2 227.1 1.959 1.995 2905 743.3079892 3 228.3 1.9945 2.0175 3333 828.3001302 1 219.6 1.9165 1.999 763 199.1603681 1 2 218.1 1.9035 1.993 830 218.785186 3 217.9 1.9555 1.9915 674 173.0699888 1 218.1 1.966 1.982 2802 719.0862216 2 2 218.5 1.9575 1.989 2094 537.823932 3 218.8 1.9685 1.986 1699 434.588986 1 212.4 1.9705 1.9835 1901 486.3774964 3 2 208.7 1.979 1.9805 1625 414.6032712 3 220.2 1.977 1.976 2011 514.7762016 1 222.5 1.999 1.9815 1776 448.3695292 4 2 222.5 1.9735 2.0115 2209 556.4658899 3 223.7 1.983 2.001 2157 543.6011193 1 217.8 1.969 1.9865 1345 343.8650198 5 2 221.6 1.9695 1.9775 1796 461.1411253 3 218.4 1.949 1.9745 1306 339.3705871
74 Table A.4: WVT initial measurements
Sample Weight D1 D2 Area Depth Mix Number [g] [mm] [mm] [m2] [mm]
1 88.7 50.9 50.9 0.00259 20.66 C 2 92.6 50.93 50.85 0.00259 21.98 3 91.8 50.78 50.98 0.00259 20.84 1 86.8 50.85 50.78 0.00258 20.38 1 2 90.2 50.92 50.64 0.00258 20.81 3 86.9 51.02 50.95 0.00260 19.35 1 85.5 50.84 50.97 0.00259 20.43 2 2 82.7 50.78 50.79 0.00258 19.12 3 78.9 50.68 50.61 0.00256 18.89 1 84.8 50.54 50.55 0.00255 20.38 3 2 85.3 50.32 50.4 0.00254 20.56 3 85.1 50.59 51.23 0.00259 20.65 1 82.3 50.66 50.93 0.00258 19.98 4 2 89.1 50.59 50.83 0.00257 20.58 3 89.8 50.76 50.9 0.00258 20.38 1 87.3 51.03 50.66 0.00259 20.69 5 2 87.8 50.85 50.62 0.00257 20.18 3 84.4 50.83 50.95 0.00259 19.37
75 Table A.5: WVT test data
Days Apparatus Mix Number Wt [g] 1 2 5 6 7 8 24 48 120 144 168 192 1 357.7 357.5 357.2 356.6 356.2 355.9 355.7 C 2 313.4 313.2 313.1 312.4 311.9 311.9 311.5 3 297.1 296.9 296.7 295.9 295.7 295.3 295.1 1 300.8 300.3 299.9 298.3 297.9 297.3 296.8 1 2 300.6 300.3 299.9 298.8 298.4 297.9 297.4 3 320.6 320.2 319.9 318.7 318.2 317.9 317.3 1 303.6 303.3 303 302.1 301.8 301.5 301 2 2 299.5 299.3 299 298 297.6 297.2 296.8 3 283.9 283.7 283.4 282.4 282.1 281.7 281.4 1 318 317.7 317.4 316.3 315.8 315.3 314.8 3 2 316 315.7 315.5 314.4 314 313.5 313.2 3 298.5 298.2 297.9 296.9 296.4 296 295.7 1 306.5 306.1 305.8 304.7 304.1 303.7 303.2 4 2 328.6 328.3 328.1 327.2 326.8 326.4 326.1 3 311.6 311.2 311 310.1 309.6 309.3 308.8 1 304.9 304.8 304.6 303.7 303.3 303 302.6 5 2 307.7 307.4 307.3 306.5 306.1 305.9 305.5 3 304 303.8 303.5 302.5 302 301.7 301.3 Days Apparatus Mix Number Wt [g] 9 12 13 14 16 19 216 288 312 336 384 456 1 357.7 355.4 354.8 354.3 354 353.4 352.4 C 2 313.4 311.3 310.4 310.3 309.9 309.2 308.4 3 297.1 294.9 294 293.8 293.5 292.8 291.9 1 300.8 296.3 295 294.5 294 292.8 291.5 1 2 300.6 297.1 296.1 295.6 295.2 294.3 292.9 3 320.6 316.9 315.9 315.5 315.1 314.1 312.9 1 303.6 300.7 299.7 299.3 299 298.2 297.1 2 2 299.5 296.5 295.5 295.2 294.8 293.9 292.9 3 283.9 281.1 280.1 279.8 279.4 278.7 277.6 1 318 314.6 313.3 312.9 312.4 311.3 309.9 3 2 316 312.9 311.7 311.3 310.9 310.1 308.9
76 3 298.5 295.3 294.3 293.9 293.5 292.7 291.6 1 306.5 302.8 301.7 301.3 300.8 299.9 298.6 4 2 328.6 325.7 324.9 324.5 324.1 323.3 322.3 3 311.6 308.6 307.7 307.3 306.8 306.1 304.9 1 304.9 302.3 301.4 301 300.6 300 298.9 5 2 307.7 305.3 304.5 304.4 303.9 303.4 302.5 3 304 301 300.1 299.8 299.3 298.5 297.5
Table A.6: Water absorption data
Sample Weight 15 30 60 120 240 360 1440 Mix Number [g] min min min min min min min 1 215.6 221.8 225.8 232 242.9 254.1 254.9 255.5 C 2 216.5 218.5 219.7 221.6 225.5 234.1 241.3 254.7 3 217.2 220.2 222.3 225.5 232.1 246.8 255.1 256.4 1 214.6 228.1 237.4 247.2 249.4 249.7 249.9 250.2 1 2 213.9 221.8 228.4 237.6 247.2 247.6 247.8 248.3 3 214.3 231.3 243 247.2 247.6 247.8 247.9 248.4 1 213.2 215.7 217.1 219 222.4 229.9 236.7 251.7 2 2 211.4 211.8 212 212.3 212.8 213.5 214 218.1 3 212.4 213.9 214.9 216.2 218.7 223.4 227.8 251.2 1 206.2 210.4 213 216.6 223.8 237.7 246.4 247.2 3 2 215.4 216.1 216.6 216.9 218.2 220.7 223.1 239.8 3 206.5 207.5 208 208.9 210.8 214.3 217.6 241.4 1 217.1 223.8 227.8 233.6 244.1 257.2 257.6 258.1 4 2 218.7 222.5 224.6 227.7 233.7 245.5 255.3 259.1 3 206.6 212.9 216.4 221.3 229.8 244.1 246.2 246.6 1 215.7 217.7 218.8 220.3 223.1 228.3 232.9 253 5 2 214.2 216.1 217.3 219.1 222.1 227.3 232 250.4 3 218.6 220 220.7 221.7 223.8 227.4 230.4 248.9
77 Table A.6: Pier compression data
Height Width Length Net Corrected Sample Compressive Load 1 2 1 2 1 2 Strength 1 12.38 12.38 3.83 3.821 8.1305 8.26 83046 2649 2 12.38 12.38 3.721 3.8875 8.2775 8.1975 76243 2433 3 12.38 12.38 3.847 3.8165 8.208 8.3205 80890 2554 4 12.38 12.38 3.9155 3.819 8.228 8.3705 71998 2243 5 12.38 12.38 3.778 3.844 8.341 8.3655 67029 2106 6 12.38 12.38 3.8455 3.689 8.1905 8.1855 72859 2362 7 12.38 12.38 3.6585 3.7875 8.2305 8.37 78198 2531 8 12.38 12.38 3.7695 3.7875 8.174 8.2415 74890 2415 Average 2412
78 Table A.7: Bond strength data
Flexural Applied Tensile Width Length Load Strength Prism No. (in) (in) (lb.) (psi) Mix C 3.828 8.26325 22 13 3.8425 8.371 61 36 3.81825 8.2995 10 6 3.829 8.33875 29 17 3.76625 8.29375 90 56 3.83975 8.2155 106 64 3.829 8.33875 124 75 3.889 8.3395 14 8 3.838 8.32075 66 39 3.81 8.34475 112 68 3.902 8.3655 68 39 3.8905 8.37875 88 51 3.833 8.28475 41 24 3.8815 8.323 0 xx 3.85825 8.28325 100 59 3.85075 8.2985 24 14 3.7585 8.30375 58 36 3.858 8.2985 10 6 3.855 8.3965 23 13 3.849 8.3025 43 25 Average 34 Mix 2 3.84825 8.3545 186 111 3.76975 8.27925 147 92 3.802 8.23475 131 81 3.8605 8.39375 125 73 3.8565 8.372 101 59 3.78225 8.2575 100 62 3.8415 8.30875 128 77 3.85725 8.32925 157 93 3.83525 8.404 127 75 3.84 8.42625 210 124 3.80725 8.25025 212 131 3.8165 8.29025 195 119 3.84125 8.32475 139 83 3.81025 8.2645 167 102 3.8855 8.4435 176 102 3.76725 8.2595 180 113 3.8705 8.3365 142 84 3.82575 8.35125 188 113
79 3.863 8.345 113 67 3.81575 8.27875 199 122 Average 94
Table A.8: Seawater data
Mix C 1 Time [days] 7 28 60 90 180 7 28 60 90 180 1586.0 Ca 1370.51 4 1465.38 1997.75 1142.62 1691.50 2986.91 3044.77 3281.41 2629.83 1863.5 Mg 2997.36 2 831.74 481.61 52.15 3298.15 1494.34 182.90 280.48 434.49 2599.7 S 2796.31 8 2303.17 2501.54 2207.02 2893.77 2526.57 2072.82 2037.33 2263.94 31078. 28902.1 26467.6 37421.2 34263.4 33257.5 33131.8 33076.2 Cl 34131.87 98 28539.18 9 8 2 1 4 8 8
Mix 2 3 Time [days] 7 28 60 90 180 7 28 60 90 180 1391. Ca 1446.82 1517.47 22 1475.70 1110.26 1653.29 2023.03 2367.07 2377.74 2097.36 211.2 Mg 3067.13 1291.53 3 38.92 9.18 2077.29 26.58 9.99 4.77 6.38 2005. S 2715.11 2312.84 98 2070.84 1544.57 2544.03 1963.97 1844.39 1739.54 1331.20 2625 27891.5 26863.7 32139.8 26655.1 30948.5 28814.8 29887.8 Cl 34057.7 28834.68 7.42 0 4 9 6 6 0 3
Mix 4 5 Time [days] 7 28 60 90 180 7 28 60 90 180 2055. Ca 1374.32 1708.97 24 1732.72 2044.64 1187.40 1093.96 1583.28 1680.63 1978.11 Mg 2053.50 443.17 21.30 5.90 5.74 2528.60 2181.82 556.79 172.44 22.24 2155. S 2271.18 2131.39 64 1791.70 1388.46 2410.17 2357.03 1957.32 1939.29 1427.97 2738 26032.9 30228.1 30877.2 29564.0 26781.9 24909.6 26610.8 Cl 28136.5 26646.23 6.70 2 6 2 7 3 0 5
80 REFERENCES
1. L. Edison, 'Perspectives: The Reintroduction of Natural Cement,' Journal of ASTM International, 4 [8] (2007).
2. P. Maravelaki-Kalaitzaki, 'Hydraulic lime mortars with siloxane for waterproofing historic masonry,' Cem. Concr. Res., 37 [2] 283-290 (2007).
3. B. Lubelli, R. P. J. van Hees and H. J. P. Brocken, 'Experimental research on hygroscopic behaviour of porous specimens contaminated with salts,' Constr. Build. Mater., 18 [5] 339-348 (2004).
4. M. P. Edison, 'Formulating with Rosendale Natural Cement,' Journal of ASTM International, 4 [1] 32-43 (2007).
5. R. Vyskocilova, W. Schwarz, D. Mucha, D. Hughes, R. Kozlowski and J. Weber, 'Hydration Processes in Pastes of Roman and American Natural Cements,' Journal of ASTM International, 4 [2] 96-105 (2007).
6. E. Eckel, 'Natural Cements'; pp. 200-247 in Cements, Limes, and Plasters. John Wiley & Sons, Inc., New York, 1928.
7. J. J. Walsh, 'Petrography: Distinguishing Natural Cement from Other Binders in Historical Masonry Construction Using Forensic MicroscopyTechniques,' Journal of ASTM International, 4 [1] 20-31 (2007).
8. R. Boynton and K. Gutschick, 'Durability of Mortar and Masonry'. National Lime Association, Arlington, VA, 1964.
9. U. Cummings, 'American Cements'. Rogers & Manson, Boston, 1898.
10. R. Hanley, 'A study of the workability of natural hydraulic lime mortars and its influence on strength,' Mater Struct, 41 [2] 373 (2008).
11. I. Papayianni and M. Stefanidou, 'Strength–porosity relationships in lime–pozzolan mortars,' Constr. Build. Mater., 20 [9] 700-705 (2006).
12. S. Mindess, J. Young and D. Darwin, 'Concrete'. Pearson Education, Inc., Upper Saddle River, NJ, 2003.
13. A. Çavdar and Ş Yetgin, 'Investigation of mechanical and mineralogical properties of mortars subjected to sulfate,' Constr. Build. Mater., 24 [11] 2231-2242 (2010).
81 14. S. U. Al-Dulaijan, 'Sulfate resistance of plain and blended cements exposed to magnesium sulfate solutions,' Constr. Build. Mater., 21 [8] 1792-1802 (2007).
15. C. Ftikos and G. Parissakis, 'The combined action of Mg2+ and Cl− ions in cement pastes,' Cem. Concr. Res., 15 [4] 593-599 (1985).
16. J. Goncalves, R. Toledo Filho and E. Fairbairn, 'Evaluation of magnesium sulfate attack in mortar-metakaolin system by thermal analysis,' J Therm Anal Calorim, 94 [2] 511-516 (2008).
17. M. Santhanam, M. Cohen and J. Olek, 'Differentiating seawater and groundwater sulfate attack in Portland cement mortars,' Cem. Concr. Res., 36 [12] 2132-2137 (2006).
18. A. Moropoulou, A. Bakolas and E. Aggelakopoulou, 'Evaluation of pozzolanic activity of natural and artificial pozzolans by thermal analysis,' Thermochimica Acta [420] 135-140 (2004).
19. C. Ouyang, A. Nanni and W. F. Chang, 'Internal and external sources of sulfate ions in portland cement mortar: two types of chemical attack,' Cem. Concr. Res., 18 [5] 699- 709 (1988).
20. A. Arizzi, H. Viles and G. Cultrone, 'Experimental testing of the durability of lime- based mortars used for rendering historic buildings,' Constr. Build. Mater., 28 [1] 807- 818 (2012).
21. P. Lopez-Arce, E. Doehne, J. Greenshields, D. Benavente and D. Young, 'Treatment of rising damp and salt decay: the historic masonry buildings of Adelaide, South Australia,' Mater Struct, 42 [6] 827-848 (2009).
22. J. Dweck, P. Ferreira da Silva, P. Buchler and F. Cartledge, 'Study by thermogravimetry of the evolution of ettringite phase during type II Portland cement hydration,' J Therm Anal Calorim, 69 [1] 179-186 (2002).
23. S. Antao, 'DTA, TG, and XRD studies of sturmanite and ettringite,' Canadian Mineralogist, 40 [5] 1403-1409 (2002).
24. D. A. St John, A. B. Poole and I. Sims, 'Salt and sea water attack'; pp. 259-261 in Concrete Petrography: A handbook of investigative techniques. Arnold, London, 1998.
25. O. S. B. Al-Amoudi, M. Maslehuddin and Y. A. B. Abdul-Al, 'Role of chloride ions on expansion and strength reduction in plain and blended cements in sulfate environments,' Constr. Build. Mater., 9 [1] 25-33 (1995).
82 26. B. Lubelli, R. P. J. van Hees, H. P. Huinink and C. J. W. P. Groot, 'Irreversible dilation of NaCl contaminated lime–cement mortar due to crystallization cycles,' Cem. Concr. Res., 36 [4] 678-687 (2006).
27. J. Ahl, 'Salt diffusion in brick structures Part II The effect of temperature, concentration and salt,' J Mater Sci, 39 [13] 4247-4254 (2004).
28. M. Nasraoui, W. Nowik and B. Lubelli, 'A comparative study of hygroscopic moisture content, electrical conductivity and ion chromatography for salt assessment in plasters of historical buildings,' Constr. Build. Mater., 23 [5] 1731-1735 (2009).
83