TGA (thermogravimetric analysis) is an important method to analyze the hydration products of cementitious materials. When a TGA test for cementitious materials sample is performed, the initial TGA curve may appears as such:

A typical TGA test result shows the buoyancy effect. Portland cement past, w/c=0.4, cured 1 day.

As can be seen from the curve, the mass exceeded the original mass, which is not reasonable, since the decomposition of the sample always decrease the mass of testing sample. So why does this happen? This is because of the buoyancy effect in the TGA equipment.

Every substance found in a gas atmosphere is subjected to a buoyant force. This results in “apparent” mass changes. The degree of buoyancy, and thus the degree of mass change as well, are basically dependent on the volume of the substance and the density of the prevailing gas.

However, for thermogravimetric investigations, it is not the absolute value of the buoyancy at a certain temperature that is decisive, but rather its change as a function of temperature. A typical buoyancy curve is shown as follows.

A blank TGA test.

It was recorded using a NETZSCH STA 409 with a DSC sample carrier in a static N2 atmosphere. There is no sample in the crucible, but the recorded result shows the mass is always over ZERO.

Due to the volume-dependence of the buoyancy, it is important to specify the sample carrier type in addition to the gas atmosphere, since different sample carrier/crucible combinations give rise to different “apparent” mass changes.

How to remove the Buoyancy Effect?

Knowing this, we can remove the buoyance effect by running a blank test. Repeat the test with the same crucible but without anything in it or with inert material, then the buoyancy effect is recorded. Subtracting the buoyancy result from the first measurement result, the true mass change of the test is obtained.

A typical TGA test result after the subtracted the blank test, removing the buoyancy effect. Portland cement past, w/c=0.4, cured 1 day.

Why does the buoyancy curve initially show such a drastic rise?

You may find the buoyancy curve initially shows a drastic rise, why is that? This phenomenon arises from the design of TGA device.

At low temperatures, heat transfer from the furnace to the sample thermocouple is exclusively through convection. Since a certain amount of time is required before the crucible material is evenly heated and the warm air preferably flows along the inside of the protective tube, and thus does not reach the crucible immediately, the actual temperature of the gas atmosphere is far higher, particularly at the beginning of a measurement, than the temperature sensor indicates. Therefore, the gas density changes dramatically in the beginning of the test, thus subjects dramatic “apparent mass” change.

A schematic figure showing the temperature difference

Since samples that undergo MIP are usually irregular, the bulk volume is not possible to measure without being immersed in liquid. By the help of MIP, the bulk volume can be measured, thus the bulk density can be calculated simply dividing the mass by bulk volume.

At the end of MIP, the intruded volume of mercury at corresponding pressure is recorded, which means the pore volume is known. As soon as the bulk volume and pore volume are known, the solid volume including closed fine pores is also known. Then the porosity and apparent density are easily calculated.

The density of cement particles is commonly referred as 3.1 g/cm^3, and that of slag and limestone are 2.6 and 2.7 g/cm^3, respectively. You may be curious to know how about the bulk density of cement paste. Are they higher and lower than the raw materials? Are they stable values as curing ages extend and thus more hydration products formed?

Recently, several cement and blended paste samples of mine have been tested using MIP. I list the bulk density results as below. All the pastes are mixed at water : powder = 0.4, sealed and cured at 20 °C. As each scheduled curing age (1, 3, 7, 14, 28 and 91 days) reaches, the hydration were stopped by liquid nitrogen, and later freeze dried to remove the frozen free water.

Bulk density of hydrating (blended) cement pastes

From the results, the bulk density of cement or blended cement ranges between 1.4 to 1.8 g/cm^3, much lower than raw material, and blended cement paste, such as the ternary blended cement paste, constantly has lower bulk density. This show that the hydrating blended cement pastes are more porous, which are further confirmed by tested porosity results below.

Porosity of hydrating (blended) cement pastes

However, this is pretty complicated in the case of cementitious materials. The mass of a certain amount of harden cement (or concrete) paste is a finite value, but how about the volume? since the harden cement paste is porous, how do we consider the open and close pores inside the paste as we want to determine the volume?

I met the problem when I performed Mercury Intrusion Porosimetry (MIP) test, especially when I read the output result from the MIP machine, there are bulk density, skeletal density, envelope density and apparent density.

It is not difficult to find these definitions from either textbooks or internet. First I list them below:

Apparent particle density: The mass of a particle divided by its apparent (particle) volume (BSI).

Bulk density: (also called Bulk powder density): The apparent powder density under defined conditions.

The mass of the particles divided by the volume they occupy that includes the space between the particles (ASTM D5004).

The ratio of the mass of a collection of discrete pieces of solid material to the sum of the volumes of: the solids in each piece, the voids within the pieces, and the voids among the pieces of the particular collection (ASTM D3766).

Envelope density: The ratio of the mass of a particle to the sum of the volumes of: the solid in each piece and the voids within each piece, that is, within close-fitting imaginary envelopes completely surrounding each piece (ASTM D3766). The ratio of the mass of a particle to the envelope volume of the particle (implied by BSI).

Skeletal density: The ratio of the mass of discrete pieces of solid material to the sum of the volumes of: the solid material in the pieces and closed (or blind) pores within the pieces (ASTM D3766).

True density (also called True particle density): The mass of a particle divided by its volume, excluding open pores and closed pores (BSI).

It is a headache to me to well understand them and distinguish these different density definitions. I guess it is still not clear for people to understand these densities either. Fortunately, I found a schematic picture (below) well illustrate the physical meaning of these density definitions, which you may no longer misunderstand them any more.

Is it now clear?

Representative sampling

The mineralogy and glass content of slags depend largely on the mode of cooling of the slags, e.g. slow cooling in a slag pot can result in a large amount of crystalline phases, and fast granulation in water can result in high glass content. A mineralogical analysis will thus only represent the mineralogy of a slag system for a certain grade of cooling and strongly different results can be obtained when cooling conditions are different between the different batches.

Even in a single slag pot, mineralogy and glass content can vary strongly. Taking a sample representative for a slag pot can be done by performing the sampling after a first size reduction of the bulk material or, by mixing representative parts of the slag pot (sides of the pot, center, near cracks, in center of slags, etc.)

Crushing

A grain size of <10 micron is required for quantitative X-ray diffraction (Rietveld). To avoid amorphisation of the sample during grinding, wet milling is done in a McCrone Micronizing mill. Before grinding in the McCrone Mill, the samples are crushed and passed through a 500 micron sieve.

Take representative amounts of sample (50-100 g). Crush the sample by hand in a porcelain mortar. Use shock impact for grinding, avoid shearing. A jaw crusher can be used, but automatic milling devices which could induce shear stress of amorphisation such as ball mills should be avoided!

Micronizing of the sample

- Weigh 2.7 g of sample; add 0.3 g (10%) of ZnO internal standard. Note down the exact weights, they will be used in the calculations

- Micronize the samples in a MeCrone Micronizing mill using 5 ml of ethanol (methanol) as grinding agent and a grinding time of 5 minutes for soft material (e. g. limestones) up to 10 minutes for hard materials (eg. quartzites, slags). Since methanol tends to react with some artificial minerals, and as it is toxic, ethanol is preferred. To ensure that samples are ground up to < 10 micron, the size required for X-ray quantification, the grain size is best checked by (wet) laser diffractometry.

- After micronizing, recuperate the sample in porcelain cups. Cover the cups with plastic foil, because recovery of powder from the porcelain when dried is difficult. Wash with methanol to recuperate as much as possible of the sample.

- Dry for one—two days under a fume hood (methanol is toxic).

Preparation for X-ray diffraction

- Dried samples are gently disaggregated in an agate mortar and passed through a 250 micron sieve, to ensure good mixing of sample and ZnO standard.

- +/- 0.5 g of sample is needed for the X-ray analysis. Side-loading with frosted glass is recommended to fill sample holders, to prevent preferential orientation of fibrous zeolites and clay minerals. Sample holders are gently tapped while filling, to ensure good packing of the grains. Alternatively, back loading is used.

The content is offered by Dr. Lieven Machiels

A typical chemical reaction of a slag could be like the following [NIST, D.P Bentz],

C7.88S7.39M3A + 2.6CH + bH → 7.39C1.42SHmA0.046 + 0.66M4.6AHd

C=CaO, S=SiO2, M=MgO, A=Al2 O3, CH=Ca(OH)2, H=H2O.

The method EDTA (Ethylenediaminetetraacetic acid) has been described by Erntroy [Erntry, 1987]. The following is a short description of the method.

• 1. 93.0g of disodium EDTA 2H2O is dissolved in a mixture of 250 ml triethanolamine and 500 ml water. The solution is transferred to an volumetric flask;
• 2. 173 ml of diethylamine added and the mixture made up to 1000 ml with water;
• 3. For the extraction, 50 ml of the above solution is pipetted into a beaker and diluted to approximately 800 ml with water;
• 4. The solution is brought to a temperature of 20.0±2 ◦C and 0.5 g of the dried and powdered sample paste, weighed to the nearest 0.0001 g, sprinkled over its surface;
• 5. The solution is stirred for 120 ± 5 min while maintaining the stated temperature and is then filtered under vacuum through a 90 mm diameter Whatman GF/C filter which had been previously washed with 100 ml of distilled water, dried and weighed;
• 6. The residue is then washed 5 times with 10 ml lots of distilled water, fried at 105 ◦ C for 1 hour and weighed to the nearest 0.0001 g.
• 7. Calculation of reaction degree of slag.

2 Calculation of reaction degree of slag

Take Msl g slag and MCH g Ca(OH)2 and MH g H2O as reactant. When the test age reaches, say 3 days, take two pieces of paste, dry and weigh
them, then,

• 1. one is ground to powder for EDTA test;
• 2. the other piece undergoes a ignition loss test (LOI), from the LOI, the slag mass fraction in the dried paste powder can be determined, the ignition loss of CH should be taken into account, for Ca(OH)2 is decomposed to CaO and H2O.

The LOI fraction is: fLOI, thus, 1 − fLOI is the mass fraction of CaO+slag, so the original slag content in the powder can be determined.

2.1 Correction of hydrotalcite

h= Mass of dried hydrotalcite formed from 1 g of MgO in the slag glass.

The value of h could be 2.35 g based on the assumption in the paper “Degrees of reaction of the slag in some blends with Portland cement”, there the hydrotalcite is considered as M5ACH7.

Therefore, the residue of slag should minus the residue formed from MgO. The content of MgO can be determined by the content of slag of the tested powder.

2.2 Correction of slag

A small amount of slag may be solved in the reagent EDTA solution.

It is better do a pure unreacted slag test using EDTA, to determine the fraction of slag dissolution.

If the dissolution of slag in EDTA is quite small, then there is no need to take account for correction for solved unreacted slag in EDTA reagent, assuming no unreacted slag is solved in the EDTA solution.

A well formatted file (PDF) of the content on this webpage may be downloaded via this link.

1 Topochemical Reaction Concept

Immediately after the first contact of the cement with water a calcium-rich silicious clinker would liberate Ca2+ ions into the solution. A calcium-poor skeleton is left which reacts with the calcium-rich solution which is accompanied by swelling of the hydration products compared with the original volume of the anhydrous cement.

The topochemical concept was launched by Michaelis at the beginning of last century. Since then topochemical phenomena have been reported for the hydration of both C3S and Portland cement.

Schematic representation of proposed hydration mechanisms

2 Through-Solution Concept

In the through-solution concept dissolution of the anhydrous grain after contact of the cement with water is considered to be followed by hydration in the solution. The hydration products then precipitate on the grain surface.

The through-solution concept was first formulated by Le Chatelier for plaster and Portland cement in the first decade of this century. The concept has been supported by many reserachers, Regourd et al., Brunauer and Gauglitz, Williamson and Dron et al.

The debate on whether cement hydration proceeds topochemically or according to a through-solution mechanism dates back to the very beginning of cement chemistry. The subject is of significant importance in view of mathematical modeling of both the hydration process and structural formation. As such, a so-called Simultaneously Operating Mechanisms is proposed to account for the debate.

3 Simultaneously Operating Mechanisms

According to Neville the controversy between the topochemical and through-solution concept can largely be reduced to a matter of terminology. Shebl et al. explains that hydration of C3S involves both through-solution reactions and topochemical reactions (solid-state reaction). The water/solid ratio would be an important factor in this respect. For low water/solid ratios the reaction would be predominantly topochemical, whereas for high water/solid ratios the through-solution mechanism would be more important. Both reactions could occur simultaneously.

The concept of simultaneously operating mechanisms is plausible indeed if the outer products, i.e. the products which are formed outside the original grain boundaries in a relatively water-rich environment, are formed by a through-solution mechanism, while the inner products, formed inside the original grain boundaries, are the result of a topochemical reaction.

From a literature survey on this topic Daimon concluded that it is very difficult to determine which mechanism has marred. The fact that we are dealing with a poly-size system significantly contributes to the complexity of the subject. Probably the effect of the particle size distribution also explains why Kalousek found many contradictions between different authors on this point.

The conclusion is that the mechanism of cement hydration depends on the water/solid ratio, viz. low water/solid ratios lead to topochemical reaction, while through-solution mechanism is more important at high water/solid ratios, and both reactions could occur simultaneously.

Reference: Simulation of hydration and formation of structure in hardening cement-based materials, by K. Van Breugel, TuDelft.

Even in the 21st century, to build a house right now is still a slow, labour-intensive, dangerous process, and almost always over-budget. Unlike the motoring or technology industries which use automated production methods to complete routine construction tasks, housing construction is one of the only industry that still does things manually till now, as said by Professor Khoshnevis from USC.

It is time to change. With the achievement of the research in 3D printing filed, Professor Khoshnevis scaled up 3D printing to make this technology being able to construct buildings by using a process called Contour Crafting. He hope to use this method to improve the basic concept of house construction so that it was accessible to everyone, because with better shelter comes a more civilized society.

How does 3D printer work to print a house?

Here is how 3D printing works to “print” a house:

A CAD design is sent to a large-scale 3D printer that is mounted to a block of land. The printer lays out the concrete-like foundation of the home through a nozzle that can move anywhere on the property. Like any 3D print-out, the house is made layer-by-layer and reinforced with various materials — like electrical, plumbing and communication infrastructure — as the build progresses.

The process is super fast. In a TED presentation, Prof. Khoshnevis said that “we anticipate that an average house, like 2500 square foot house, can be built in about 20 hours from a custom design”.

As for the material used as the “printer ink”, concrete is the solution. The concrete used in the 3D house printer is a mixture of concrete and fibre polymers, meaning that it is more than three times stronger than traditional concrete used in today’s houses. The traditional concrete can withstand roughly 3000 pounds per square inch of pressure, while the new printed concrete can withstand around 10,000 pounds per square inch.

Is it the future of cementitious materials in construction industry?

Professor Khoshnevis says the method could be used to construct emergency or low-income housing. What is more, the concept is currently being supported by NASA so that the space agency can one day be used to build a colony on the Moon, and perhaps on the Mars, who knows.

As a cement researcher, I think the method is really revolutional in the construction industry, even just consider it from the theoretical concept. If the method is quite popular in the future, does it mean cementitious materials are the future of construction materials compared with steel? since the mixture “concrete” is the main material in the 3D printing machine used as “printing ink”.

GGBS (Ground Granulated Blastfurnace Slag) is used all over the world as a direct replacement for Portland cement. It is added to the concrete mixer along with ordinary cement, aggregates and water.

The normal ratios and proportions of aggregates and water to cementitious material in the mix remain unchanged. Mixing times are the same as for ordinary cement. Both wet mixing and dry mixing processes can be used for making concrete with GGBS.

What is the substitution rates of GGBS in concrete?

Usually, replacement rates for GGBS vary from 30% to up to 85%, depending on the application and technical requirements. Typically 50% is used in most instances. Higher replacement rates up to 85% are used in specialist applications such as in aggressive environments and to reduce heat of hydration.

According the EU code “EN-197-1-2000 Cement. Composition, specifications and conformity criteria for common cements“, slag in Blastfurnace cement CEM III/C can be used at a substitution rate as high as 95%.

Below are some typical Substitution Rates

Ready-mix Concrete:
50% GGBS is used for most mixes.

Precast Concrete:
30% to 50% GGBS primarily depending on concrete curing conditions.

Special Concrete:
Mass concrete typically has at least 70% GGBS for temperature control. Concrete in aggressive environments usually contains from 50% to 70% GGBS for enhanced durability performance.

• 1.- Production process chemistry and engineering
• 2.- Sustainable production
• 3.- New cementitious matrix
• 4.- Hydration and microstructure
• 5.- Hydration and thermodynamics
• 6.- Modelling
• 7.- Properties of fresh and hardened concrete
• 8.- Concrete durability
• 9.- Standardization

When reading papers, I saw numerous researchers refer to the information from the papers of ICCC, unfortunately, those proceedings of earlier age are not easy to find, since they are no digital version (like PDF files), or they are not open to the public. Even for the latest proceedings, you have to attend the meetings to get the proceedings, but the cost to attend these meetings is expensive or at least not cheap. What is even worse is that every meeting is definitely not available for all people in the field, due to the difficulties such as travelling and time schedule.

The purpose of conference is to gather people who are interested on some issues in the same field to discuss and share latest results. As a result of meetings, the proceedings paper is a good media to store and spread these information. So we should make these proceedings open for the related researchers rather than close them for a small group of researchers, which is not the REAL Science, Science should be OPEN, and must be OPEN.

I have attended the 13th ICCC thanks to the financial support from my promoter, and got a digital proceedings of the conference. I think it is a good idea to share them with you who are also a researcher on cement science. Just click the link Proceedings of 13th International Congress on the Chemistry of Cement to download them onto your computer or copy the whole folder to your DropBox account.

Share them and make science OPEN.

Concrete made with GGBS cement is much better than concrete made with 100% OPC at maintaining its compressive strength when exposed to high temperatures.

OPC specimens heated to 400 °C or above have been shown to exhibit severe cracking to the point of disintegration after a few days. However, it was found that concrete made with blends of 35%, 50% and 65% GGBS performed much better. In the GGBS specimens there was no visible cracking after exposure to the higher temperatures. The concrete with 100% OPC degraded to powder over the following year, while the concrete made with GGBS maintained its strength over the same period.

Further studies showed that after exposure to 900 °C concrete with 0% GGBS maintained 6% of its original strength, concrete with 30% GGBS maintained 54% of its original strength and concrete with 70% GGBS maintained 70% of its original strength.

Date source: Ecocem

How does slag improve the resistance to fire damage?

Why concrete made with GGBS has better resistance to fire damage? It is generally agreed that above 400 °C Ca(OH)2 which is the hydration products of OPC decomposes into CaO and H2O, then on cooling CaO and left H2O rehydrates into Ca(OH)2 which needs more space than CaO causing the concrete to crack and degrade. With the addition of GGBS into concrete, hydration product Ca(OH)2 is generally reduced compared with concrete made with 100% OPC, thus improving the performance of resisting fire damage.

Reference: Ecocem report–Technical Performance