A Critical Review of the paper "Ultra-green construction: reactive MgO masonry products" (Liska, M. and A. Al-Tabbaa (2009).

During the process of reviewing two papers from Cambridge university for the British Institute of Chartered Engineers John came to understand a lot more about what the group at Cambridge are trying to achieve, the merit of it, where they have gone wrong in relation to our own carbonating Eco-Cement and carbonation generally. On this page we examine the paper published in Waste and Resource Management titled "Ultra-green construction: reactive MgO masonry products"[1]

The paper starts well with a call for action I have often made and thanks for the accolades. This is followed by some confusion over the terms Reactive Magnesia and Reactive Magnesia Cements however and the highlighted first section on page 185 defines Reactive Magnesia Cements thus "Reactive magnesia cements are blends of PC and reactive magnesia in different proportions depending on the intended application, from high-density concrete to porous masonry units."

To my recollection, this is not what we have defined reactive magnesia cements as at all, at least in recent times. They do not have to contain PC and if we have said so then we now retract it. In recent times we have defined them as "Cements containing reactive magnesia that can be and often are blended in any proportion with other hydraulic cements such as Portland cement because they will hydrate sufficiently rapidly so as to not cause dimensional distress." They could also be defined as cements containing reactive magnesia but such a lame definition would give no hint of the fundamental requirement - that they hydrate quickly and can be blended with hydraulic cements. In most cases this would mean that they have the ability to hydrate in the same rate order as Portland cement. Technically magnesia is reactive if it has a high proportion of unsaturated co-ordination sites (and thus relatively low lattice energy) and this can be argued thermodynamically to be related to low temperature calcination. See technical.reactive_magnesia.php.

I was annoyed when in the next paragraph they say "An extensive investigation into the strength development of reactive magnesia cement masonry units, in which the PC is either partially or completely replaced by reactive magnesia, has been carried out." The small amount of work they have done could not be called extensive and it came to little because starting off with the wrong hypothesis they reached the wrong conclusions. They could not get blended formulations to carbonate because they did not and probably still do not understand carbonation kinetics and the importance of appropriate particle packing.

The scetch below delineates the fundamental differences in strength with concretes made with reactive magnesia. It has been around since about 2003 in various forms and makes it quite clear that gas permeability leads to greater strength through improved carbonation kinetics.

Strength ~ Permeability and Proportion of MgO in the different kinds of Reactive Magnesia Cements

We reviewed earlier papers by the Cambridge team [2] and the improvements in this paper are noticeable. The team have at last realised that the use of carbonated reactive magnesia can contribute significant strength. They are no longer saying carbonation cannot be achieved or strength gained. This is progress. What they have not realised, at least when they wrote the paper is that there are many ways in which carbonation can be achieved of which forced carbonation is only one way. TecEco achieve rapid carbonation by chemical and physical optimisation, appropriate particle packing resulting in gas permeability and ideal wet-dry carbonation conditions ("optimal carbonation"). We have proved that products like blocks made using our methods can be made, transported and erected in a week. See http://www.tececo.com/exemplar.eco-masonry.hockings.php or http://www.tececo.com/files/newsletters/Newsletter91.php Cambridge on the other hand have not yet been able to achieve satisfactory carbonation of any mix regardless of the proportion of the MgO to other binders without using forcing and even in this paper achieved only just over 70% conversion with forced carbonation. This is significantly better than their initial findings which were that magnesium cements (presumably all of them) did not carbonate or contribute sufficient strength much at all. The order of events reminds me of Imperial College London who participated in the dumping of our technology only to later claim they invented reactive magnesium cements. They had to apologise [2]. Hopefully I am wrong.

With such a poor early carbonation results (See the documents at http://www.tececo.com/rdandd.thirdparty.php) there were two major possible hypotheses they could have arrived at and one minor corollary. Either their formulations and methods were wrong or that PC-MgO blends do not gain strength without forced carbonation. They have supported the wrong hypothesis in a very unscientific way, consistently arrived at the wrong conclusion by adopting the latter without questioning the former. The corollary is that if you apply a process as blunt as forced carbonation it will work regardless whether you can get unforced natural optimised carbonation to work or not.

Because they have not been able to achieve optimal carbonation their results show that the carbonation of pure magnesia and magnesia PC blends is not achievable without forcing. This is the corollary and of course it is true in the context of their thinking! Designing experiments to achieve the required result is totally the wrong way to go about science.

The scientific method demands a review of their work because the conclusion they have reached do not support the evidence we have provided to prove optimised carbonation works. The clues are even in their own work that they are going about carbonation in the wrong way. They have only been able to achieve a little over 70% conversion using forcing for a considerable time and should have realised that their formulations are very non optimal.

We can achieve sufficient carbonation in days to allow transport and erection See http://www.tececo.com/exemplar.eco-masonry.hockings.php or http://www.tececo.com/files/newsletters/Newsletter91.php. They claim they have worked with us [3]. They have not. There has been no discourse whatsoever of a scientific nature between us other than banter relating to reviews I have done of papers they have written! [3]. I hope they stop dumping our technology at the cost of their scientific reputations even if they are doing so though ignorance as we should really all be working together.

For a sustainable future we need new greener cements in the market place and improvements to the process of making and using existing cements. Elsewhere I have discussed at great length our carbonate built environment solution to global warming. There is room for both high reactive magnesia with no other hydraulic components, high reactive magnesia with guaging salts or phosphates and high reactive magnesia with or without guaging salts or phosphates blended with with hydraulic binder components such as for example PC as well as geopolymers. What matters in relation to all magnesia and lime based cements that rely on re adsorption or non release of CO2 is that we fix the supply chain so that there are no CO2 releases during manufacture so the source reactive magnesia is made sustainably and at much lower cost and I am working on that. See http://www.tececo.com/files/newsletters/Newsletter92.php

Carbonation as the authors state leads to greater strength and forced carbonation, with a higher proportion of hydroxide components able to carbonate, leads to even more strength. (Both the Cambridge team and TecEco both understand this applies to hydroxide phases only - but does the reader? See later.)

The paper reveals the tremendous potential of carbonating magnesium phases but apart from the low conversions achieved because of non optimisation, forced carbonation without PC has downsides including potential alienation of the Portland cement industry and expensive process including for example carbonation rooms. There are however some advantages including the possible use of more fine waste such as fly ash and mine tailings unsuitable for other uses. In the future fly ash will be in limited supply because the coal industry is in some countries "dead man walkin"[4]. There are also producers of point sources of CO2 looking for on site in house solutions.

Thank you to the Cambridge team for the credits in the second column on page 185. Use of the term porous without further clarification concerns me however as it is technically incorrect and suggests the authors do not understand the difference between porosity and permeability which is required. For example in column 2 on page 185 they say "Reactive MgO is manufactured at a much lower temperature (600–750degC) than PC and, in porous blocks under appropriate curing conditions, reactive MgO will carbonate by absorbing CO2 and, as a result, gain significant strength. The potential for significant CO2 sequestration within porous masonry units containing reactive MgO as the cement component and the associated significant increase in strength are the subject of this paper." I made this mistake early on and so I really can't be too hard on them. For the record porous substrates may or may not be permeable and for carbon dioxide to enter the matrix of the substrate to react there must be some form of transport. Gas transport is much more rapid than solid or liquid diffusion. Porosity is a measure of the void spaces in a material, and is a fraction of the volume of voids over the total volume, between 0–1, or as a percentage between 0–100% [5]. Permeability implies that the the pores are sufficiently interconnected for gases or liquids to be able to pass though. The porosity at which this can occur is referred to as the percolation point. The following diagram should assist coming to this understanding.

In the above diagram situation 1 - 4 are matrixes. The red and white dots represent pores or solid material. In the case of white solid material and red pores, 1 is red pore permeable. 4 is red pore impermeable and in the case of red solid material and white pores, visa versa. The percolation point for both red and white pores in the above illustration is somewhere between situations 2 and 3.

At some point of relative occurrence the red or white dots representing both pores and solid material "percolate" and become connected from one side of the containing squares to the other. Think of this simple analogy in relation to a concrete matrix. It is clear that porosity does not mean permeability. If the percolation point is not exceeded at a given porosity the concrete is not permeable.

Porosity leads to permeability but does not imply permeability. It's a bit of a mince on words I know and others including us have used the term porous when the term permeable would have been more precise. The industry use the term pervious but mainly in relation to the ability to let water pass through. Because of this confusion in terms and after looking at their aggregate particle gradation diagrams in this paper we doubt whether the Cambridge team have achieved all important gas permeability essential for kinetic optimisation of carbonation. An understanding of percolation theory is required and we suggest they ask for a lecture by Geoffrey Grimmett from Cambridge University to their students as he wrote a book on the subject [6]. Remember that CO2 molecules are very very small indeed and so the space through which they need to pass does not have to be very large. It is thus possible to mono and gap grade down to very small sizes to achieve gas permeability however an understanding of the maths is required. See http://www.tececo.com/technical.particle_packing.php

It is pleasing that the Cambridge team take the view at the bottom of the second column on page 185 that reactive MgO is quite different to dead burned MgO and chose to quote us and Mark Shand's book [7]. We have refined much of what we say about reactive magnesia and more information is to be found at http://www.tececo.com/technical.reactive_magnesia.php.

The first paragraph on page 2 sets the background with a comment regarding Portland Cement emissions. The reader may or may not agree with the exact numbers; I don't but as it really all depends on the references you use I have no problem with this. What is important is that the Cambridge authors recognise that we are dealing with a thermodynamic cycle when they say "Hence, although the production of MgO through this route results in more CO2 emissions than PC production, the potential for CO2 sequestration within the MgO blocks combined with their significantly enhanced strength is expected to reverse this order, as will be demonstrated in this paper." We suggest our readers think about what could be the case with manufacture with releases of CO2 as with our Tec-Kiln

Whilst on the subject of strength, also mentioned on page 186 in the second column where microstructure is stated as considered, there is one very important, dominating factor that should have been mentioned that was not and that is the significant molar volume increase from MgO to hydrated magnesium carbonates. The total volumetric expansion from magnesium oxide to lansfordite is for example 811% and to nesquehonite 568%. This molar volume expansion is probably a much more important contributor to strength than microstructure although we agree microstructure is very important. To be fair densification is given passing mention later in the paper. Of course, under forced carbonation conditions the more MgO the more strength! the multiplier is significant.

A number of important initiatives to make MgO without releases are mentioned in column 1 of page 186 and since the time of writing by the Cambridge authors we have come to favour a process first thought of in the 40's by DOE scientists[8] and since rediscovered and improved by a team at the university of Rome [9][10]. See also http://www.tececo.com/files/newsletters/Newsletter92.php. This process is important, probably more so than all the others mentioned in the first column of page 2 in the paper because it produces nesquehonite which can be used as a microaggregate and aggregate and requires less energy to calcine to produce the all important MgO.

On page 186 in the second column the team discuss their materials and methods. It was a mistake to continue to use the Causmag material we donated as it would have been around 5 years old at the time of testing and probably already partly carbonated and they do later admit this. I also question the particle packing. I don't have a working version of our TecBatch program any more to analyse the particle packing and doing it by hand takes for ever. A good look at the particle sizes used would suggest they would however pack too well and not be very permeable. There are too many low end fines and no gap grading.

Blocks are not pressed so much as vibrated into shape in Australia. I suggest that pressing at 5Kn would probably even further reduce permeability. Whilst on this subject the reader should realise that doubling the size of an orifice increases the cross sectional area by a factor of four (at least quadrupling the gas flow) whereas doubling the CO2 concentration only doubles the CO2. See http://www.tececo.com/technical.particle_packing.php.

In table 4 the proportion of MgO to PC at 1:1 is much lower than we recommend at 66-85 MgO : 33-15 PC so the strength of the carbonated PC in figure 4 is correspondingly lower as might be expected. Our high magnesia PC mixes achieve only slightly less or the same ultimate strength through natural carbonation than pure MgO mixes through forced carbonation if the proportions are correct, packing correct and other kinetic conditions favourable. It would have been interesting to see the results in figure 4 on page 189 if this were the case.

Having made all the abovementioned mistakes setting up the experiments the science becomes rather more classic and what to expect from a university more interested in judgemental rather than creative thinking. I note with interest that the curves for strength gain in figure 3 on page 188 continue upwards even after around 300 days for the MgO - PC system in spite of the impermeability of the blocks whereas the curve for strength gain of MgO formulations alone totally flattens out and no significant strength is gained. Wearing de Bono's green hat it means something totally different to me than the Cambridge team. It possibly reinforces our contention that the Ca++ from PC is a catalyst for dissolution of MgO and that carbonation is also favoured by the high pH as a result of the production of Portlandite during the hydration of PC on the other hand the result could be entirely the result of the PC present in such impermeable blocks. For the record I suspect the fact that CSH becomes highly ionic at high pH has something to do with carbonation of our blends given the highly polar nature of magnesium compounds in the system.

Figure 4 on page 189 certainly shows the superiority of magnesium carbonation systems for strength over calcium carbonation systems under forced carbonation conditions. This can be extrapolated to not forced by optimised systems such as ours. At least on page 189 the fact that the Causmag XLM we donated was old is noted. Figure 5 is a rather meaningless graph and after spending a couple of minutes on it I choose to make no comment.

In relation to durability on page 189 I will be very interested in that. My prediction. Mg carbonates are very durable, including nesquehonite other than at low temperatures. Have a look at the carbonate fields diagram below and see Ballirano's paper[10].

Stability of Magnesium Carbonates (Source: unknown)

The XRD work is useful and I note the detection of hydrotalcite in the PC - MgO mixes as we have not found this mineral.

That the PC - MgO mix in figure 3 behaved like a Tec-Cement concrete with far too much magnesia is confirmed by the XRD results in Figure 6 where no significant carbonate peaks are to be found. The problem was of course a lack of gas permeability. If the gas had no way of getting into the matrix how could carbonation ensue? Figure 7 shows that under forced carbonation a smorgasbord of carbonates are produced. This is considered in relation to MgO phases on page 190 "It is thus clear that the degree of reactivity, the presence of impurities and the curing environment all play a strong role in the type of carbonation products formed." I would add to this short list that pH and permeability are probably more important and it is a bit rash to jump to conclusions.

The authors found on page 194 that in relation to MgO only mixes the rate limiting step for carbonation was hydration of MgO. This does not necessarily apply to all mixes and other work including our own in relation to PC-MgO mixes indicates both brucite and nesquehonite can co-exist during setting. There are a large number of variable influences so perhaps not a good idea to quickly (as I did in the beginning) jump to too many conclusions.

Also on page 194 the authors give overall carbonation results for MgO and lime systems indicating a maximum of 71% for MgO and total conversion for lime. We found no brucite or MgO after complete carbonation meaning that, in our systems at least, with all important permeability, complete carbonation ensues. The difference is of course gas permeability. The authors note the significant densification that ensues and this is a result of molar volume expansion as we point out earlier. What pores they have in their relatively impermeable mixes were as a consequence totally blocked off and so carbonation cannot proceed to completion even with forcing.

In relation to the discussion about LCA analysis on the same page I again point out that the supply chain is the weakest point in all our argument for man made carbonate. A source of cheap magnesia made without releases is essential and we should be addressing this issue. See http://www.tececo.com/files/newsletters/Newsletter92.php

Apart from confusion about porosity and its relationship to permeability there is one more serious error of terminology that persists throughout this paper and that is the use of the term “weight”. It is unfortunately a common one. I was once corrected for using this term and hope it is nowhere to be found in a scientific paper on the TecEco web site or that I have written in recent years unless in the correct context. The term “weight” is an imprecise everyday term that does not apply to “mass” which is constant fixed and used in science. Mass is often defined as the amount of matter in an object. Mass and weight are not the same thing.  Weight is the force on an object due to the gravitational pull of a planet or other heavenly body.  Mass on the other hand, remains constant, no matter where it is and the term the author should use. Most good scientists know the difference and it detracts from the positive aspects of the paper.

This paper makes significant contributions to the overall science but is too concerned with fulfilling the main hypothesis of the Cambridge group rather than exploring what is or could be. Gas permeability is very much the limiting factor from the very beginning but never recognised as such in for example the discussion about conversion. Carbonation of hydroxide phases, even in PC mixes is always useful and at least the Cambridge team possibly recognise that there is little point in carbonating silicate phases thought a good idea by for example Halifax Canada-based Carbon Sense Solutions [11].