proposals before in a number of instances, such as ‘resin‐modified’ glass ionomer cement (GIC). If GIC has some good properties but is weak and moisture‐sensitive, whilst light‐cured materials are strong and insensitive, surely we can do both? No. Firstly, the one function replaces the other: in a given volume, something has to go to make space for something else. All too often, ‘additions’ are made that are not recognized as the replacements they are. Given that, something must be diminished even if something is gained. Secondly, by including competing reactions that have no chemistry in common, the trade‐off depends on the relative rates and timings: it is a very fine balance, the probability of attaining which is low [17]. This is a complicated and messy system that falls between two stools. It is well recognized that ‘compomers’, where a GIC‐type glass was used as the core in a resin matrix, failed to work as hoped [13, 18] – so why now do we see supposedly light‐cured hydraulic silicate cement? Similar arguments apply, similar outcomes are to be expected. The triumph of advertising over substance? Wishful thinking is the bane of dentistry.
We see similar failures of appreciation in the seemingly random selections of additives regularly studied and proposed for many applications. For example, a material is too weak for a certain use, but a strong material is known that can be made as a powder – why not add this? Again, the replacement aspect of such a design is not recognized, but a key requirement of such composite structures is missing: bonding. Composite structures require a bond – that is, a chemical bond – between the matrix and the core (alias the ‘filler’, a term that betrays a less than honourable economic incentive in some contexts) for stress transfer to occur and the benefit to be realized. This is ‘matrix constraint’. With it, there are remarkable improvements. Without it, the material behaves as if it were full of holes, with the obvious outcome. This was seen in the attempts to strengthen silver amalgam with (silver‐plated!) sapphire whiskers, GIC with zirconia powder, and GIC with amalgam alloy powder (‘miracle mix’) to name but three egregious examples. Now we have ‘microsilica’ added to HSC. In fact, we should be careful to distinguish between materials that are included to do a job – such as reactants or bonded core – and those that have no other purpose than to dilute the system, which is all a true filler actually is. Of course, the inclusion of pharmaceutically active substances such as antimicrobials must also be treated rationally, because they are then part of the matrix (occupying volume) and therefore affect all its properties, always – poor discriminatory power experiments notwithstanding.
It is often the case that such an additive will be explored at a range of proportions (and, sadly, when it matters, not accounting for the consequent changes in other, more important, ratios). Then, through the sequential application of Student's t‐test, the maximum amount that does not give a statistically significant result (i.e. a ‘nonsignificant P > 0.05’, ‘N.S.’) will be decided upon. This is fallacious in several respects. Firstly, anything that interferes with the setting reaction or the resulting structure must, by definition, cause a deterioration. Secondly, the ‘failure to detect’ is not the same as an assertion of no effect; it is the entirely expected consequence of the poor discriminatory power of testing with a small sample size, large scatter, and relatively weak effect. Whether it matters is not the focus of attention as it should be, but the claim is made that the addition is safe because ‘nonsignificant’ is enough. In fact, with a large enough sample size, you will always get a ‘significant’ result. In addition, the test is weak because it is piecemeal instead of looking for the covariance of the outcome with input, when the full power can be obtained. It is also wrong because it is trawling without multiple test protection. What is proper is to determine the size of the effect, then determine – from other considerations – how much is tolerable. If reviewers do not understand all this, what hope is there?
1.6 Classification of HSCs
As enthusiasm for this kind of material has grown, a range of variations has been produced, but many of them make reference to the original (nonsensical) labelling (v.s.). Consequently, there is a deal of confusion as to the nature of the various formulations and what behaviour might be expected from them. There are numerous products available now, and some systematic classification would be helpful to inform product selection. The chemistry is essential to understanding what is going on – how could it be otherwise? – and thus to enabling the statement of a simple, informative, and (most importantly) accurate classification. As already indicated, an HSC is defined by the reaction of a silicate system with water (a simple classification of HSCs based on their chemistry is shown in Figure 1.1). That silicate system might be a Portland cement‐like (PC) complex mixture (regardless of its provenance, i.e. whether it is derived directly from a commercial PC or created in a similar fashion from less‐contaminated feedstock) or a simpler calcium silicate (CS) or mixture of related compounds that has been synthesized for ‘purity’ (to avoid heavy metals) or for other reasons. We therefore have a very simple division into PC and CS. The main subdivision then is based on the water that is needed for the material hydration reaction. Some HSCs are to be mixed with water (i.e. they are an aqueous mixture at the point of application), and some are presented as a slurry or paste in a nonaqueous vehicle, which is subsequently lost entirely, diffusing away. In the latter case, setting is still specifically hydraulic, as it relies on water from the surroundings (v.s. re: drying) diffusing into the water‐miscible liquid continuous phase: the setting chemistry is essentially unaffected. (Incidentally, this gives the lie to the description of such products as ‘pre‐mixed’ – quite clearly, they are not, as no reaction water is yet present. Such a presentation is a rather neat approach, so why is it trivialized by a nonsense label? This is advertising, not science.)
Figure 1.1 Fundamental classification of hydraulic silicate cements at the point of use. The types referred to for clinical purposes elsewhere in this book are shown. Type 1 is the original formulation of the class, without any additive. All others have one or more additives of various kinds, for various purposes [19].
Whether or not a radio‐opacifier is added to any of these does not affect the fundamental behaviour: the essential chemistry – hydraulic reaction – remains as the defining characteristic, whether or not that additive modifies the setting reaction rate or outcome in any way (whether deliberately or accidentally, and no matter how drastically) [20]. Indeed, whether any other additive is included or the formulation is tweaked for any reason (e.g. rate modification), the basic classification must remain if the key chemistry persists. Attempts to create ‘generations’ on a historical basis are as pointless as they are uninformative in the absence of logic, consistency, and relevance. There are no alternatives known at present for true HSCs, all of which are essentially based on CSs (as stressed elsewhere in this book).
The trouble is that what may reasonably be called ‘fake HSCs’ are also offered. Along the lines of compomer (v.s.), a PC material powder in a light‐cured resin matrix has been produced: this is a filled (composite) resin (FR), no more, no less. It is simply irrational, and highly misleading (if not culpably misrepresented), to call this an HSC, or to imply that it is by association, irrespective of any beneficial effects, perceived or claimed, from a high pH at the surface, available calcium, and so on. If setting does not depend on water, it is not an HSC. In fact, water that does diffuse into the cured material (through the matrix, as is normal and expected for such resin systems, despite popular belief) must react with the PC material: such chemistry is unavoidable. Thus, one could reasonably expect this to expand (reaction product volume is necessarily greater than PC volume), with perhaps unfortunate results, albeit slowly – but this is not ‘setting’. Similarly, a PC‐ or CS‐containing material
