Abstract
Objectives
To characterize the development of strength during the setting process of dental silver amalgam in the context of ‘early strength’ measurements for standards compliance testing in relation to patient instructions, and demonstrate the applicability of the Hertzian ‘ball on disc’ method.
Materials and methods
Sixteen dental silver amalgam products were tested using the ‘ball on disc’ protocol at 1, 2, 3, 4 and 24 h after setting at 37 °C in air. The mixed materials were packed into a tapered steel disc mold (10 mm diameter, 3 mm thick) resting on a glass surface, slightly overfilled and carved level with a sharp edge, then ejected at ∼10 min and placed immediately into an incubator at 37 °C. Testing was in Hertzian mode, using a 20 mm steel ball, with the specimen resting on a disc of glass-filled polyamide ( E = 10 GPa) at a cross-head speed of 0.2 mm/min on a universal testing machine (E3000, Instron). The load at first crack was recorded, as was the number of radial cracks produced.
Results
Radial cracking into 2–5 pieces, in a clinically-relevant (non-explosive) mode was observed in all cases. Considerable variation in setting rate between products, as indicated by the development of load at failure with time, was found. The distribution of normalized failure load values overall was lognormal (Weibull was excluded). The RMS coefficient of variation overall was 12.4%.
Significance
The ball-on-disc test provides a facile, relevant measure of the strength of dental silver amalgam, and is viable as a standards compliance test. Early strength testing at a minimum of 2 h is suggested.
1
Introduction
Dental silver amalgam continues as a mainstay of economical restorative treatment in various jurisdictions, despite intensifying discussion of the collateral environmental problems arising from mercury-containing waste . Safety has been debated since the 1840s, and continues unabated today , although the recognition of the value of the material remains . A key aspect has been whether filled-resin materials can replace amalgam , but this is considered by some experts not yet to be the case .
Accordingly, it is appropriate that the properties and behavior of such products continue to be subject to scrutiny vis-à-vis treatment efficacy, and thus tested for compliance with agreed standards. It follows immediately from this that the work of the International Organization for Standardization (ISO) Technical Committee for Dentistry (TC106), and of similar national bodies, in this area ought to continue to provide guidance as to the minimally-acceptable performance in the various aspects, and as to the tests that may be used to ascertain compliance . Early strength is one such key property.
Unlike many materials used in dentistry which set through a chemical reaction, dental silver amalgam is (relatively) slow. For example, as judged by dimensional change at 37 °C, the process has not stopped at 150 h . Certainly, tensile strength has been shown to increase to at least 100 h , and to 7 days in one instance at least , similar to results for ‘compressive strength’ . Transverse bend strength for one alloy increased to at least 3 months . However, some results are far from clear for the interval 1–7 days . Very early strength has been noted to be relevant to first contact as part of the treatment process, i.e . for occlusal adjustment, at 15–20 min . Other than that, little data has been published ( Figs. 1 and 2 ) , a point noted elsewhere. Nevertheless, it is the early strength that is of primary importance, no matter the final value, since it is the vulnerability to premature masticatory forces that is presumed to be of most concern . This low early strength is generally recognized, but the implications clinically only rarely get mentioned . The instructions to the patient regarding when it might be ‘safe’ to bite on a new restoration seem to be part of the oral tradition of teaching in dentistry, and vary between operator, but are always vague, with no known documentation, although “everybody says it”. This situation may arise from the absence of specific evidence, but such instructions do appear to be eminently sensible and nobody disputes their wisdom in principle. Even so, those few published studies of the development of strength have not been focused on this topic and so have made no apposite remarks. For comparison, Philips is reported to have advised that 8 h should elapse before biting, while on limited data a possible reduction to 6 h was suggested . Secondly, since there is no absolute means of determining a minimum appropriate strength for amalgam (or indeed any material) there can be no absolute threshold for the strength at any chosen time after mixing or placement: how strong is strong enough? Ultimately, of course, it is unanswerable because in the engineering terms appropriate to the assessment of an object, it is the combination of strength and section in the context of the mode of loading that controls the outcome. Section depends on many factors outwith material considerations, and although it is to be hoped that the appropriate design is made to suit the material (or vice versa ), this engineering of restorations is not part of clinical teaching as such, only rule of thumb learnt (effectively) by apprenticeship is applicable or practiced.
This absence of clarity is therefore not surprising. Nevertheless, the relevant ISO standard refers to a minimum “compressive strength” of 80 MPa at 1 h (and 300 MPa at 24 h, to put this in perspective). Notwithstanding the pragmatic and political aspects of the drafting of standards , the choice of testing at 1 h appears to be unsupported by any documentation or other evidence (never mind the value). It would seem that the only explanation is that it is a ‘convenient whole number’: it has no intrinsic meaning, as with many such test times. This is perhaps more related to the (understandable) desire to create simple work schedules in the test house as it is to clinical reality (for example, there are not many studies that report values for anything at all between about 8 and 16 h after preparation, nor requirements to obtain such values).
The use of “compressive strength” is another matter of convenience. As has been pointed out , there is no such thing, and although the terms “bearing capacity” or “crushing strength” could be used it does not affect the fact that crack initiation is internal and in shear. So-called indirect tensile strength ( alias diametral compression), often vaunted as the solution to the logical difficulty, is in fact the same class of test in all respects and offers no escape (other than scale, Figs. 1 and 2 show much the same proportion of the 24 h value at 1 h) . The problem, however, is that the mode of failure in any such a test bears no relation to the clinical context, where explosive fragmentation is not known ever to have occurred. There are other problems. Parasitic stresses at the platen interfaces, due to various irregularities (wear, distortion or lack of parallelism of the platens; irregularities or non-parallelism of the specimen faces or axial meridians), cannot be avoided without going to extraordinary lengths. Platen-specimen friction is also instrumental in changing the outcome: any attempt to modify this further complicates the behavior .
A further difficulty with the existing ISO standard , which was derived from the equivalent American Dental Association standard , is that the specimen preparation requires an elaborate ‘one-shot’ procedure, in a complex mold, to eliminate the inevitable variation in the product between operators and the consequent sensitivity of the test result to the internal condition of the specimen (especially porosity). Primarily, this is to obtain consistency between test houses. Unfortunately, this also results in a specimen of variable length (which on its own affects the test result) as the consolidation and consequent mercury expression are dependent on both mixing ratio and particle interactions, how well they ‘settle’ under a static load. The product thus bears no relation to the material as employed in service.
This problem of realistic test loading to obtain realistic failures has been discussed elsewhere . Essentially, simple tensile failure in a bending mode from cuspal loading over a relatively compliant substrate (compliant, that is, in comparison with the hard alloy platens ordinarily used in testing machines) better represents the ordinary outcome and service conditions. This may be mimicked by using the so-called “Hertzian contact” or ball-on-disc test ( Fig. 3 ), which has been used extensively for ceramics ( e.g . ) but has also been shown to be effective on a variety of other restorative materials and especially silver amalgam . Radial dissection into 2–5 pieces, initiated on the lower surface, is obtained (rather than top surface-initiated cone-cracking) .
Fig. 3 shows the principal features of this mode of testing. A disc of the test material, of a diameter large enough in relation to the thickness that the test result is little affected by increase, is rested on a substrate which is relatively compliant, i.e. so that E sub < E test , here nominally intended to represent dentin. The dimensions are, of course, essentially arbitrary, but partly constrained by the need to control failure mode and the notion that failure loads should as far as possible be of the magnitude of those relevant in the mouth. However, the nature of the test is such as to render calculation of the stress at the point of failure impossible without resorting to extremely elaborate and individual calculations, and this would defeat the object of a routine quality-control test. Nevertheless, the failure load, in the sense of the bearing capacity of the disc under the chosen conditions, provides a fair figure of merit that, one assumes, is a reasonable proxy measure for the tensile stress then.
There are several features of the test that increase its usefulness. Firstly, the lower surface – at which the desired radial cracking is initiated – is prepared directly against some suitable matrix (such as a glass microscope slide) such that it represents the natural condition as set. That is, it shows the material’s intrinsic surface structure and texture, without spurious flaws being introduced from any finishing procedure. In this sense it represents the state in actual service at a cavity floor, albeit free of gross irregularities. In the case of silver amalgam this is complicated by the ‘loss of gloss’ phenomenon due to volumetric changes on setting , but similar behavior can be expected in a number of other cases. Secondly, and in a complementary sense, the upper surface is not involved in the failure initiation process and therefore can be finished in any convenient way without having to worry about the flaws thereby created. Thirdly, the diameter of the specimen is chosen to be large enough that packing flaws and chips on removal from the mold in its early weak state have no bearing on the crack-initiation process through any appreciable change in the stress state of the central region of the lower surface. In addition, there is no critical alignment concern, either in terms of parallelism or exact centrality of the ball contact point. Crucially, the method is not technique-sensitive beyond a standard skill required of a dentist, and does not require the completely unrealistic ‘one-shot’ specimen preparation technique of current standards . Finally, the apparatus required is rather simple, the only special aspect being the substrate. This set of attributes distinguish it from many other tests where specimen perfection, particular expertise, and strict (but irrelevant) protocol are critical to the outcome.
The present purpose therefore is to employ the Hertzian contact, ball-on-disc test to ascertain its suitability for standards compliance testing of dental silver amalgam, and in particular for early strength, and to identify, if possible, an appropriate time for such a test to be applied.
2
Materials and methods
The dental silver amalgam products tested are shown in Table 1 . Pre-dosed encapsulated products were mixed as supplied. Where instructions were available for bulk powder and tablet products, they were followed using a screw-cap reusable capsule (SS White) and a 7.97 mm diameter steel ball as a pestle for efficiency (Chap 15, section 7.3 ). Some were very old ‘archival’ materials and no instructions were available, in which case a suitable mixing ratio was identified by preliminary trials. All products were mixed under the same conditions (Silamat Plus; Vivadent, Liechtenstein) for 8 s, on the ‘slow’ setting, having ascertained that a clinically-usable mix was obtained in each case. The mixed material was packed by hand into a steel disc mold, 10.0 mm diameter, 3.00 mm thick, resting on a glass plate, using a conventional 2 mm diameter condensing point, overfilling slightly, then carving back smooth and level using the edge of a glass microscope slide. In the case of Megalloy EZ and Septalloy NG50 the product was available in both 1- and 2-spill capsules, and a full series of specimens was prepared for each. The material was allowed to set for ∼10 min, then carefully ejected from the mold and placed immediately into an incubator, in air, at 37.0 °C.
| Brand | Manufacturer’s description | Manufacturer | Batch id. |
|---|---|---|---|
| Amalcap Plus | 1 Spill, non-gamma2, regular | Ivoclar Vivadent, Sweden | F55151, NT4008 |
| Ease | Dispersed phase, pellets, type 2 | Caulk, USA | 112177, 011178 |
| DCL Amalgam Cap | 2 Spill, gamma 2-free, dispersed phase | Dental Composite, UK | 037335 |
| Dispersalloy | 2 Spill, regular set, admixed | Dentsply Caulk, USA | 090828 |
| Dispersalloy | Dispersed phase, fast set, zinc, tablets | Johnson & Johnson, USA | 6E063 |
| Epoque 80 | Non-spherical, non-gamma 2 | Scania Dental, Sweden | 7-217 |
| Megalloy EZ | 1 Spill, high strength, high copper, spherical | Dentsply Caulk, USA | 0803241, 091102 |
| Megalloy EZ | 2 Spill, high strength, high copper, spherical | Dentsply Caulk, USA | 1101251 |
| Scanalloy 68 | Fine grain | Scania Dental, Sweden | – |
| Septalloy NG 50 | 1 Spill, single composition, non-spherical, non-gamma 2 | Septodont, France | 07079C, 09002A |
| Septalloy NG 50 | 2 Spill, high copper, single composition, non-spherical, non-gamma 2 | Septodont, France | 07084B |
| Supralloy | High copper, spherical blend, type 1, class III | Unident, USA | 032282 |
| Token No 1 | Zinc, non-spherical | Toyo Chemical Laboratories, Japan | 2042 |
| Toyo DP | Dispersed phase, spherical, non-zinc | Toyo Chemical Laboratories, Japan | 910-1021 |
| Toyo Atomised | Spherical fine cut, non-zinc | Toyo Chemical Laboratories, Japan | 891-1112 |
| Vivacap | 2 Spill, high copper, single composition, non-gamma2 | Ivoclar Vivadent, Sweden | MT4065, MT 4072 |
Specimens were removed from the incubator 10 min before the time of test in order to cool. They were then tested using the apparatus shown in Fig. 3 on a universal testing machine (Model E3000; Instron, High Wycombe, UK) at 0.2 mm/min cross-head speed in air at 23 °C, resting freely on the substrate, a 5 mm-thick disc, 10 mm diameter, of 30% glass fiber-filled polyamide (LS360362 SJP; Goodfellow Cambridge, Huntington, UK) 1
1 The catalogue number appears to have been changed; it is now at the time of writing AM367910 (but the same order code, 791-942-57), described as “Polyamide – Nylon 6, 6–30% Glass Fiber Reinforced – Rod, PA 6,6 30% GFR.
, faced flat. The relative dimensions of specimen, substrate and ball were chosen to ensure radial cracking initiated from the lower surface of the specimen ( i.e . at the interface with the substrate). That is, Hertzian cone-cracks and ring cracks must not (and do not) occur for validity. Testing was done at 1, 2, 3, 4 and 24 h after the start of mixing, all started within −0, +2 min of the target time. Substrate disks were replaced if any wear or deformation was detected (which was rare). The data acquisition rate was set to 100 s −1 . This enabled the detection of the first crack through the ringing of the load signal ( Fig. 4 ) without employing the acoustic emission apparatus previously described , although it was generally audible. Occasionally, for the weakest specimens, the crack was so ‘soft’ that no ringing was detected, although radial dissection was found to have occurred. Inspection of the load-time plot mostly allowed the identification of the fracture point by a slope discontinuity. Very rarely, even this was not possible and no fracture load value could be recorded. The number of pieces into which the specimen fractured was recorded, confirming that radial dissection had indeed occurred, i.e . bottom-initiated cracking. This required the use of some finger pressure as the pieces were often held together by the intimate tortuosity of the fracture surfaces; incomplete fracture could not be excluded in such cases (the crack was often not detectable visually).
Statistical analysis was in software (SigmaPlot 11.2; Systat Software, San Jose, CA, USA).
2
Materials and methods
The dental silver amalgam products tested are shown in Table 1 . Pre-dosed encapsulated products were mixed as supplied. Where instructions were available for bulk powder and tablet products, they were followed using a screw-cap reusable capsule (SS White) and a 7.97 mm diameter steel ball as a pestle for efficiency (Chap 15, section 7.3 ). Some were very old ‘archival’ materials and no instructions were available, in which case a suitable mixing ratio was identified by preliminary trials. All products were mixed under the same conditions (Silamat Plus; Vivadent, Liechtenstein) for 8 s, on the ‘slow’ setting, having ascertained that a clinically-usable mix was obtained in each case. The mixed material was packed by hand into a steel disc mold, 10.0 mm diameter, 3.00 mm thick, resting on a glass plate, using a conventional 2 mm diameter condensing point, overfilling slightly, then carving back smooth and level using the edge of a glass microscope slide. In the case of Megalloy EZ and Septalloy NG50 the product was available in both 1- and 2-spill capsules, and a full series of specimens was prepared for each. The material was allowed to set for ∼10 min, then carefully ejected from the mold and placed immediately into an incubator, in air, at 37.0 °C.
