The author would like to thank Kate van Lookeren-Campagne and Lisya Bacaci, both ceramic conservators and instructors at ICN and now at the UvA for their assistance in the initial development of the course, and the development of the mechanical testing specimens, Kate Seymour and her students for their assistance with adhesives at the UvA, and Robert van Langh, now head of conservation at the Rijksmuseum in Amsterdam, for introducing materials science and mechanical testing into the metals conservation curriculum at ICN back in 1996.
1When one thinks of scientific issues in conservation practice, one generally thinks of issues dealing with chemistry. The conservator is confronted with questions about which solutions are safe for treating and cleaning objects, what the chemical composition of materials are, how objects and their materials age, and what the effects of indoor or outdoor environments are on objects and collections.
2However, there are many conservation questions which concern how objects respond to physical loads, that is, questions which deal with their mechanical properties. Examples include:
Which adhesive has sufficient but not too much strength for a particular repair?
What levels of shock or vibration are allowable for a given object in transport, or for construction work near a museum?
How should an object be packaged for transport, or be protected from vibrations on exhibition?
What is the safest way to hang a wall tapestry?
How should one repair a tear in a painting canvas?
What type of material is the best for use as a loose-lining for paintings?
What are safe climate conditions to avoid cracking of wood panel paintings or furniture?
How does one apply the results of academic research and finite element (computer) modelling found in the literature to conservation practice?
3Although there are many such questions, the concepts of the mechanical properties and mechanical testing of materials play at most a minor role in most conservation science curricula. One can debate the reasons for this, but generally one can say that this is because such concepts of the field of the mechanics of materials fall outside the backgrounds of most professionals and instructors involved in art and conservation. This has resulted in many misconceptions concerning the effect of physical loads on objects, in practice and in the existing literature.
4The role that mechanical properties plays in conservation training in The Netherlands is no larger than anywhere else. However, over the past decade, the Cultural Heritage Agency of the Netherlands (RCE), and its predecessor, the Netherlands Institute for Cultural Heritage (ICN) have developed a short one to two day workshop to teach the most important and basic engineering concepts of mechanical properties necessary for conservation practice. In this workshop, the importance of understanding and using a basic vocabulary for mechanical properties and testing is emphasized. Simple mechanical tests are introduced which can also be used for many applications in the conservation studio without the need for expensive testing equipment. This combination of material provides students with sufficient knowledge to understand basic practical test results, and make informed conservation decisions in daily practice. This paper briefly describes the teaching methods used in this workshop, its advantages, but also its disadvantages when dealing with the much more complex material often found in the literature.
5As with many other subjects which a student or professional sees for the first time, the field of mechanics of materials is filled with strange new terms. It is essentially a foreign language. The workshop described here thus begins by asking students to think of
6words which they already know which could be related to mechanical properties. Some groups come up with a long list in a short time, other groups need a few hints and prodding. However, most groups come up with words such as
elastic / elasticity
7The students are then introduced to a number of basic terms in the engineering mechanics of materials. These include the concepts of stress, strain, stiffness, and the effects of temperature/expansion. These are defined, and then compared with the words which the students came up with. They are shown how improper use of terminology can lead to confusion or outright errors in calculations and the interpretation of results, and in communication with external engineering consultants. Two examples are discussed here to illustrate this.
8The first example concerns the use of the word “stress”. In the art and conservation world, the word “stress” is all too often used interchangeably with the words “force” and “load”. In engineering terms, they are not the same. The difference between the two terms can be illustrated by taking two rods of the exact same material, for example, a steel, both with the same length but one thicker than the other, see Fig. 1. One could ask, which rod is stronger? One would assume that the thicker rod (A) will carry more load before it breaks. However, the material in both rods is exactly the same, so the question arises, how can the two rods have different strengths?
Fig.1 Strength of two rods
Strength of two rods made with exactly the same material. Rod A can carry more force or load before it breaks than Rod B. That is because Rod A has a larger cross-section area to carry load; this is a geometric property of the rod. However, the stress (maximum load divided by cross-section area) is the same in both rods when they break. That is a materials property.
Credits W.(Bill) Wei
9The answer can be found by loading the rods until they break. One finds, in fact, that Rod A can carry more load before it breaks than Rod B. The force or load which one needs to break a rod depends on how thick the rod is. That is a geometric property. The value is often given by conservators in metric units in grams (g) or kilograms (kg), which is actually a unit of mass. The technical unit is Newtons (N), which is mass times the acceleration of gravity, 9.8 m/s2, also mistakenly called “g-force”.
10However, if one were to divide the maximum load that rod A could carry by its cross-section area, and compare that with the result if one divided the maximum load that rod B could carry by its cross-section area, one would get the same answer. That calculated value is the material property, strength, which is given in units of stress = force / area, that is, Newtons per square millimeter (N/mm2) or MegaPascals (MPa). A stress value is the answer one gets if one asks a manufacturer for the strength of a material they produce, or one looks up materials strength in engineering tables of materials properties.
11In practice, whether in conservation or in engineering, one of course deals with an object or product with a particular geometry, which is supposed to carry a certain force or load. However, objects come in all shapes and sizes. Although the conservator would like to know if a particular material can handle the load, materials manufacturers cannot be expected to provide a maximum load which their materials can carry for every possible geometry that comes up. They therefore give materials strength in units of stress, how much load their material can carry per unit area. In order to select the proper material to carry that load, the conservator thus needs to first calculate what the expected or allowable stress is.
12Students are given simple examples for calculations which they may encounter in practice. For example, the question is posed what the proper glue is for a joint in an historic chair or table. The expected stress on the joint is the expected load, for example, the portion of the weight of the heaviest person expected to sit in the chair and carried by the joint, divided by the area to be glued. One would then look for glues that have enough strength (stress = force/area) to handle the stress at the joint.
13It is noted at this point that the conservation literature provides many examples of interpretation problems associated with the improper use of the terms force/load and stress. Many published tests to compare the strengths of glues give results in terms of force or load, in metric terms, in grams or kilograms. The conservators or conservation scientists are often careful to use the same sized specimens, which would be fine were it not for the fact that they do not consider the possibility that a particular glue may not cover the entire surface area as intended. Some glues do not flow well, others do not wet the surface well, and others are full of bubbles. This means that the actual glued areas may be different than the intended glue area, see Fig. 2. There can be a large scatter in maximum loads that are measured for different specimens with the same glue because the handling of the glue was difficult. Furthermore, a comparison of the mechanical strength of glues using maximum load is meaningless if one glue covers the entire glue area, and another is full of bubbles and covers much less. By dividing all of the measured loads by the real glued area, one would come to a better comparison of the strengths of the glues. The final selection of the proper glue would then also consider how easy it is to work with the various types of glues.
Fig.2 Glue A and Glue B
Glue A and Glue B may have the same strength, but because Glue B is harder to apply, it does not cover the entire joint area (gray) as Glue A does. In an experiment, one would find that Glue B carries less load, but the failure stress (load divided by area) would be the same as for Glue A. The reason to reject it would therefore not be because it has less strength, but because it is more difficult to apply for this application.
Credits W.(Bill) Wei
14The second example of the confusion between engineering and cultural heritage terms is the use of the words “elastic or elasticity”, and “plastic or plasticity”. In the art world in particular, the terms elastic and plastic are descriptive words, which say something about the form and/or feel of an object, and how it is perceived by the viewer. One talks, for example, about the elasticity or plasticity of a particular sculpture or an abstract painting. The use of this language implies that elasticity or plasticity have something to do with motion or deformation.
15In engineering language, elasticity and plasticity are also words used to define deformation. However, the engineering definitions are very specific. A material, which is elastic in mechanics terms, returns to its original shape after it is unloaded. If it is permanently damaged, for example, bent, it is said to be plastically deformed. Plastic deformation is a synonym for permanent deformation.
16All materials, in particular, metals and many plastics exhibit a combination of both types of deformation. They first deform elastically, and then plastically (permanently). This property of metals has direct application in restoration practice. If one wants to bend a metal part back to its original form, one most bend it (remove the plastic deformation) beyond its desired form because it springs back slightly. That springing back is the elastic part of the deformation. A simple demonstration of this property is to ask students to bend the leg of a paperclip exactly 90° in one movement, no cheating. Experience shows that it is pretty much a matter of luck that a student achieves exactly 90° in one movement, but it becomes immediately obvious what the concept of elasticity means in practice.
17During the workshop, the concepts of stress, strain, and stiffness are demonstrated by having the students conduct simple mechanical tests. Such tests can also be used in the conservation studio for measuring the strength of materials for simple applications, avoiding the need to go to a specialized laboratory. Two types of tests have been introduced over the past few years for this purpose. Both are tensile tests, that is, they measure mechanical properties while pulling a specimen apart.
18The first type of test is for determining the tensile strength of glues, see Fig. 3a. In this test, specimens, ceramic specimens in Fig. 3a, are prepared with two holes drilled in each end. The specimens are broken in half, and then glued back together. Different types of glues are used, and the specimens are either glued across the entire fracture surface, or just half, Fig. 3b. The glues are prepared and allowed to dry according to the standards for the particular glue.
Fig.3 Specimens for testing the tensile strength of glues
a) Ceramic specimen broken in half and reglued over b) 100% or 50% of the fracture surface
Credits W.(Bill) Wei
19For testing, the specimens are hung on hooks going through one of the drilled holes, Fig. 4. A large bucket is hung on the specimen using the other hole. Water and various metal weights are slowly and carefully added until the specimen breaks. This is generally quite an exciting occurrence, even for the experienced instructor, and it is recommended that a sufficient number of bath towels be available nearby if the bucket is not stabilized in time.
Fig.4 Test for the strength of a glue
Simple test for the strength of a glue joint using water and weights as the load. Arrow points to ceramic specimen, similar to that shown in Fig. 3a. B is the bucket being filled with water and metal weights.
Credits W.(Bill) Wei
20In order to calculate the stress at fracture, that is, the strength of the glue, the students must first convert the weight required to break the specimens to force in Newtons by multiplying the weight in kilograms times the acceleration due to gravity as described earlier. The weight must include the weight of the hook and the bucket.
21The students are then instructed to divide the force by the area covered by the glue. For teaching purposes, they are asked to use two types of areas. First, they divide the force by the area that they were instructed to glue, that is, either the total area, or half of it. This gives them a “theoretical” strength, that is, the strength of the glue if it had covered the entire surface which they were supposed to glue. They are then instructed to look at the fracture surfaces of the glue in the microscope to see what the actual area was. That is, they are asked to subtract areas where glue is missing. These are corresponding areas on both sides of the fracture surface that show no glue. This calculation gives them the actual fracture stress, that is, the actual tensile strength of the glue. If the coverage is, in fact incomplete, the actual strength will be higher than the theoretical strength because the actual area covered is smaller. Note in this regard that it is better to use materials such as ceramics with low porosity or metals for this experiment so that the glue stays on the glued surface. Experience using wood samples shows that the results are much more difficult to interpret because many glues soak into the wood.
22The results of the tests are then evaluated with the students. It should be noted that because of the limited time for the tests, students generally compare only two glues with two very different strengths. There is then enough time for them to look at the effect of glued area, and to repeat one or two of the tests for a minimum of statistics. Even so, there can still be a large scatter in the results depending on how well the students prepared the samples and conducted the tests. However, experience has shown that this is a useful test for demonstrating the difference between a test result based on load, and a test result based on stress.
23The second type of tensile test which can be used demonstrates the theory of stress and strain. In this test, students experimentally determine a so-called stress-strain diagram, a diagram which shows how much a material deforms as it is increasingly loaded, see Fig. 5. With this test, they can again understand the importance of the difference between the terms load and stress, but also learn that it is important to distinguish between deformation as a change in length (for example, in millimeters), and the engineering term strain, which is percent deformation.
24In order to produce reliable results with such a test performed by hand, rubber bands are used instead of ceramics. Rubber bands deform, of course, much more than ceramics under the same loads. This makes it easier for students to take measurements by hand. Test specimens are prepared from rubber bands available at the stationary store. Wide rubber bands of a given length are preferred. Students test as-received rubber bands with the original width, and rubber bands with half of the cross-section area, that is half of a wide rubber band cut lengthwise, see Fig. 5. By testing two cross section areas, the students again see the difference between the terms load and stress.
Rubber bands with reference (gauge) lengths used for determining a stress-strain curve. The bands are made of the same material and have the same length, but different widths (cross-section areas).
Credits W.(Bill) Wei
25For the deformation/strain measurements, a reference length known as a gauge length is marked on the rubber band. This is first measured and recorded, as is the width of the specimen within the gauge length. As with the ceramics, the rubber band is hung on a hook with the reference length visible. The first load is actually the hook and bucket hung on the other end of the rubber band. The weight of the hook and bucket must therefore be known, and students are then instructed to measure the length between the two reference stripes. Subsequently, students add stepwise 100 grams of weight (100 milliliters of water), measuring the new length with each addition. In this way, they build up a table of load versus length. Six to eight data points are sufficient. It is not necessary to test the rubber bands to failure. (Rubber bands can stretch to more than their own length, that is, more than 100% strain.)
26The students then calculate and plot two kinds of graphs, one a plot of load against change in length, and one of stress against strain. (For the purposes of this publication, the mathematical details of the calculations are not reported here.) A comparison of the two types of graphs is shown in Fig. 6. The point is made again that the results shown as a load/length diagram depend on the geometry (size) of the specimen, Figs. 6a. By using stress and strain, the materials properties are visible since the curves fall together, Fig. 6b.
Fig.6 Results from tensile testing on rubber bands of the same length but different widths: Load-Deformation
Plot of load versus change in length for a thick rubber band (solid curve) and a thin rubber band (dotted curve). The difference between the two curves is due to the different cross-section areas of the rubber bands.
Fig.7 Results from tensile testing on rubber bands of the same length but different widths: Stress-Strain
Plot of stress versus strain for the thick rubber band (solid curve) and a thin rubber band (dotted curve). The calculated curves fall together because they show a material property which is the same in both cases.
Credits W.(Bill) Wei
27This paper has briefly discussed how one can teach the basics of the mechanical properties and testing of materials in a one or two day workshop. These basics include an understanding of what mechanical properties are, how they are defined, and what they mean for basic conservation practice. The key to the success of the workshop is keeping things simple, practical, and as related to conservation practice as possible. Sometimes this comes at the risk of providing an explanation which is not quite technically correct. For example, glues are usually not loaded in tension, but in shear or bending. The latter two tests are, however, more difficult to carry out and interpret for a short workshop. However, once students understand the concepts, experience shows that they catch the small deviations from truth and understand enough of the concept to ask.
28Unfortunately, not enough time is provided in conservation curricula to properly teach more complex issues in mechanical properties. Each workshop thus closes with a presentation of a number of other more complex concepts including internal stresses due to temperature and humidity changes, shock loading, and the cumulative effect of vibrations. Practical examples are discussed in brief, including the cracking of cradled panel paintings, the hanging of (heavy) textile objects, the restoration of slashed paintings, and the effect of transport or construction vibrations on the condition of sensitive objects. The goal is to show students that a good, basic understanding of the mechanical properties of materials can often lead to relatively simple solutions for many such complex mechanical problems. The students are encouraged to consider these issues in their further studies, possibly including them in their final diploma projects.
29The basic concepts of the workshop are now well established. Each year new examples and variations on the concepts are introduced in order to improve the course but also to provide variety from year to year. The workshop has been offered at the University of Amsterdam (UvA) and its predecessor curriculum at the Netherlands Institute for Cultural Heritage since 1996, and has been given in the conservation curriculum at the University of Applied Sciences (HTW) in Berlin since 2011.
30A short course on the mechanical properties and testing of materials in conservation has been developed over the past decade by the Netherlands Institute for Cultural Heritage (ICN), now part of the Cultural Heritage Agency of the Netherlands (RCE). The goal is to provide students with a basic understanding of what mechanical properties are, how they are used in engineering practice, and how this can be translated to their conservation practice. The course has been successful due to its simplicity, and its emphasis on proper terminology and practical experience. It is hoped that other conservation curricula will recognize the need for teaching in mechanical properties, and provide more than just one or two days for teaching the necessary material.