Book and Paper Group Annual
Volume 17 1998
The American Institute for Conservation

Predicting Failure in Hinges: Measurement of a Lap/Shear Bond between Japanese Tissue, Wheat Starch Paste, and Oil-Saturated Newsprint

Ann M. Baldwin

Abstract

Traditional wheat starch paste is a reliable and reversible adhesive used for archival hinging of works of art on paper. Yet, this time-tested, water-based adhesive does not always bond well to oil-saturated paper, as do several nonaqueous synthetic adhesives. However, due to problems with reversibility, aging characteristics, and the increased awareness of occupational risks associated with organic solvents, conservators do not generally recommend synthetic adhesives for long-term use on paper supports. In response to this problem, an experiment, modified from TAPPI Tensile Test T 813 om-91, was designed to measure the strength of wheat starch paste bonds when used to attach Japanese tissue hinges to aged, oil-saturated, newsprint. Commercial iron-oxide oil paints were applied by brush in a painterly, variegated pattern to sheets of wood pulp paper using two different dilutions: straight from the tube, and thinned with a commercial oil painting medium. The uneven paint coverage allowed comparison of the paste bond to oil-saturated paper relative to that of plain, unpainted, paper. Twenty-five 16" X 12" samples were light-aged naturally for eighteen months, and dark-aged for almost thirty-three months. Japanese tissue hinges were then attached to samples using cooked, strained, and thinned wheat starch paste. To secure the attachment, pressure was initially applied with a bone folder. Hinged samples were then placed under weights, using non-woven polyester webbing as interleaving, and desiccated blotters to absorb moisture. Lap/shear tensile tests were conducted at the Testing Facility of International Paper Company, Tuxedo Park, New York, using an Instron Model 1122 with a 5500R Series system interface, and Series IX software. Results indicate that while wheat starch paste remains a preferred adhesive for most paper supports, it is inadequate when used on oil-saturated paper.

1. Introduction

As a medium, oil paint has been employed on paper throughout the ages by artists ranging from Tintoretto, Jordaens, and Rubens, to deKooning, Kline, and Frankenthaler (Kosek and Green 1992), (Cove 1992). Realized as sketches, studies, or finished works, oil paintings on paper are found on both primed and unprimed paper supports.

Graph
Fig. 1a. Side view of hinged work on paper
Graph
Fig. 1b. Verso of work of art with two hinges attached

Graph

Fig. 2. Work of art on paper hinged with Japanese tissue and wheat starch paste, attached to rag mat board backer

When displaying these works, presentation formats may vary according to artistic intent, custom, and curatorial preferences. Works on paper have often been exhibited as traditional canvas easel paintings, by being mounted directly to paper, board, or canvas, and then wrapped around stretchers, or strip-lined and secured to the reverse of a solid support. Alternatively, such works may hang, suspended only from their top edges by strips of tissue, which are in turn, attached to a secondary support such as four-ply rag board. Typically, two or more long-fibered Japanese tissue "hinges" will be used to suspend the work of art on paper within a window mat. In this hinging process, the Japanese tissue is secured to the reverse of the painting with cooked, strained, and thinned wheat starch paste (Figures 1a-b and Figure 2).

Starch pastes have a long and successful tradition as adhesives in Eastern and Western art conservation and mounting (Winter 1984). Owing to their steric compatibility with cellulosics, and their relative ease of reversibility, starch pastes remain the preferred adhesive in most paper conservation procedures (Swanson 1972). Yet, empirical results show that this water-based carbohydrate adhesive does not consistently bond well to oil-saturated paper (Keynan and Weingarten 1991).

In order to hinge or line oil-saturated paper, conservators utilize a number of natural and synthetic adhesive alternatives, in combination or alone. For these hard-to-adhere joins, conservators have suggested mixtures of starch paste and gelatin (Fishman 1993); or mixtures of starch paste and poly(vinyl acetate) emulsions (Zukor 1998). Heat- and solvent-activated synthetic adhesives such as Beva 371 (Perry and Townshend 1992) and Lascaux 498-20X (Maheux 1992) have also been advocated. In addition, the use of desiccated blotters is recommended to absorb excess moisture at the bonding site (Volent 1994). Conservators advocate using non-toxic, reversible adhesives for the health and safety interests of both the conservator and the works of art.

One-of-a-kind or specially made paper/synthetic adhesive hinge systems may be indicated for temporary use on supports to which wheat starch paste will not adhere, or where, because of the nature of the support, moisture cannot be used for hinge application or later hinge removal (severe cockling) or for display of transparent objects. Caution: Synthetic adhesives are not recommended for permanent hinging and should only be used if they can be safely removed from the paper, i.e., solvents used for removal are safe to apply to the object. (Book and Paper Group/AIC 1988).

A preliminary literature search indicates that no adhesive is specifically recommended for hinging oil paintings on unprimed, i.e., oil saturated, paper supports.

2. Experimental

This experiment was designed to determine whether wheat starch paste, prepared according to standard practice, will bond and hold a Japanese tissue hinge to aged, unprimed, and oil-infused, newsprint paper. The objective of these tests is to recommend, or rule out, the reliability of wheat starch paste for use on oil-saturated paper supports, and to indicate other techniques and adhesives for future testing.

TAPPI Tensile Test T 813 om-91, tensile test for the manufacturer's joint of fiberboard shipping containers, was modified to test the efficacy of lap/shear joints in the archival hinge.

Graph

Fig. 3. Dimensions of test specimen

The arrangement of a "T-hinge" attached to a paper support with paste closely resembles the configuration of a lap/shear joint (Bikales 1971). (See Figure 3). [The "V-hinge" is another type of hinging configuration. It is subject to peel forces, and was not tested in this study.] The performance of conservation-grade adhesives has been studied in both lap/shear and peel tests as published by paintings conservators (Berger 1972 ), (Hedley 1984), and (Katz 1985)

2.1 Sample Preparation and Materials: Paper and Oil Paint

Focusing the study on modern and contemporary works of art on paper, the test samples were prepared using commercially-manufactured artist materials. Sheets from a pad of 16" X 12" Morilla—newsprint paper by Canson were designated as the support paper; Utrecht oil paint, prepared with linseed oil and packaged in tubes, was chosen as the test medium. The paints included two different iron-oxide pigment types, Yellow Ochre and English Red Light. They were applied by brush in two dilutions—in one group, paint was delivered straight from the tube, and in the other group, paint was first thinned with Winsor Newton brand oil painting medium, a mixture of stand oil and petroleum distillate.

Yellow ochre is a natural earth pigment with the chemical composition, Fe2O3 . H20, and may contain impurities such as gypsum and magnesium carbonate (Gettens and Stout 1966). With a medium oil index of approximately seventy-six pounds of oil per one-hundred pounds of pigment, it forms an excellent, strong paint film, and has a slow-to-average drying time. English Red Light is a descriptive name for maroon-colored anhydrous iron oxide, Fe2O3. It is a naturally-occurring earth pigment, and can be prepared chemically by heating ferrous sulfate with chalk; it usually contains gypsum. The label on Utrecht English Red Light describes the pigment as artificially-prepared. With a low to moderate oil index of about sixty-three pounds of oil per one-hundred pounds of pigment, English Red Light forms a good to fair (sometimes brittle) paint film, and has an average drying time (Mayer 1970), (Gottsegen 1993).

Working on a surface of prepared palette paper, the tube paints were thinned with Winsor Newton painting medium using a palette knife. During this process, it was observed that English Red Light wetted more easily, and became a less-viscous, more easily-manipulated paint mixture compared to the Yellow Ochre tube paint when prepared in a similar manner. Thinned English Red Light paint resembled an ink, in both consistency and flow properties, in contrast to the Yellow Ochre paint, which was found to be pasty, stringy, and more difficult both to dilute and to apply by brush. The greater 'wettability' of English Red Light, a synthetically-prepared pigment, seems to be due to its minute particle size.

For each paint type and dilution, a sheet of 16" X 12" newsprint was divided into four 8" X 6" quadrants, and lightly ruled with soft graphite pencil. Each sheet was oriented with the grain in the vertical direction, and painted. To replicate a painterly application, media was applied with a 2" wide flat bristle brush in a variegated fashion, i.e., in some cases the media reached the edges of the paper, and in some cases, the media did not. There was a range of paint loading and saturation.

2.2 Natural Light Aging

Painted Samples

Fig. 4. Four panels of painted samples at the studio of Rustin Levenson Art Conservation Associates, New York, New York, before light-aging

Painted samples were stapled onto the walls of the studio of Rustin Levenson Art Conservation Associates, New York, New York in order to be light-aged (fig. 4). Facing south, the room receives ample sunlight throughout the day, and is exposed to a range of relative humidity and temperature levels. The samples were naturally light-aged from October 15, 1993 to April 11, 1995 for a total of eighteen months, four days. The samples were then removed, placed in flat storage with glassine interleaving, and dark aged from April 5, 1995 to December 26, 1997—thirty-two months and twenty-one days.

2.3 Hinge and Paste Preparation

In December, 1997, the 16"(H) X 12"(W) sheets were each cut into four 8"(H) X 6" (W) pieces using a Fletcher brand table mat cutter # 2100. Based on weight, strength, and handling properties, light-weight Sekishu Hanshi-mare tissue manufactured by Hiromi was selected as the hinge paper for this study. After aligning chain lines to correspond with the grain direction of the sample support paper, 2 3/4" (H) X 6" (W) strips of the Japanese tissue were water-torn. These dimensions allowed for a hinge/paste/newsprint lap-join with a 3/8" overlap, running the width of each trimmed sample. [Recommended paste coverage varies in height from one-quarter to one-half inch (Book and Paper Group/AIC 1988)].

A mixture of Zin Sho fu wheat starch and distilled water was prepared in a ratio of 1:4.5 starch to water by volume, first through hydrating, and then by cooking using a Tefal® Cook & Stir—electric gravy maker. The cooking time was approximately forty-five minutes. During the cook, the temperature control was increased gradually to a setting midway between "4" and "5". After monitoring three distinct stages of the cook: 1) transformation from opaque to translucent mixture, 2) bursting of starch grains, and 3) change in viscosity from thicker to less thick gel, the heat setting was gradually lowered to zero. The cooked paste was then transferred to a sterilized glass container, and cooled in a water bath (Clapp 1987). After straining through a fine mesh plastic sieve, the paste was thinned with distilled water to a moderate-thin, heavy cream consistency (Phibbs 1994). Paste was made fresh on the day of hinging.

2.4 Hinging

Hinging took place in winter, 1998, in a heated room with an atmosphere approximating the preconditioning environment outlined in TAPPI T 402 om-88, standard conditioning and testing atmospheres for paper, board, pulp handsheets, and related products. Using humidity indicator strips, the relative humidity was measured at nearly forty per cent, just above the recommended relative humidity levels of 10-35%. The temperature measured 24.4 degrees C, within the specified range of 22-40 degrees C (72-104 degrees F).

Using a Mylar® template under the water-torn edge of the Sekishu Hanshi-mare hinging tissue, paste was applied by brush to the top three-eighths of an inch of the hinge paper. After air-drying for a few seconds, the pasted tissue was positioned on the top edge, verso, of the painted sample. Pressure was applied with a bone folder to help secure the bond. Polyester webbing was then positioned as a non-stick interleaf, followed by placement of desiccated blotter squares, used to wick off excess water. During the first hour, blotter squares were changed five times; in the five subsequent hours, blotter squares were changed one per hour; the hinges remained under weights overnight.

Hinges were applied four days prior to tensile testing. Batches were numbered and separated into two groups—one for current testing and one for future testing—by a coin toss. The samples were placed into an air-tight plastic bag separated by one-ply rag board.

3. Testing

The experiment was conducted at the Testing Facility of International Paper Company, Tuxedo Park, New York, on January 6, 1998. Testing was performed under the direction of paper scientist, Dennis Crawshaw, and with the help of tensile test technician, John Conklin.

3.1 Conditioning Environment

The experiment was undertaken in the test environment of 50% and 73% F, within the range specified in TAPPI T 402 om-88, standard conditioning and testing atmospheres [relative humidity of 50% ±2%, and a temperature of 23.0 ±1.0 degrees C (73.4 ±1.8 degrees F)].

3.2 Sample Trimming

Prior to tensile testing, each hinged sample was trimmed uniformly using a Thwing Albert 1" precision cutter model JDC 1-12, resulting in five one-inch-wide (25.0 ±0.5 mm) replicants per each 8" X 6" sample. Within each group type (I-IV), samples were labeled at the top of the hinge, from left to right, with letters "a" through "e" using a soft graphite pencil. The trimmed specimens were then conditioned to ambient levels of temperature and humidity for three hours.

3.3 Test Groups

The test groups are depicted in Table 1 and as follows:

I YELLOW OCHRE STRAIGHT
II YELLOW OCHRE THINNED
III ENGLISH RED LIGHT—STRAIGHT
IV ENGLISH RED LIGHT—THINNED

As suggested by conservation literature, a pre-test trial was conducted (Bradley 1984). This initial trial batch of twenty specimens—five of each variant—was run to test the effects of different grips (both the one-inch flat-head and the half-moon shaped grippers), at varying distances from the lap-joint (1/8" and 1"), and at different cross-head speeds (0.25 in/minute and 0.5 in/minute).

3.4 Procedure

The test group of forty specimens—ten specimens of four variants each—were then tested. Lap/shear tensile tests were carried out on an Instron Tensile Tester, Model 1122 with a 5500R Series system interface, and Series IX software. The trimmed samples were positioned between two 1" flathead grippers outfitted with textured rubberized padding. Initial distance between jaws measured seven inches. Cross-head speed was set at 0.5 inch per minute, and the load cell was calibrated for a full scale load range of 224.809 lbf.

During each test, the cross-head movement is programmed to stop at the moment of material failure—whether hinge, adhesive, or painted support. The force required to create the failure is measured in pound-force per inch (lbf/in) or kilo-Newtons per meter (kN/m), and recorded for each sample. During operation, the lower cross head moved in a downward motion, drawing the painted newsprint support away from the hinge, which was held stationary at top. The cross heads return automatically to the initial distance of seven inches apart, the broken sample is removed, and the next sample is inserted.

4. Results

After testing, all samples were examined and data were recorded according to TAPPI Test T 813 om-91, Tensile test for the manufacturer's joint of fiberboard shipping containers. Observations and calculations included type of failure, average load at failure with standard deviation, degree of oil saturation in sample, and experimenter's notes.

Graph

Fig. 5. Rate of failure per forty test separations—number of separations per type of failure

Graph

Fig. 6. Average load at failure—load in kN/m vs. type of failure

Of the forty tensile test separations, thirty resulted in adhesive failure, seven in hinge failure, and three in support failure (See Figure 5). In most instances, it took less load to break the adhesive bond than to produce other failure types. Results reported in Table I list samples in order of increasing load-force (in kN/m and lbf/in) required to cause failure. Load forces at failure ranged from a low of 1.323 kN/m, for an adhesive failure, to a high of 2.778 for a support failure. The mean or average failure of all types was 2.010 kN/m; the mean adhesive failure measured 1.921 kN/m; the mean hinge failure, 2.142 kN/m; and the mean support failure, 2.585 kN/m. (See Figure 6) This last value corresponds to industry tensile testing results, where at similar cross head speeds, 1" wide strips of commercially-manufactured, wood fiber-based paper typically undergo cohesive failure at loads from 14 to 15 lbf/in, or 2.452 to 2.628 kN/m. (Conklin 1998. Personal communication) [A published standard is not available]

In examination of the samples, it became evident that the degree of oil saturation is closely related to failure type and average load at failure (Table VI). Based on paint coverage, samples were divided into four classes of oil saturation: 1) full, 2) variegated, 3) low, and 4) negligible or none. The designation 'variegated' describes samples with an uneven paint application with three distinct zones: 1) oil saturation, usually full, directly associated with paint coverage, 2) oil haloing emanating from the painted region, and 3) plain, unpainted newsprint paper. In variegated samples, failure occurring in the wheat starch bond corresponds directly to painted surfaces, and is discerned readily from intact bonds corresponding to unpainted surfaces (Figs. 7-8).

Fig. 7-8. Recto and verso of a variegated sample #II.20.d. Failure occurring in the wheat starch bond corresponds directly to painted surfaces and is discerned readily from intact bonds corresponding to unpainted surfaces.
Photograph
Fig. 7. Adhesive Failure and Support Failure
Tear at interface between painted and unpainted paper
Photograph
Fig. 8. Adhesive Failure and Support Failure
Tear at interface between painted and unpainted paper

 

Fig. 9-10. Recto and verso of a sample #II.22.e. Full oil saturation corresponds with complete adhesive failure
Photograph
Fig. 9. Complete Adhesive Failure
Photograph
Fig. 10. Complete Adhesive Failure

 

Fig 11-12. Recto and verso of sample #II.21.a. Negligible oil saturation corresponds with support failure at high load
Photograph
Diagonal Break—1" from bottom edge of hinge. High load at break
Photograph
Diagonal Break—1" from bottom edge of hinge. High load at break

When load was applied to oil-saturated samples (full and variegated), the lap/shear joint failed thirty out of thirty-three times. Adhesive failure clearly correlates with oil coverage. For example, in samples with full oil coverage, eight out of eight test pulls resulted in adhesive failure (Fig. 9-10). In samples with variegated oil saturation, twenty-two out of twenty-five test pulls resulted in adhesive failure. This is in contrast to the seven samples with low to negligible oil saturation. In samples with low oil saturation, four out of five test pulls resulted in hinge failures; and one resulted in support failure (Fig. 11-12). Lap/shear testing of the two samples with negligible or no oil saturation resulted in one hinge failure, and one support failure.

Each of the three support failures occurs in Yellow Ochre specimens. Ruptures take place within 1-2" from the upper edge of the support—from ~1/4" to 1" below the bottom edge of the hinge joint. At bonding sites, these supports range in degree of oil saturation: one is characterized as variegated, one as low, and one as negligible. Two of the three failures occur at high load forces: 2.726 kN/m (15.567 lbf/in) and 2.778 kN/m (15.862 lbf/in).

Photograph

Fig. 13. Average load at failure—load in kN/m vs. degree of oil saturation

Test data reveal another consistent relationship between degree of oil saturation and average load at failure. As noted in Figure 13, the greater the oil saturation, the lower the average load at failure; and the lower the oil saturation, the greater the average load at failure. The average load at failure for full oil saturation is 1.763 kN/m; for variegated oil saturation: 2.003 kN/m; for low oil saturation: 2.237; and for negligible/no oil saturation: 2.509.

As demonstrated in Table VII—Summary of Lap/Shear Tests of Wheat Starch Paste on Oil-Saturated Paper—loads at failure were greater on average for Yellow Ochre samples than for English Red Light samples. The combined average load at failure for Yellow Ochre samples, Groups I and II, amounts to 2.152 kN/m (12.285 lbf/in). The combined average load at failure for English Red Light samples, Groups III and IV, is 1.877 kN/m (10.715 lbf/in), a 12.8% decrease in load force.

Analysis of data in Table VII—Summary of Lap/shear Tests of Wheat Starch Paste on Oil-Saturated Paper—revealed an inverse relationship between degree of dilution (thinning) of the paint and the average load at failure, measured in kN/m. In the Yellow Ochre samples, the average load at failure for paint straight out of the tube (group I) is 2.354 kN/m, while the load of the thinned paint samples (group II) averaged 2.778 kN/m. If this trend had been consistent, the English Red Light groups should follow the same pattern. However, for samples with paint straight out of the tube (group III), the average failure measured 1.786 kN/m, less than that of thinned paint samples (group IV) with an average load at failure of 1.967 kN/m.

Photograph

Fig. 14. UV/VIS Fluorescence. From left to right: Recto of samples I.13.d.—straight, II.20.d.—thinned,III.22.c.—straight, IV.21.a.—thinned.

Photograph

Photograph

Fig. 15-16. UV/VIS Fluorescence. From left to right: Verso of samples I.13.d., II.20.d.,III.22.c., IV.21.a. The more intense UV/VIS fluoresce reaction indicates that paint straight from the tube has a greater oil concentration than paint thinned iwth Winsor Newton oil painting medium.

Photograph

Fig. 17. UV/VIS Fluorescence. From left to right: Recto of samples I.17.b.—straight, II.20.c.—thinned, III.21.d.—straight, IV.15.b.—thinned.

Photograph

Fig. 18. UV/VIS Fluorescence. From left to right: Recto of samples I.17.b., II.20.c., III.21.d., IV.15.b. The more intense UV/VIS fluoresce reaction indicates that paint straight from the tube has a greater oil concentration than paint thinned iwth Winsor Newton oil painting medium.

This discrepancy prompted further investigation. Based on ultraviolet/visible fluorescence examination, it becomes apparent that paint straight out of the tube has a higher oil content than paint thinned with Winsor Newton painting medium (figs. 14-18). The visible fluorescence reaction, a golden yellow hue, suggests that the addition of painting medium composed of stand oil and turpentine disperses the paint and produces a paint film with a lower oil content. Paint straight out of the tube is more concentrated, and thus, holds more oil. This observation is consistent with the behavior of groups III and IV.

Since the same trend does not occur with the Yellow Ochre samples it is likely to be due to a much lower degree of oil saturation. As seen in tables I and VII, four out of ten samples with paint straight out of the tube (group I), have a low oil saturation, and higher loads at failure—2.103, 2.334, 2.360, and 2.726 kN/m. In future investigations, the oil saturation of the samples could be selected to be more consistent with other test groups.

Photograph

Photograph

Fig. 19-20. Recto and verso of sample # IV.21.d. Adhesive failure with corresponding support failure.

Photograph

Top: sample #II.21.d. Bottom: sample #IV.21.b. Adhesive failure corresponds with oil saturation. Wheat starch paste bond to unpainted paper remains intact.

Other observations include the frequency of ancillary damage to the specimens. From a group of thirty samples with adhesive failures, six—or 20%—also underwent support failures (19-20); fifteen—or 50%—showed some stress on the hinge (fig. 21).

5. Discussion

It is significant that 75% of all lap/shear tensile tests resulted in adhesive failure. Bonds were observed to break at loads from a low of 1.323 kN/m (7.554 lbf/in) to a high of 2.778 kN/m (15.862 lbf/in). These numbers have immediate relevance for weighty, large-scale works on paper where oil-saturated paper is to be—or has been for some time—held in shear with only a few hinges, particularly for those works being shipped upright in traveling exhibitions, for they may well be subjected to sudden downward jolts.

Results also show the unbalanced adhesive strengths of a wheat starch bond to aged and oil-saturated newsprint paper, relative to a bond made to aged, but unpainted, newsprint paper. The paste bond to oil-saturated paper fails at loads well below those of plain, unpainted paper, which, in all test pulls, remained intact.

In addition, it is clear that one type of damage can lead to another. Under lap/shear stresses, when adhesive joints fail, tears in the support may also occur. As mentioned above, six (or 20%) of the thirty test separations resulting in adhesive failures also produced tears in the newsprint support. On those six samples, the paint coverage is variegated, and the hinge/wheat starch bond covers both painted and unpainted paper. Ruptures occur where stresses are uneven. Vertical tears form in the well-adhered unpainted areas adjacent to the failed bonds of painted regions, and horizontal tears course along the bottom edge of hinges. This raises concerns for curators, conservators, and matter/framers alike, for the risk of additional injury to a work of art resulting from poorly-placed or inadequately-adhered hinges, is both undesirable and avoidable.

Other considerations in hinging such works include the cohesive strength of oil films, and the relationship of paper strength to specific pigments in oil. As noted above, all three support failures occurred in Yellow Ochre samples. It is difficult to state confidently whether these failures are the result of low oil-saturation or pigment type. However, there appears to be a correlation between pigment type and load force at failure. The average load at failure for Yellow Ochre samples, Groups I and II, is 2.154 kN/m (12.285 lbf/in), and that of the English Red Light samples, Groups III and IV, is lower—1.877 kN/m (10.715 lbf/in).

The greater tenacity of the Yellow Ochre samples, relative to English Red Light specimens, seems associated to unique physical properties of the Yellow Ochre in linseed oil. As a hydrate of iron oxide, the pigment binds well to oil, and produces a tough, good-to-excellent, paint film. English Red Light, on the other hand, actually powders and sheds, indicating a less-integrated, weaker paint film. Due to factors such as oil index, wettability, drying rate, and pH range, conservators would expect to encounter differences in adhesive bonding and paper durability, relative to pigment type, in oil paintings on paper.

6. Conclusion

Results indicate the inadequacy of wheat starch bonding to aged and oil-saturated newsprint paper when paste is prepared in a 1:4.5 ratio and thinned with distilled water. Ultimately what is adequate for plain paper does not appear acceptable for oil-saturated paper.

Accordingly, conservators may wish to inspect works of art on display or in storage to assess the condition of the hinge bonds.

Further testing is indicated. Proposals for future investigation would include examination of a range of other adhesive formulations, such as more concentrated starch-to-water mixtures, starch paste in combination with other adhesives, cellulose ethers, other synthetic adhesives such as Lascaux acrylic dispersions, vinyl acetate emulsions, and resin mixtures such as Beva 371.

8. Acknowledgments

This project began with the guidance and encouragement of paper conservator Daria Keynan and paintings conservator Carol Weingarten, and was supported by Margaret Holben Ellis, Chairman of the Conservation Center, Institute of Fine Arts, New York University. The experiment was made possible through the generosity of International Paper Company, providing advanced testing equipment at their Research Facilities at Tuxedo Park, New York, as well as technical support from paper scientist, Dennis Crawshaw, and technician, John Conklin. I wish to thank Ann Timberlake at the Institute for Paper Science and Technology; paper conservators Antoinette Owen, Rachel Danzing, and Konstanze Bachmann; paintings conservators Harriet Irgang, Rustin Levenson and Carolyn Tomkiewicz for their support, and James Frantz, Adjunct Professor of Conservation, for assistance with the Instron Machine at the Conservation Center.

Tables

Table 1. Lap/Shear Test Results—All Samples
  kN/m lbsf/in Paint Type of Failure Oil Saturation Other failure
1 1.323 7.554 Black Triangle A Full No
2 1.477 8.434 Gray Triangle A Full No
3 1.482 8.460 Black Square A Full No
4 1.527 8.718 Gray Triangle A Variegated Support
5 1.583 9.038 Black Square A Full No
6 1.618 9.241 Gray Triangle A Variegated Stress on hinge
7 1.648 9.413 Black Triangle A Variegated No
8 1.648 9.413 Black Triangle A Variegated Support
9 1.660 9.478 Black Square H Low No
10 1.668 9.525 Black Square A Full No
11 1.714 9.788 Black Square A Variegated Stress on hinge
12 1.778 10.154 Gray Triangle A Variegated Stress on hinge
13 1.782 10.176 Black Square A Variegated Support
14 1.788 10.211 Gray Triangle A Variegated No
15 1.814 10.360 Gray Triangle A Variegated Support
16 1.861 10.626 Black Square A Variegated Stress on hinge
17 1.872 1.0.688 Black Triangle A Variegated Stress on hinge
18 1.895 10.823 Black Triangle A Variegated Stress on hinge
19 1.942 11.091 Black Square A Full No
20 1.945 11.108 Black Triangle A Variegated Stress on hinge
Sigma 34.025 194.299 Black Square      

KEY: Gray Square=Ochre Straight Black Square=English Red Light Straight
Gray Triangle=Ochre Thinned Black Triangle=English Red Light Thinned

kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

Table 1 continued. Lap/Shear Test Results—All Samples
  kN/m lbsf/in Paint Type of Failure Oil Saturation Other failure
21 2.009 11.469 Black Square H Variegated Slight adhesive
22 2.103 12.007 Gray Square H Low No
23 2.134 12.188 Gray Triangle A Variegated Stress on hinge
24 2.163 12.351 Black Square A Variegated Stress on hinge
25 2.167 12.371 Black Triangle A Variegated Stress on hinge
26 2.199: 12:556 Gray Triangle A Variegated Stress on hinge
27 2.240 12.792 Black Triangle H Negligible/None No
28 2.252 12.860 Gray Square S Variegated No
29 2.262 12.914 Gray Square A Variegated Stress on hinge
30 2.274 12,985 Gray Square A Full No
31 2.289 13.071 Gray Square H Variegated No
32 2.334 13.329 Gray Square H Low . Slight adhesive
33 2.344 13.384 Black Triangle A Variegated Stress on hinge
34 2.358 13.465 Gray Square A Full Stress on hinge
35 2.360 13.475 Gray Square H Low No
36 2.375 13.561 Gray Triangle A Variegated Support
37 2.416 13.798 Gray Square A Variegated Stress on hinge
38 2.584 14.753 Black Triangle A Variegated Support
39 2.726 15.567 Gray Square S Low No
40 2.778 15.862 Gray Triangle S Negligible/None No
Sigma 80.392 459.046        
Mean 2.010 11.476        
Median 1.977 11.289        
S.D. 0.361          

KEY: Gray Square=Ochre Straight Black Square=English Red Light Straight
Gray Triangle=Ochre Thinned Black Triangle=English Red Light Thinned

KEY: kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

 

Gray
Square Ochre Straight

Table 2. Lap/Shear Test Results—Ochre Straight
  Sample No. Min lbsf/in Type of Failure Oil Saturation
22 I. 13. a 2.103 12.007 H Low
28 I. 13. b 2.252 12.860 S Variegated
29 I. 17. d 2.262 12.914 A Variegated
30 I. 9. b 2.274 12.985 A Full
31 I. 13. d 2.289 13.071 H Variegated
32 I. 17. a 2.334 13.329 H Low
34 I. 17. b 2.358 13.465 A Full
35 I. 17. a 2.360 13.475 H Low
36 1. 17. c 2.416 13.798 A Variegated
39 I. 9. e* 2.726 15.567 S Low
Sigma   23.374 133.651    
Mean   2.337 13.361    
Median   2.312 13.200    
SD   0.161      

* Note: When this sample was run, test grippers were positioned closer to the top edge of the support than in other samples. It was included to complete the sample set of ten. Had it not been included, the mean would be 2.294 Min, a difference of 0.043.

KEY: kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

Gray
Triangle Ochre Thinned

Table 3. Lap/Shear Test Results—Ochre Thinned
Sample No. kN/m lbsf/in Type of Failure Oil Saturation  
2 II. 10. d 1.477 8.434 A Full
4 II. 20. d 1.527 8.718 A Variegated
6 II. 20. b 1.618 9.241 A Variegated
12 II. 10. b 1.778 10.154 A Variegated
14 11. 10. c 1.788 10.211 A Variegated
15 II. 20. a 1.814 10.360 A Variegated
23 II. 20. c 2.134 12.188 A Variegated
26 II. 21. d 2.199 12.556 A Variegated
36 II. 21. c 2.375 13.561 A Variegated
40 II. 21. a 2.778 15.862 S Negligible/None
Sigma   19.488 111.285    
Mean   1.949 11.129    
Median   1.801 10.285    
SD   0.411      

KEY: kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

Black Square English Red Light Straight

Table 4. Lap/Shear Test Results—English Red Light Straight
Sample No. kN/m lbsf/in Type of Failure Oil Saturation  
3 III. 22. c 1.482 8.460 A Full
5 III. 22. d 1.583 9.038 A Full
9 III. 21. a 1.660 9.478 H Low
10 III. 21. d 1.668 9.525 A Full
11 III. 21. c 1.714 9.788 A Variegated
13 III. 22. a 1.782 10.176 A Variegated
16 III. 21. a 1.861 10.626 A Variegated
19 III. 22. a 1.942 11.091 A Full
21 III. 21. b 2.009 11.469 H Variegated
24 III. 22. b 2.163 12.351 A Variegated
Sigma   17.864 102.002    
Mean   1.786 10.200    
Median   1.748 9.982    
SD   0.209      

KEY: kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

Black Triangle English Red Light Thinned

Table 5. Lap/Shear Test Results—English Red Light Thinned
  Sample No. kN/m lbsf/in Type of Failure Oil Saturation
1 IV. 15. d 1.323 7.554 A Full
7 IV. 21. a 1.648 9.413 A Variegated
8 IV. 21. d 1.648 9.413 A Variegated
7 IV. 21. c 1.872 10.688 A Variegated
8 IV. 15. b 1.895 10.823 A Variegated
0 IV. 21. a 1.945 11.108 A Variegated
5 IV. 21. b 2.167 12.371 A Variegated
7 IV. 15. c 2.240 12.792 H Negligible/None
3 IV. 17. a 2.344 13.384 A Variegated
8 IV. 16. a 2.584 14.753 A Variegated
Sigma   19.666 126.997    
Mean   1.967 12.670    
Median   1.920 10.965    
SD   0.375      

KEY: kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

Table 6. Correlation between Oil Saturation and Bond Failure
Degree of Saturation
Type of Failure FULL VARIEGATED LOW NEGLIGIBLE/NONE
Adhesive 8 22 0 0
Hinge 0 2 4 1
Support 0 1 1 1
Total 8 25 5 2

Table 7. Summary of Lap-Shear Tests of Wheat Starch Paste on Oil-Saturated Paper
  MIN kN/m MAX kN/m AVERAGE kN/m MEDIAN # Failures A—H—S SD Oil Saturation F—V—L—N
ALL SAMPLES 1.323 2.778 2.010 1.977 30—7—3 .361 8—25—5—2
Gray Square YELLOW OCHRE STRAIGHT 2.103 2.726 2.354 2.314 4—4—2 .161 2—4—4—0
Gray Triangle YELLOW OCHRE THINNED 1.477 2.778 1.949 1.801 9—0—1 .411 1—8—0—1
Black Square ENGLISH RED LIGHT—STRAIGHT 1.482 2.163 1.786 1.748 8—2—0 .209 4—5—1—0
Black Triangle ENGLISH RED LIGHT - THINNED 1.323 2.584 1.967 1.920 9—1—0 .375 1—8—0 -1

KEY: Gray Square=Ochre Straight Black Square=English Red Light Straight
Gray Triangle=Ochre Thinned Black Triangle=English Red Light Thinned

kN/m=kilo-Newtons per meter lbsf/in=pounds-force/inch
Type of Failure: a=Adhesive h=Hinge s=Support

References

Berger, G. 1972. Testing adhesives for the consolidation of paintings. Studies in Conservation, (17), 173-194.

Bikerman, J.J. 1971. Adhesion and bonding. N. Bikales, ed. Wiley Interscience, New York. 38-41

Book and Paper Group, American Institute for Conservation. 1988. Matting and framing. Chapter 40. Paper Conservation Catalog. S. Bertalan and H. Phibbs, Compilers. ©Book and Paper Group, American Institute for Conservation. 40.4.2 A.5. "Number and location of hinges", 20-21; and 40.4.7. "Non-paste alternatives", 40-41

Bradley, S. 1984. Strength testing of adhesives and consolidants for conservation purposes. Adhesives and Consolidants. The International Institute for Conservation of Historic and Artistic Works. 22-25

Clapp, A.F. 1987. Curatorial care of works of art on paper. © Intermuseum Conservation Association. Nick Lyons Books. Recipe number 2. 148

Conklin, John. 1998. Personal Communication.

Cove, S. 1992. Constable's oil sketches on paperboard and millboard. Conference Papers, Manchester 92. © The Institute of Paper Conservation. 123-127

Fishman, R. 1993. Personal communication.

Gettens, R.J. and G.L. Stout. 1966. Painting materials: a short encyclopedia. © Dover Publications. Unabridged and corrected republication of 1942 publication © D. Van Nostrand. 1942. 114, 134.

Gottsegen, M. 1993. The painter's handbook. Watson-Guptill Publications. Revised edition of 1987 version. 145, 154

Hedley, G. 1984. Lining without heat or moisture. ICOM Committee for Conservation, 7th Triannual Meeting, Copenhagen. 84.2.38-44

Katz, K. 1985. Testing and comparisons of peel and lap/sheer for Lascaux 360 and Beva 371. Journal of the American Institute for Conservation of Historic and Artistic Works. (24) 60-68

Keynan, D. and C. Weingarten. 1991. Paintings on paper: collaboration between paper and paintings conservators. 1991 Paintings Specialty Group Postprints. Nineteenth Annual Meeting of the American Institute for Conservation of Historic and Artistic Works. Albuquerque, New Mexico. Compiled by Chris Stavroudis. 11-21

Kosek, J. and L.R. Green. 1992. Survey of oil painting and sketches on paper in the collection of the British Museum—an assessment of stability. Conference Papers, Manchester 92. © The Institute of Paper Conservation. 96-102

Maheux, A. 1992. Persistent dilemmas: the conservation of large works of art on paper at the National Gallery of Canada. Conference Papers, Manchester 92. © The Institute of Paper Conservation. 73-80.

Mayer, R. 1970. The artist's handbook. The Viking Press, New York. 3rd edition. Eighth printing, 1976. 140-144.

Phibbs, H. 1994. The hinging of works of art on paper. Preservation Hinging, A Picture Framing Magazine Supplement.

Swanson, J. W. 1972. An informal discussion on the cooking of starch pastes and the use of enzymes for removing degraded starch films. Second Seminar on the Scientific Approach to the Preservation of Paper Artifacts. Institute of Paper Chemistry. Transcribed from tapes by Barbara Gould, Library of Congress.

Townsend, P. and R.A. Perry. Conservation of Modern Oil Paintings on Paper. Conference Papers, Manchester 92. © The Institute of Paper Conservation. 129-32.

Volent, P. 1994. Personal communication.

Winter, J. 1984. Natural adhesives in East Asian paintings. Adhesives and Consolidants. The International Institute for Conservation of Historic and Artistic Works. 117-120

Zukor, K. 1998. Communication and persuasion: Soviet wartime posters as visual propaganda. Twenty-sixth AIC Annual Meeting, Book and Paper Session, June 1-7, 1998.

Suppliers

Blotter Paper

Archivart Products

7 Caeser Place
P.O. Box 428
Moonachie, New Jersey

Fletcher brand table mat cutter # 2100

United Picture Framing Wholesales
Long Island, New York

Hinge Tissue

Hiromi HPO7, Sekishu Hanshi-mare
New York Central Art Supply
62 Third Avenue
New York, New York 10003

Hollytex Polyester Webbing

Tefal Cook & Stir Gravy/Paste Maker
University Products
517 Main Street
P.O. Box 101
Holyoke, Massachusetts 01041-0101

Humidity Indicator Strips

Multiform Desiccant Products, Inc.
Buffalo, New York 14213

Morilla—Newsprint Paper by Canson
Utrecht Oil Paints

Winsor Newton Oil Painting Medium
Utrecht Art Supply
111 Fourth Avenue
New York, New York 10003

Zin Sho fu wheat starch

TALAS
568 Broadway
New York, New York 10012

Ann M. Baldwin
Conservator of Works of Art on Paper
115 Havemeyer Place
Greenwich, CT 06830

Publication History

Received: Fall 1998

This paper was submitted independently by the author, and was not delivered at the Book and Paper specialty group session of the AIC Annual Meeting. It has not received peer-review



Volume Contents
Publication Info
Notes from the Compiler
Guidelines for Authors
BPG Annual Series Contents Book & Paper Group

[Search all CoOL documents]


URL: http://www.cool.conservation-us.org/coolaic/sg/bpg/annual/v17/bp17-03.html
Timestamp: Wednesday, 03-Aug-2011 10:44:35 PDT
Retrieved: Monday, 24-Jul-2017 10:50:04 GMT