2 EXPERIMENTAL PROCEDURE

2.1 Materials

THREE DIFFERENT TYPES of paper were selected, based on an estimate of permanence. Permanence can be defined for the purposes of this study as the expected ability of the paper to retain certain properties (i.e., those properties tested, as described below) during the various experimental treatments. The papers selected for testing were:

1. Whatman Chromatography #1,3 basis weight 87g/m2, thickness 0.17mm, supplied in a roll �″�300′. This paper was expected to have moderate permanence. It is made from pure cellulose without additives, and has an ash content of 5μg/cm2, consisting largely of sodium sulfate, sodium carbonate, and silica.4 A number of rolls were tested for folding endurance and pH, both unaged and after two weeks of dry oven accelerated aging at 100�C. The paper was found to lack uniform aging characteristics from roll to roll. For this reason, the Whatman sample population was limited to a single roll. Samples were cut to measure �″�6″.
2. Arches Cover Infinity,5 a paper designed for permanence (hence the name), was expected to have excellent permanence. This paper is made of 100% rag fiber, is hard-sized with a synthetic size, and buffered with 3% calcium carbonate.6 Strips measuring�″�6″ were cut, avoiding the deckle, from two sheets, with the 6″ dimension parallel to the machine direction of the paper.7
3. Aquabee Newsprint Rough 887-R,8 was chosen to give some overlap with the sample population in the Tang and Jones study. It is an acid paper of groundwood and chemical woodpulp, with alum-rosin size,9 and was expected to be the least permanent of the papers tested.

Three waters were selected to give a range of purity:

1. New York City tap water was obtained in the Conservation Center of the Institute of Fine Arts;
2. deionized water was prepared by passing tap water through a Barnstead Bantam deionizer, equipped with a new High Capacity (multi-bed) filter;10
3. distilled water was prepared from tap water using a single column glass distillation apparatus and standard distillation procedure, and was stored in glass flasks with ground glass stoppers.

All three types of water were tested for free and total chlorine content and for hardness as calcium carbonate.11 Both the deionized and distilled waters showed no perceptible amounts of chlorine or calcium carbonate, while the tap water did contain significant amounts of both contaminants (0.1mg/1 chlorine, and 102.6mg/1 calcium carbonate).

Each of the waters was tested for pH (see Table I). Both the tap and deionized waters had values that varied from day to day, but the pH of the individual samples of the water remained stable for periods of two to three hours. The pH of the distilled water, however, became increasingly acidic on exposure to the atmosphere (to a minimum of pH 4.0), resulting from reabsorption of gaseous carbon dioxide that had been expelled by the distillation procedure. The pH of the distilled water was stabilized by bubbling nitrogen gas through it for the duration of the washing treatments.12

Table I. pH of Wash Waters

2.2 Methods of Treatment

Ten samples of each paper were subjected to each of the various combinations of washing and aging procedures. To minimize differences in handling, the ten samples for each treatment were sewn to a strip of Whatman Chromatography #1 paper, leaving a gap of approximately 3/8″ between the samples to allow free circulation of the water on all surfaces.13

Nylon net was used as a washing support. Preliminary tests of the effect of the net when used as a washing support for Whatman paper samples revealed an increase in pH and in the folding endurance of the paper samples after washing. The net itself was considerably more flexible after the treatment. This strengthening effect on the paper was attributed to the action of a water-soluble sizing on the net. The infrared spectrum14 of a sample of the sizing showed a strong resemblance to polyvinyl alcohol, which has been found to cause a large increase in the folding endurance of paper.15 The net was considered suitable for use as a washing support when further preliminary washing did not result in a change in its flexibility or in the pH of the wash water. This was achieved by bathing the net for one half-hour in hot tap water.

Folders of the net were made by folding in half sheets approximately 14″�14″, and sewing with white cotton thread along the fold. Before each treatment, the prepared net folders were soaked for about 15 minutes in the type of water to be used in the treatment, then drained on clean blotters, and allowed to air dry. The paper samples were placed in the net folders with the support strip of Whatman paper adjacent to the folded seam. This system provided consistency in handling, as all ten samples could be picked up at the same time, being held through the two layers of net by the corners of the support strip.

The washing trays were constructed of Mylar, a chemically inert film of polyethylene terephthalate.16 Just before each treatment, the washing trays were rinsed with tap water, then with deionized water for the deionized and distilled water treatments, and thoroughly drained. Each washing tray held two liters of water, maintained at 21�-24�C for the tap water and deionized water, and at 18�C for the distilled water. (Temperature was monitored with recording hydrothermographs. Nonuniformity in the treatment temperature resulted from changes in outside temperature for which the heating system could not adequately compensate.)

The washing treatment consisted of submersion of each folder of samples in three successive baths. “Normal” wash time was one half-hour (ten minutes per bath); “overnight” wash time was 16� hours (ten minutes each in the first two baths, and sixteen hours ten minutes in the last bath).

Repeated immersion in still water baths was chosen over a single immersion in running water primarily for the sake of convenience. The solvation of material during washing has been shown to be dependent more on its diffusion into the wash water than on the amount of flow of the water, a bath of still water being adequate for washing if the water-paper system is allowed to reach equilibrium.17 We felt that if equilibrium were not reached with two liters of water and ten minutes wash time per bath, then a) these conditions at least approximate common practice and are thus reasonable, and b) this would show up in a difference in effect between the “normal” and “overnight” wash treatments. After washing, the samples were removed from the baths, held aloft briefly to allow excess water to drain, and laid on clean, acid-free blotters to air dry overnight before aging and testing.

Samples were aged in a Precision Scientific Company Model 17 circulating dry oven at 100�3�C. The samples were hung vertically from a rack, very loosely packed to ensure free circulation of air around all surfaces of each sample.18 To separate the effects of washing from effects of dry oven aging, unwashed control samples of each type of paper were given aging treatments identical to those of the experimental samples. Whatman paper samples were subjected to a range of aging times: 0 or 1 week before washing; and 0, 1, or 2 weeks after washing; referred to as pre- and post-aging respectively. Arches paper samples were post-aged only, for 0, 1, and 2 weeks. Newsprint samples were post-aged only, for 0, 3, and 7 days.

Dry oven aging was chosen as a variable to determine whether effects of washing not immediately apparent would be manifested with the increase in temperature. It should be noted that this was not intended to approximate natural aging.19 The reactions that occur during dry oven aging, affecting folding endurance, brightness, and pH, proved to be sensitive to water washing, and as such are good indicators of differences in the effects of the various waters.

2.3 Methods of Testing

Folding endurance was measured on a Tinius-Olsen Model No. 2 machine,20 with a dead weight of 250g on a #2 spring for Newsprint, 500g on a #2 spring for Whatman, and 750g on a #3 spring for Arches samples. Results are reported as the average of ten measurements with standard deviation.21

Tensile strength measurements were made on an Instron Table Model TM-M 1101 Universal Tester,22 with crosshead speed of 0.5cm/min, and load scale of 10kg for Whatman and Newsprint, and 20kg for Arches samples. Determination of tensile strength was made on portions of the samples not previously subjected to tension during folding endurance testing. Results are reported as the average of three to ten measurements with standard deviation.

Color was subjectively measured by comparing samples to each other and to Munsell Color Company glossy surface sample chips23 under illumination from the Macbeth Spectral Light Machine at the setting which revealed the most differences in color: “daylight” setting for Whatman and Arches samples, and “minus red” for Newsprint samples.

Reflectance was measured as percent-transmittance at 457nm on a Bausch & Lomb Spectronic 20 Colorimeter-Spectrophotometer.24 For each type of paper, 100% transmittance was defined by the reflectance of the unaged, unwashed control samples. Results are reported as the average of three measurements with standard deviation.

Surface pH was measured on a Beckman Expandomatic SS-2 pH meter25 with an Ingold silver chloride reference glass electrode.26 Readings were taken on samples laid out on a sheet of Mylar over a sponge pad, rinsed with distilled water and dried between readings. The electrode was also rinsed with distilled water and dried between readings. A value for surface pH was obtained as follows: one drop of nitrogenstabilized distilled water was placed on the sample. Whatman paper, which wet easily, was tested by placing the electrode on the wetted surface and immediately commencing the reading. With Arches and Newsprint samples, the drop of distilled water sat on the surface and was moved in a small arc with the electrode to enhance wetting before the reading was taken. A value for each reading was obtained when the pH meter needle stabilized, usually after about 10 seconds.27 Results are reported as the average of three to ten measurements with standard deviation.