JAIC 1983, Volume 23, Number 1, Article 5 (pp. 47 to 62)
JAIC online
Journal of the American Institute for Conservation
JAIC 1983, Volume 23, Number 1, Article 5 (pp. 47 to 62)


Daniel Clement

ABSTRACT—An experiment is described which relates the amount of damage to certain degraded papers caused by the formation of gas-incurred delaminations during aqueous hydrogen peroxide bleaching to variations in treatment conditions. While some treatment modifications decreased the severity of the blistering, none prevented the problem entirely.


MANY PAPER CONSERVATORS USE hydrogen peroxide to bleach paper which has become discolored or stained. This bleach offers advantages over some of the other common bleaches. It decomposes into oxygen and water so that there is less chance that dangerous chemicals will be left in the paper; for example, it does not contain chlorine and so cannot leave chlorine residues. According to Burgess and Hanlan, hydrogen peroxide tends to degrade cellulose less than other common bleaches.1 While some conservators have noted problems of color reversion, Burgess reports results which indicate that paper bleached with hydrogen peroxide is less prone to color reversion than paper bleached with other common bleaches.2 Also, information presented in a TAPPI publication suggests that hydrogen peroxide improves the color stability of both bleached chemical paper pulps and bleached mechanical paper pulps.3

One problem that conservators sometimes encounter with aqueous hydrogen peroxide bleaching, however, is that blisters may form in the paper during bleaching. These blisters result when gas is produced and captured within the paper structure, causing small pockets. These gas-produced delaminations are visible because they stand in relief from the plane formed by the sheet of paper. While it is true that during actual treatments the bleaching can be terminated when the blistering starts to occur and the blisters can usually be brought back into plane as the paper dries, such mechanical disruption of the paper certainly is to be avoided.

The active bleaching species4 associated with hydrogen peroxide is the perhydroxyl anion, HOO−. The reversible equilibrium which determines the concentration of this anion is described by

Fig. .
On the other hand, the anion's partner, H3O+ is also in equilibrium with OH− from the dissociation of water:
Fig. .
Thus any increase in OH− concentration depresses the level of H3O+(eq.2) and consequently raises the concentration of the HOO− in equilibrium with it (eq.1). This implies that bleaching will occur more rapidly as the solutions of hydrogen peroxide are made more alkaline.

The irreversible decomposition of hydrogen peroxide, on the other hand, is described by

Fig. .
This decomposition is strongly catalyzed, especially under alkaline conditions, by transition metal ions, such as copper and iron, and by a host of other materials.5 Bleaching is not caused by the decomposition reaction. If the associated production of oxygen gas within the paper structure is vigorous enough, blisters may form in the paper.6 The problem of blistering does not arise with all papers, but conservators must always be aware that blistering may occur.

The present study reports results obtained in an experimental investigation into the extent of blistering during the bleaching of selected old papers when treated in a number of different hydrogen peroxide baths. The objective was to see which bleaching conditions and pretreatment procedures produced the least blistering for a given amount of bleaching.


Seven degraded expendable nineteenth and early twentieth century lithographs were chosen as subjects for the experiment. These prints were cut into small pieces and the pieces evenly distributed into several groups which would each receive a different treatment. In this way, a number of treatments could be performed on parts of the same print and the variations in the amount of blistering and bleaching could be noted. Variations in bleaching procedures included changes in pH or concentration, variation in the kind of alkaline wash pretreatment, and the addition of ethanol or stabilizers to the bleaching bath. The amount of blistering and bleaching was measured after each treatment.

The expendable artifacts used are described briefly in Table 1. These prints were chosen because in a preliminary screening they were all shown to be susceptible to the formation of blisters during bleaching.7 Old prints were chosen in preference to other possible subjects, like sheets of artificially aged new paper, because it was believed that since these were examples of actual artwork, they would more closely represent some of the combinations of fiber furnish, sizings, coatings, staining and degradation that the conservator may encounter in practice.

Table 1. Description of the Papers Used

The purpose of the investigation was not to determine what kinds of paper were more likely to blister, but to see whether the problem of blistering could be abated by varying bleaching conditions.

The treatment procedures are described below and they are also outlined by the flow chart in Figure 1.

Fig. 1. Flow chart for treatments.

  1. All prints were drycleaned with Opaline eraser and then cut into 1 ″ 3″ pieces and distributed into groups. The smallest sheet yielded 25 sections, while the largest provided 135. There were a total of nine treatment and control groups. While the larger prints provided enough pieces to be evenly distributed among all nine groups, the smaller prints yielded fewer pieces which were only used in certain of the treatment groups. Therefore, some of the experimental and control variations were performed using paper from all seven prints, while others were done using paper from only three or four prints. A total of 370 small sections of prints were used.Since it was observed in preliminary tests that the margins of papers were more prone to blistering than the rest of the sheets, note was made of the pieces which were taken from the edges of the prints and they were evenly distributed among the groups. These were later analyzed both as part of the entire group and separately.After the sheets were cut up and distributed into groups, each group was washed in tap water for hour and allowed to air dry.
  2. All groups were given an alkaline wash for hour at 25C and allowed to air dry. (Reflectance measurements mentioned below in section F were taken here.) With eight of the nine groups this involved immersion in a bath of water brought to pH 10 with ammonium hydroxide, while the remaining group was bathed in a 50% saturated solution of calcium hydroxide. Margaret Hey has suggested that such a pretreatment may prevent damage to paper caused by hydrogen peroxide bleaching.8 The calcium hydroxide solution had a pH of about 12.9 Deionized water was used for all alkaline and bleaching baths.10
  3. Six of the nine groups were then each bleached in a particular hydrogen peroxide bleaching solution for hour at 25C All bleaching was done in non-metallic trays. For each treatment variation, all small pieces from a particular print were treated in the same tray, with 100 ml of solution being added for each piece. Fisher 30% hydrogen peroxide, stabilized with sodium stannate by the manufacturer, was the source of peroxide.The formulations of the various bleaching solutions are listed below.2% hydrogen peroxide brought to pH 10 with ammonium hydroxide. This was the standard bleaching solution to which the others were compared. The others each varied from this formulation in a particular way, for example, in pH or concentration.112% hydrogen peroxide in water brought to pH 9 (instead of 10) with ammonium hydroxide.1% hydrogen peroxide (instead of 2%) brought to pH 10 with ammonium hydroxide.2% hydrogen peroxide in water. A pH of 10 was maintained with sodium hydroxide, sodium carbonate and acetic acid. Magnesium sulfate and sodium silicate were added as stabilizers against decomposition.This formulation was suggested by Helen Burgess.12 While five different chemicals were added here, the pH and concentration remained the same as with the standard and it was decided to consider the experimental variable simply as the addition of stabilizers.A 33% volume equivalent of ethanol was added to a 2% solution of hydrogen peroxide in water.13 The pH was raised to 10 with ammonium hydroxide.
  4. The three remaining groups were alkaline wash controls. Since the total brightening resulting from an alkaline bleaching bath is presumably the combined effect of both bleaching and alkaline washing, alkaline baths were used as controls so that a clearer picture could be obtained of the brightening due to bleaching alone.The formulations of the alkaline wash controls are listed below. Each corresponds to one of the experimental bleaching solutions in pH and/or in the addition of ethanol.pH 10 ammonium hydroxide bath.pH 9 ammonium hydroxide bath.A 33% volume equivalent of ethanol is added to water and the pH raised to 10 with ammonium hyroxide.
  5. After bleaching, all papers were washed for five minutes in tap water and the blistering which occurred on each small piece was recorded with black and white raking light photographs taken on 35 mm film. The photography was performed while the paper was wet. The amount of blistering was later to be judged from these photographs. Figure 2 is a diagram of the box which was constructed in order to allow all of the photographs to be taken under precisely the same conditions of illumination and magnification, with the effects of ambient light eliminated. The blistering recorded on film was then compared to the kinds of blistering depicted on Figure 3A. This figure indicates how the amount of damage was quantified on a scale of 0 to 4. A rating of 0 meant that no blistering occurred at all, and 1 meant that very little blistering occurred. 2 was an intermediate rating described by either of the sketches in the third row. Severe blistering was given a rating of 3. This blistering might be evenly distributed over the surface, but it was also observed that the blistering often occurred preferentially over a limited part of the surface. A rating of 4 was used to describe the greatest amount of damage where the entire surface had blistered, or had blistered and delaminated. Although the categories may appear somewhat arbitrary, they seemed reasonable after observing the ways in which paper tended to be damaged by gas formation within the paper structure. One of the actual photographs taken using the setup described previously is reproduced for comparison in Figure 3B. 2 This box was constructed to make the raking light photographs used to record the blistering on the small bleached pieces of paper. The box was made of wood and painted flat black. 3A Degrees of damage from blistering. 3B Photograph of blistered paper taken using the photography box described in Fig. 2.In order to rate the blister damage to paper, contact prints were made of the 35 mm negatives and the photographic paper was cut ot the size of the individual frames. Three judges, all second-year Cooperstown paper conservation students, then grouped the photographs into the categories using the diagram as a guide. For purposes of analysis, the average of the three ratings, one from each judge, was used. Only with five photographs out of approximately 270 did the ratings of the three judges differ from one another by more than one unit. In these five cases, the judges were given the actual pieces of paper to evaluate, and the two-unit discrepancy was eliminated. It was considered acceptable, for example, if two judges gave a rating of 3 while the other gave a rating of 2. (The damage recorded by the photo in Figure 3B was given a rating of 3 by all judges.)
  6. After the five minute water rinse and photography, the papers were rinsed again for 55 minutes in tap water and allowed to air dry. The brightening of the paper was measured by taking total reflectance measurements of the surface of the dry paper at a wavelength of 416 nm both before the final bleaching step and again after the final tap water wash.14 The change in brightness was also measured for the alkaline wash controls, but no photographs were taken because no blistering occurred.


Data obtained in the experiment are summarized in the form of bar graphs (Figures 4–8). Each figure shows, for a number of different papers, the amount of bleaching and blistering that took place under the standard procedure as opposed to that which occurred under a modified procedure. The results obtained with each paper under the standard conditions are always represented by the bar (marked A) at the left in each pair of bars. The bar at the right (marked B) in each pair is the experimental variation which is being compared to the standard pH 10—2% procedure.

Fig. 4A. Damage from blistering as a function of pretreatment. A = NH4OH (standard), B = Ca(OH)2.

Fig. 4B. Change in percentage reflectance as a function of pretreatment. A = NH4OH (standard), B = Ca(OH)2.

Fig. 5A. Change in blistering as a function of pH. A = pH 10 (standard), B = pH 9.

Fig. 5B. Change in percent reflectance as a function of pH. A = pH 10 (standard), B = pH 9.

Fig. 6A. Damage from blistering as a function of concentration. A = 2% H2O2 (standard), B = 1% H2O2.

Fig. 6B. Change in percent reflectance as a function of concentration. A = 2% H2O2 (standard), B = 1% H2O2.

Fig. 7A. Damage from blistering as a function of the use of stabilizers. A = unstabilized (standard), B = stabilized.

Fig. 7B. Change in percent reflectance as a function of the use of stabilizers. A = unstabilized (standard), B = stabilized.

Fig. 8A. Damage from blistering as a function of the addition of ethanol, A = H2O2 in water (standard), B = H2O2 in water and ethanol.

Fig. 8B. Change in percent reflectance as a function of the addition of ethanol, A = H2O2 in water (standard), B = H2O2 in water and ethanol.

3.1 A. Calcium Hydroxide vs. Ammonia Pre-treatment

Figure 4A shows the amount of blistering that occured during bleaching and Figure 4B shows the increase in brightness. The amount of blistering was reduced in most, but not all cases. The amount of bleaching was slightly reduced in all cases, possibly because more discoloration was removed in pre-treatment due to the higher pH of the calcium hydroxide bath.

3.2 B. pH 9 vs. pH 10

Figures 5A and 5B show that lowering the pH reduced both the amount of blistering and the amount of bleaching. This is consistent with the fact that the concentration of the active bleaching species and the rate of hydrogen peroxide decomposition are both less at lower pH.

It is notable that the various papers reacted quite differently to the same variation in bleaching conditions. Blistering was reduced significantly with papers #3 and #4 without equally great reductions in bleaching. Little reduction in blistering occurred with paper #5. Although blistering was reduced significantly with paper #8, the reduction in bleaching was considered unacceptably great. The papers were obviously quite different in their reactions to changes in bleaching conditions, but the differences would probably not have been predictable by the paper conservator prior to bleaching.

3.3 C. 1% vs. 2% Concentration

The results given in Figures 6A and 6b show how three different papers reacted when the hydrogen peroxide concentration was lowered to 1% from the standard of 2%. With papers #3 and #5 blistering was substantially reduced, but not with paper #7. The amount of bleaching, however, was reduced in all cases. It is apparent that the amount of oxygen produced can be very substantial even with lower concentrations of hydrogen peroxide. The decomposition of even a small amount of hydrogen peroxide can produce a surprisingly large volume of oxygen. (The decomposition of 1 ml of a 1% solution can produce approximately 5 ml of gas.)

3.4 D. Stabilized vs. Unstabilized

Figures 7a and 7B show that with paper #3 the amount of blistering was greatly reduced while the amount of bleaching remained about the same. Unfortunately, such encouraging results were not produced with the other four papers. These results seem consistent with a statement by Schumb, Satterfield and Wentworth: “… it appears reasonably certain that in the presence of more than minute proportions of catalytically active contaminants it is impossible to restrain the decomposition of hydrogen peroxide by the addition of stabilizers …”15 It seems likely that the amount of impurities present in papers like these may be beyond the limit within which stabilizers are effective.

It is worth noting that each time this bleach bath was prepared a certain amount of the white chemicals remained undissolved. This material clouded the solution and partially obstructed the view of the artwork. This phenomenon might restrict the usefulness of the bath for conservation procedures since the paper must be observed carefully during the bleaching process so that if blistering starts to occur the bleaching can be terminated. Also, the insoluble material might become trapped in the interstices of the paper.

3.5 E. Addition of Ethanol

In the final variation of treatment, ethanol was added to the bleaching bath. It was hoped that in the presence of ethanol the paper might remain stronger because of the smaller percentage of water. Ethanol reduces the surface tension, thus increasing the possibility that the gas produced could escape from the paper structure without damaging the paper. Oxygen is also much more soluble in ethanol than in water, a fact which might reduce the extent of bubble formation.

Figure 8A shows that the addition of ethanol reduced the amount of blistering in most but not all cases. Again, the various papers reacted to the same change in procedure quite differently. The addition of ethanol practically eliminated the blistering with paper #3, but did not substantially reduce the blistering with papers #4 and #8. Blistering was quite reduced also with papers #5, #6 and #7. The amount of bleaching was reduced in all cases.

3.6 F. Marginal Pieces

As mentioned earlier, note was made of which pieces came from the margins of the prints and these pieces were distributed among the different bleaching baths. Indeed, it was found that the amount of blistering on these pieces was greater than average by about half a point in the 0 to 4 scale used to describe damage.

3.7 G. Alkaline Wash Controls

The amount of brightening due to alkalinity alone was always very small compared to the effect due to bleaching.16

3.8 H. pH Changes During Bleaching

The alkalinity of the bleaching baths never changed by more than 0.15 pH unit during treatment, and generally the changes were much less than this. These fluctuations were judged not to be significant.


Several conclusions about the blistering of paper from aqueous hydrogen peroxide bleaching can be drawn:

  1. The more degraded paper found in the margins of these prints was more prone to blistering than paper taken from the rest of the sheets. This may have occurred because the marginal paper contained more catalysts for decomposition, perhaps as a result of handling, or possibly since this paper was weaker it was more susceptible to mechanical deformation.17 Unfortunately, it is generally the more degraded papers which are more disfigured by staining and discoloration and which could be more visually improved by bleaching.
  2. Peforming a calcium hydroxide pretreatment wash or adding stabilizers to the bleaching bath may reduce the blistering somewhat, but this did not consistently cause significant reductions in blistering with the papers which were used in this study. Additionally, undissolved chemicals in the stabilized bath may interfere with the inspection of artwork during bleaching.
  3. Using a less vigorous bleaching solution, as would be obtained by lowering the pH, seems to lessen the possibility of severe damage. Although less bleaching occurred during a fixed period of time, it would be interesting to see whether a comparable amount of bleaching might not occur if the bleaching time were extended.
  4. Lowering the concentration of hydrogen peroxide seems also to reduce the amount of blistering, although perhaps not as consistently as does lowering the pH.
  5. The addition of ethanol reduced the blistering quite well with most, but not all, of the papers studied. The amount of bleaching was also reduced. The results were fairly similar to those produced by lowering the pH form 10 to 9.
  6. Caution is advised when drawing conclusions about hydrogen peroxide bleaching based on behavior observed with one particular paper. Quantitatively, the responses of the seven different papers in this experiment to the same treatment modifications were surprisingly varied.


THE VARIATIONS IN bleach bath composition and paper pretreatment considered in this experiment reduced the tendency of the paper to blister in certain cases. While the use of a less alkaline solution and the addition of ethanol showed promise, no test procedure eliminated the problem of blistering. No “miracle cure” was discovered for this problem, and the conservator should always be aware that blistering may occur even if precautions are taken. Certainly, further research is warranted.


THIS RESEARCH WAS performed while the author was a student at the Cooperstown Graduate Program. The author would like to thank the faculty, especially Dr. Christopher Tahk and Ms. Cathleen Baker, and his fellow students for their help and guidance.


Burgess, H. and Hanlan, J., “The Degradation of Cellulose by Bleaching,” Journal of the International Institute for Conservation—Canadian Group, Vol. 4, No. 2 (1979): pp. 15–22.

Burgess, H., “The Colour Reversion of Paper After Bleaching,” preprints of the Institute of Paper Conservation Cambridge Conference, 1980: pp. 171–183.

Rapson, W. H., ed., The Bleaching of Pulp, Technical Association of the Pulp and Paper Industry, New York (1963): p. 301.

The dissociation constant of hydrogen peroxide in an aqueous solution at 25C is 2.4 10−12. Discussion of the perhydroxyl anion as the active bleaching species is presented in Rapson, The Bleaching of Pulp (op. cit.): p. 180.

Schumb, W.; Satterfield, C.; Wentworth, R., Hydrogen Peroxide, Rheinhold Publishing Corp., New York (1955): Chapter 8.

When the gas which formed within the blisters was exposed to a small glowing splint of wood, the wood immediately underwent very rapid combustion, indicating the presence of oxygen.

To see whether a particular print was prone to blistering during hydrogen peroxide bleaching, a small section of the sheet was placed in a 2% solution of peroxide in water (pH = 10). The papers which blistered were chosen for the experiment.

The recommendation for calcium hydroxide (or magnesium bicarbonate) pretreatment is found in: Hey, M., “Paper Bleaching: Its Simple Chemistry and Working Procedures,” The Paper Conservator, Vol. 2 (1977) P. 20. The use of a 50% saturated solution of calcium hydroxide is suggested in:Hey, M., “The Washing and Aqueous Deacidification of Paper,” The Paper Conservator, Vol. 4 (1979): p. 70.

In the author's opinion, this solution is too alkaline for routine treatment of works of art.

The water was deionized with a Barnstead Combination Cartridge. All chemicals were Fisher Scientific Reagent Grade.

A pH of 10 was chosen because it allowed comparison with Ms. Burgess' stabilized bleaching bath, where a pH of 10 is specified. (See #1 of #2 above.)

Burgess, H., “The Colour Reversion of Paper After Bleaching,” (op. cit.): p. 172.

In this case, the resulting concentration of hydrogen peroxide in water and ethanol was less than 2%. It was decided for the purposes of the experiment not to consider the alcohol as part of the solvent since in preliminary tests it seemed to be a relatively inert component in the peroxide-water-alcohol-ammonia bleaching bath.

Diffuse reflectance measurements were taken using a Pye Unicam SP8-100 spectrophotometer equipted with a diffuse reflectance accessory (an integrating sphere type conforming to C.I.E. recommendations) and a barium sulfate white reflectance standard. The reflectance measurements were the same whether the supporting layer behind each sample was white or black. The wavelength of 416 nm was chosen because it allowed comparison with work done by Ms. Burgess. Also, when some full spectrum readings of paper and after bleaching were studied, it was observed that the greatest changes in reflectance did occur in the blue part of the spectrum near 416 nm.

Schumb, W.; Satterfield, C.; Wentworth, R., Hydrogen Peroxide (op. cit.): p. 535.

In the interest of brevity, and because the amount of brightening due to alkalinity was small, the data obtained with these controls are not considered in this paper. Complete data may be obtained from the author.

P.Holladay and R.Solari mention in The Bleaching of Pulp (op. cit.: p. 191) that catalase is formed by the metabolism of bacteria which are found in some paper pulps. The production of this enzyme is of concern to papermakers because it accelerates hydrogen peroxide decomposition and therefore reduces bleaching efficiency. It would be interesting to see whether this enzyme plays a role in the more vigorous decomposition of hydrogen peroxide noted in degraded paper.

Section Index

Copyright 1983 American Institute of Historic and Artistic Works