![[CoOL]](/icons/sm_cool.gif){kind=icon}
This section of the report summarizes technical data which should be useful criteria by which to judge the performance of the Bookkeeper process and its prospects to meet the needs of the Library of Congress. Because of the very limited set of sample materials, it is premature to consider the results of these evaluations to be typical or to represent the fullest capability of the technology. Instead, these examinations should be taken as a screening, to highlight successful performance or failures for these particular materials and to target important process performance tests that should be considered in future, comprehensive evaluations.
This section of the report addresses four key issues:
According to the vendor, the Bookkeeper process is based on the application of fine magnesium oxide particles into the paper sheet. These particles are applied by immersion of the book into a suspension of the magnesium oxide in an inert fluid carrier, chosen to minimize possible interactions with the book materials. Mechanical agitation of the text allows the slurry to penetrate the book structure and deposit MgO on the surfaces of and within the pages. Removal of the book from the reactor followed by evaporation of the carrier fluid completes the treatment, and only then is the neutralization of the paper acidity thought to begin. The magnesium oxide particles rapidly absorb water and form magnesium hydroxide, which is claimed to be the active alkaline agent in this treatment. This alkaline salt being sparingly soluble in water, the paper acidity must migrate to the solid particles and react to form the neutralization products.
The Bookkeeper process is exceptional in two respects. It is by its nature "clean," introducing only very pure magnesium oxide into the paper and exposing the book to relatively unreactive fluids and surfactants. It also does not subject the books to "preconditioning" treatments which may risk incidental changes to the book materials or structure. The process as it exists today (in August 1993) also has the advantage of using materials which are not likely to become subject to environmental regulation. A second key feature of the process is the inherent simplicity of the treatment. No chemical reactions are necessary to produce the desired combination of ingredients in the paper, and consequently the treatment is unlikely to create an unexpected outcome due to poor process control. This treatment puts magnesium oxide into paper, and the only apparent variability in the treatment application is the quantity and distribution of the magnesium oxide deposited in the book.
At the same time that the process enjoys advantages of simplicity and safety, it also has two intrinsic problems. The first is the application of the particles to the paper by transfer from a fluid suspension. This procedure relies both on the fluid transport through the book structure and on the sticking efficiency of the particles on a sheet to determine the amount, distribution, and rate of particle deposition. Both of these factors depend on characteristics of the individual books and remain poorly understood; more research is needed before the treatment process will be able to guarantee that a particular specification can be met for every treated volume. The operation as it exists (in August 1993) has almost no control over the outcome of the process (i.e., the deposition in each book), but only over the external variables such as the suspension formulation and processing conditions. While this may be typical for "mass" treatment services in general, only after more experience with the current technology will the vendor be able to identify useful bench marks, such as difficult-to-treat materials, which should be used to develop conservative processing parameters.
The second inherent problem facing this process (although not necessarily peculiar to this process) is the fact that the chemistries leading to acid neutralization are poorly understood and occur over some indefinite time period following the treatment itself. The formation of active alkaline agents and their reaction with the resident paper acidity are left to the forces of thermodynamics to carry out; there is no control over, and very little current evidence to indicate, when the proximity of alkaline particles and diffusing acidity will achieve the desired neutralization. The vendor's claims that such reactions will occur are plausible, but at issue is whether, for a variety of papers in ambient environments, acids will be neutralized faster than they will degrade the papers. The uncertainty surrounding this issue is no different than the uncertainty, that some paper chemists raise, that alkaline reserves will actually provide substantial benefits. However, unlike some processes which promise some neutralization chemistry during the treatment, the Bookkeeper process is utterly dependent on these slow, uncontrolled chemistries.
There are two separate chemistries which must be established to have a needed understanding of the process capabilities and risks. The first is the formation of the various alkaline salts which will react with the paper acidity; the second is the neutralization reaction itself. The first type of chemistry is relatively well established, and the oxide will probably react with ambient water and carbon dioxide to form the basic hydroxides and carbonates which will be the active alkaline agents. The eventual formation of the carbonate salts may be likely. It should be noted that the higher solubility of these carbonate salts may greatly affect the rate and efficiency of the neutralization chemistry. To date it is unknown how quickly these transformations will take place. Evidence is needed to support the vendor's claims of what magnesium species are finally formed in the paper.
Similarly, evidence is so far lacking to support PTI's claims that acid neutralization is a rapid process in treated paper. In ordinary paper at ambient conditions, the transport of acidity to the insoluble magnesium salts, or the migration of solution phase acidity and alkalinity, will control the rate of neutralization. Conventional "proof" of acid neutralization, such as pH or alkaline reserve measurements, require adding copious amounts of water to the sheets, and this will greatly facilitate the progress of the acid-base chemistries. In fact, these measures can only serve to demonstrate that the ingredients necessary to achieve neutralization are present in the paper. To date, there is no direct evidence to demonstrate the rate of neutralization in a treated sheet at normal conditions. Furthermore, for a process which can probably only deposit alkaline agents on the surfaces of heavily sized or calendered paper fibers, the effectiveness of this treatment in rapidly neutralizing acidity present in the interior of the paper fibers is unknown.
Without the detailed information about the species formed in treated papers, it is not possible to make definite claims about other issues relevant to this process. For example, the risk of damage to the paper or other book materials from excessively high pH (above pH = 9 or so) is difficult to assess without knowledge of the salts present. The hydroxide, which is probably formed rapidly (at least on the surfaces of the oxide particles), has a relatively high pH in a saturated solution (10.4) which may be cause for concern. Conversely, if the carbonate species are the major alkaline agents eventually formed, the pH of their solutions will tend to be lower and the risk of alkaline reactions less.
Another concern for this process which cannot be resolved with current information is the nature and fate of the reaction products formed from the neutralization. Since neutralization only takes place following the completion of the treatment, there is no opportunity to remove the magnesium salt neutralization products from the paper, and the consequences of their incorporation into the papers must be addressed. Unfortunately, very little is known at present about such salt additions on paper properties or aging. One would expect that the reactions of hydroxide and carbonate salts have an intrinsic advantage in the release of some of the neutralization products (water and carbon dioxide) which will tend to encourage the reactions to proceed to completion rather than stop at some equilibrium mixture which remains acidic. The use of magnesium salts for the alkaline agents will create neutralization products such as magnesium sulfate or nitrate which tend to be soluble (so they may not inactivate the salt by forming a barrier crust on the particle surfaces), pH neutral, and have no known adverse reactions on paper strength properties.
Until more materials have been tested, the possibility of any adverse interactions with inks, sizes, glues, or other book materials can not be eliminated. The fluid carrier used has a low solvent power for most of the commonly encountered book materials, and it may be the lowest risk of the available alternatives. However, the solvent power of the surfactant/carrier mixture is not documented, and as a result there could be a risk for solubility of some book materials. Similarly, while the chemical stability of the fluid carrier is great enough that the risk of interactions of the residual fluid in the book may be low, the immediate consequences of the surfactant absorption into the book materials and its long-term aging behavior are unknown.
The focus of this evaluation is to assess whether the current process parameters are appropriate to achieve the immediate objective of the treatment: to deposit adequate alkalinity uniformly across the pages of all the books included in the study. All the books were surveyed using a pH indicator (0.04% chlorophenol red) to test for the presence of enough alkalinity to produce an alkaline reaction of the indicator. For the LC blue test books, individual pages were also examined by scanning electron microscopy, to estimate the concentration of magnesium salt particles across the page and within the paper web, and by ICP/mass spectrometry, to quantitatively measure the local salt concentration deposited across the sheets. Cold extraction pH and alkaline reserve measurements were also done to compare the page-average particle loading to the Library specification. These measurements were also performed on those books in the test batch which seemed to contain an inadequate alkaline deposit as measured by the pH indicator, so that it could be determined whether the particle loading was insufficient, or whether the loading was typical but overwhelmed by abnormally high paper acidity.
Because of the large number of materials to be tested, application of pH indicator solution was chosen to provide a rapid qualitative screening for "alkalinity" of the book pages. This test serves two functions: 1) where the color reaction differs after treatment, as a measure of the completeness of the deposition over the sheet area; and 2) where the solution reflects alkalinity, as an indication that the local area now contains enough MgO to render the pH alkaline. An acid indication on a treated page can result from either no particle deposition or a deposit so small as to leave the pH acid after the neutralization reaction has progressed as far as possible. This indicator test alone cannot determine whether there are any particles deposited on apparently incompletely treated areas. For some materials, particularly newsprint or coated papers, the indicator color was often not easily determinable because of rapid color shifts between acidic and alkaline response as the indicator solution dried. For papers containing an alkaline coating over an acidic core, such color reactions may indicate the slow penetration of the solution and eventual neutralization reaction. For treated sheets, it is not known if this occurrence also indicates zones where the local pH varies, or if the process chemistry itself is being altered as the indicator solution carries alkaline salt into regions of poorly accessible acidity. In any event, results are reported only for those tests where the color of the indicator was definite and unchanging.
The pH indicator test was performed on all the paper types included in the LC blue test books. In addition, LC blue test books which had been humidoven aged for 1 and 2 weeks prior to treatment were also surveyed. Of the six paper types included in these books, only the Clear Spring offset, alum/rosin-sized, and newsprint pages tested definitely acidic prior to treatment. Single pages of the Clear Spring offset paper and the alum/rosin-sized paper from all three treated books (oven-aged for 0 weeks, 1 week, and 2 weeks prior to treatment) tested alkaline across the entire page, except for the gutters of the Clear Spring offset in the unaged treated book (patches comprising about half the total gutter area tested acidic); the alum/rosin-sized page from the book oven-aged 1 week (40% acidic in patches in the gutter); and the alum/rosin-sized page from the book oven-aged 2 weeks (about 5% acidic in the gutter). With the exception of these gutter areas, the treatment application over the rest of the pages was apparently complete. For the newsprint pages, the color change of the indicator was not the instant conversion to purple but rather a rapid color shift between acidic and alkaline response, which eventually indicated alkalinity. Since this color shift was not apparent in the untreated newsprint, the treatment is judged to have affected the entire page surface. However, because of the complex color indication, it could not be determined whether this resulted in an alkaline pH over the page. Based on its performance on these LC blue test books, the treatment seems to be reaching most of the page surfaces, but there seems to be some difficulty in controlling the treatment in the gutters. There does not seem to be a trend in the locations in the books where deposition in the gutters is poor. Oven-aged, more acidic materials also seemed to be treated adequately (with the exception of some gutters).
The deposition of magnesium oxide particles on the surfaces and interior of the papers in the LC blue test books was examined directly in a scanning electron microscope. The surfaces of all of the papers tested (Clear Spring Offset, alum/rosin-sized, newsprint, alkaline-sized, Sterling litho gloss, and supercalendered) showed evidence of the deposit of fine particles distributed rather uniformly on the fiber surfaces and in the interstices between the fibers. The images of the cross-sections of the sheets, however, showed no obvious signs of particle deposits in the interstices within the sheets. It could not be determined from the photomicrographs how deeply the treatment was able to penetrate into the paper web. The x-ray fluorescence spectra, which provide semi-quantitative elemental analyses for the areas pictured in the SEM images, are generally consistent with this visual observation. All the surfaces of the treated papers show evidence of the presence of magnesium, but none of the analyzed areas of the interior of the papers, with the possible exception of the newsprint interior, indicate the presence of magnesium. (It should be noted, however, that this x-ray fluorescence analysis is not well suited to detect interstitial materials in cross-sectioned samples, for it will tend to probe the fiber cross-sections which make up most of the imaged surfaces rather than the interstices where the magnesium oxide would be present. A better approach for future analyses might be to analyze the fiber surfaces revealed in the interior of split sheets rather than cross-sections of papers.)
The quantitative measure of the distribution of magnesium salts across the pages of these test papers was provided by the ICP/mass spectroscopic analysis. These data for the six paper types (Clear Spring offset, alum/rosin-sized, newsprint, alkaline-sized, Sterling litho gloss, and supercalendered) illustrate the ability of the process technology to apply the magnesium oxide uniformly and thoroughly across the pages. The results are consistent with the findings from the pH indicator tests. Most of the papers contain an overall uniform deposit of magnesium oxide across the entire sheet, with the exception of the gutter areas, which generally have less (about 30-50% of the average loading). The exception was the alkaline-sized sheet, which had slightly smaller loadings on both the gutter area and the "top edge" of the sheet (i.e., the area of the sheet near the nominal top of the book). The other remarkable result of these analyses that was not apparent from the pH indicator testing was the variability of the loadings for the different sheets. The average loadings ranged from a high of 0.48% MgO for the newsprint, to about 0.2-0.3% for the three sized papers (Clear Spring offset, alum/rosin-sized, and alkaline-sized), down to 0.15% for the supercalendered sheet and 0.06% for the Sterling litho gloss sheet. This trend suggests that the smoother calendered and coated sheets may have a lesser tendency to adsorb the particles during the treatment, or the particles may be more easily detached from these sheets in subsequent handling. It is also possible that the observed variability in the particle loadings is related to the book structure (i.e., the position of the papers in these test books) rather than on the properties of the papers themselves. More testing is needed to determine if these types of sheets are inherently more resistant to treatment.
Evaluation of the twenty-five book test batch provided a better opportunity to assess the process performance on a wider variety of paper types, sizes, and book structures. As with several papers in the LC blue test books, five of the books in the test batch contained alkaline papers which could not be used to measure the process performance by a pH change from acidic to alkaline. As judged by the pH indicator color, in twelve of the remaining twenty books (60%) the pages were treated completely, and in four of the twenty (20%) the pages were treated completely except for about 1" in the gutter. In another four of the twenty (20%) larger areas of the pages (up to 95% of the page area) were left incompletely treated, and these page areas also tended to be toward the gutter. The ICP/mass spectroscopic analyses of pages from these last four books were performed to determine the magnesium particle distributions on the pages. They generally indicate the low particle loadings in the gutters of the books. More significantly, though, these analyses show that the edge areas of each of the tested pages had sizable particle loadings comparable to those obtained for the pages of the LC blue test books, about 0.3-0.75% MgO. Because the pH indicates acid conditions even in the presence of this particle loading, it suggests that these book papers may have been very acidic originally and that the treatment process parameters that were able to satisfactorily neutralize the LC blue test books papers may have been insufficient to deacidify these very acidic papers. This possibility suggests that such old, acidic papers may be more realistic bench marks for developing conservative process parameters than the less acidic papers of the LC blue test books. More experience should be gained with such old book materials in order to judge the likely outcome on more typical library holdings.
The focus of this evaluation is to attempt to answer the broad question of whether application of this treatment is effective in prolonging the useful service life for books. This seemingly simple objective is not so straightforward, however, for it requires both the specification of the quantifiable desired properties which denote "useful service life," and the choice of how to predict the future performance of the treated books. No attempt has been made to resolve these issues, which continue to elicit debate. Instead, the test protocols which the Library of Congress has chosen to perform in evaluating other deacidification processes have been followed here, with some minor additions, for these probably represent the current state of understanding of how to test for future paper performance. The three acidic papers in the LC blue test books (Clear Spring offset, alum/rosin-sized, and newsprint) were examined, and their physical properties (MIT fold endurance, zero-span tensile strength and finite-span tensile strength/stretch/stiffness/energy absorption, and internal tear resistance, each measured in the machine and cross directions of the paper), appearance (brightness, opacity, L/a/b color), and indicators of cellulose chemistry (extraction pH, alkaline reserve, hot-alkali solubility, viscosity DP) were monitored as the papers were aged in a humid oven. Since the changes in the measured properties did not usually occur linearly with time, linear rates calculated for the data are not particularly relevant. Instead, the oven-aging times required for the treated and untreated papers to reach equivalent values were used as approximate measures of the relative degradation rates for that particular property.
With the exception of the tensile stiffness which remained essentially constant, all of the other physical properties measured for the untreated papers declined during the thirty day humid oven treatment. Treated papers generally tended to show less degradation of mechanical properties than untreated. The most pronounced differences were observed in the fold endurance, which declined about 2-3 times more slowly for the treated sheets of all three paper types. Finite-span and zero-span tensile strengths also declined 3-4 times more slowly for the treated alum/rosin-sized and Clear Spring offset papers, and the other tensile properties (stretch and tensile energy absorption) also showed comparable factors of 2 differences between treated and untreated sheets. For the newsprint, the changes in tensile properties tended to be smaller and less precise, and determining significant differences between treated and untreated sheets was more difficult. However, the overall trend seems to be a slight (factor of 1.5-2) decrease in the rate of decline of these properties for the treated newsprint. For all three paper types, tear strengths for all the treated sheets showed small decreases in degradation rate (factor of 1.5-2) compared to untreated pages. Aging of these papers in a humid oven changed their appearance, causing them to darken (decrease in measured brightness and lightness parameters) and become less translucent (increase in opacity). The Bookkeeper treatment had no significant effect on these appearance changes, with the possible exception of the Clear Spring offset paper, which seemed to darken slightly less rapidly (by a factor of about 1.5) following treatment.
These measures of mechanical properties and appearance are the quantities which describe the desirable characteristics of the paper, and slowing the rate of their decline is an indicator of the long-term benefits resulting from the treatment. However, addition of the treatment process chemicals can potentially alter not only the rates of reactions but the overall degradation chemistry itself. Should this occur, the premise of these oven-aging comparisons--that treated and untreated sheets are degrading by the same process, so relationships observed at oven temperatures will also apply to room temperature aging--would be rendered invalid. Unfortunately, there are no probes which allow precise characterization of the degradation chemistries in aging paper. The chemical measures employed here, viscosity and alkali solubility, are very crude monitors of the cellulose component of the paper. Viscosity monitors the molecular weight of the cellulose, or the average length of the cellulose chains. Alkali solubility is also a measure of molecular weight, increasing as the number of chain ends increases (i. e., as the molecular weight decreases), but it is also a measure of oxidation of the cellulose polymer, increasing dramatically as the cellulose chains become oxidized. Taken alone, alkali solubility cannot confirm or disprove cellulose oxidation, so this quantity for now is merely compared during the aging of untreated and treated papers.
The changes in viscosity and alkali solubility have been measured during the humid oven aging of the papers in order to confirm that the changes in these chemical properties are also slowed by the application of the deacidification treatment. The measured cellulose viscosity decreased and the alkali solubility increased for both treated and untreated sheets of all three paper types, which indicates that the molecular weight of the cellulose was decreasing during the aging. This is to be expected, for it is this chemical change in the cellulose that is thought to result in the loss of strength and elasticity of the paper. Unfortunately, with the amount of paper required for each of these measurements, statistics are poor. The alkali solubility measurements appear very scattered and show no significant differences between treated and untreated pages, except for the Clear Spring offset paper whose alkali solubility may have increased slightly less rapidly for the treated than untreated sheet. However, the general trend for the viscosity data seems to be a viscosity loss which occurs about twice as slowly for the treated papers as for the untreated. It is worth noting that this difference in the rate for the cellulose degradation following treatment is approximately the same as that observed for the deterioration in physical properties.
The other two chemical measurements performed during the oven aging of these papers were sheet-averaged extraction pH and alkaline reserve. These quantities were monitored because they are believed to be measures of the protection afforded against future acidity, either formed in the paper or incorporated from external sources. For all the paper types, the untreated sheets were moderately acidic (pH of 5.7-6.5) and became slightly more acidic during the oven aging (pH of 4.1-5.6). Treatment of all the paper types raised the pH of the sheets to alkaline levels, about 9-9.5 for the alum/rosin-sized and Clear Spring offset papers, and about 10 for the newsprint. Subsequent oven aging of the treated sheets caused the pH to fall slightly to about 8 for the alum/rosin-sized and Clear Spring offset papers, and about 9.5 for the newsprint. It is not known whether the slight pH decreases occurring during the aging of the treated sheets are a result of further neutralization of paper acidity or merely the conversion of the magnesium salts from more alkaline to less alkaline ones.
The alkaline reserve measurements indicate this treatment process introduces excess alkalinity which persists essentially unchanged throughout the oven aging. The magnitude of the sheet-averaged alkaline reserve is somewhat low, with the alum/rosin-sized paper and the newsprint showing an alkaline reserve of about 1% (CaCO3 equivalent), the Clear Spring offset paper about 0.7% after treatment. There are reasons to view these data with some caution, however. There is considerable scatter in the data, and the acidic (by pH determination) alum/rosin-sized and Clear Spring offset papers produced measurable alkaline reserves. Alkaline reserve measurements were repeated on a separate set of unaged papers that were taken from another LC blue test book that had been treated with the same process equipment. These results were very similar to the original data (with the exception that the acidic papers had no measured alkaline reserves), and they showed the newsprint with an average alkaline reserve of 1.5%, the alum/rosin-sized about 1.2%, and the Clear Spring offset paper 0.5%.
While the overall performance in the oven tests show benefits of the treatment, it is noteworthy that for decreases in average acidity by factors of 100-10000 (increases of 2-4 pH units), the measured degradation in the oven slowed by factors of 2-4. While this could be evidence that other processes are participating in the degradation that are not slowed by, or may even be aggravated by, alkalinity, it is more likely that this behavior is a reflection of the damage done by local acidity within the fibers before those acids can migrate out to the magnesium salt particles where they can be neutralized. This technology is similar to others in the apparently slight effect it has on paper aging. This effect deserves closer scrutiny, for it calls into question two central issues in this field--the efficacy of such externally situated alkalinity in alleviating internal acid attack, and the reliability of these humid oven tests in accelerating the migration processes as well as the chemical degradation processes.
Comparison of all the physical and appearance properties measured before oven aging for the untreated and treated sheets in the LC blue test books indicates that none of the quantities were affected significantly as a result of the treatment. That is, no detectable changes in the papers' mechanical properties or appearance were produced by this treatment, which is consistent with the general lack of side effects and interactions found in the qualitative and visual evaluations described elsewhere in this report. Even the slight chalking or white residues noticed on some treated books in the twenty-five book test batch were not detected instrumentally, probably because sensing of such appearance changes on the unprinted white papers was more difficult than observations of such alterations on glossy, darkcolored materials.
While this process appeared to be free of side effects immediately following treatment, it is likely that the full effects will not be observable until the process chemistry has had the chance to progress, at least as far as the conversion of the magnesium oxide to the active alkaline agents. Since the time when this conversion occurs is not known, it is also unknown whether the evaluation for treatment side effects reported here, occurring several months after the treatment, is adequate to assess the full extent of risks. As a first effort to probe the long-term changes which might accompany this treatment, samples of papers from the treated LC blue test books were exposed to high humidity by suspending them over a water bath in an closed container for 24 hours. A slight but noticeable darkening of the treated newsprint occurred, darkening which seemed more severe for the newsprint samples which had been oven aged prior to treatment. This darkening of deacidified newsprint is a side effect frequently encountered in other treatment processes, but it only occurred following the humidification of the Bookkeeper-treated sheets. While this should not be construed as an accurate appraisal of the potential for such belated after-effects, this observation points out the need for more critical consideration in deciding when and how to assess these risks.
None of the technical evaluations performed here showed clear indications of serious side effects from the high pH values expected for the magnesium salts applied in this process. If there is alkaline damage occurring in the treated papers, the net result of the treatment on the aging behavior still seems to be a net benefit, with the loss of desirable performance properties slowing down after treatment. It should be noted, however, that more thorough study, looking specifically for chemical changes in inks, adhesives, covers, etc., is required to assess fully the possibility of alkaline reactions in these book materials.
This mass deacidification technology has advantages of simplicity of design in the treatment process, and the results of the treatment are reasonably well-characterized. While the claims about the exact nature of the process chemistry seem plausible, they should be viewed with circumspection until further research has established the nature and rates of the salt conversions and neutralization chemistries in a variety of materials. In fairness, many of these issues--the identity of alkaline agents formed, the nature and fate of reaction products, the effectiveness of alkaline salts in neutralizing acidity which is not in proximity to the alkali--are no less well understood for this treatment than for any other. Nevertheless, the answers to these questions will provide a better basis by which to judge the viability of this process and the risk of other unexplored side effects. It remains incumbent on the vendors to continue these efforts to provide evidence in support of their claims.
Clearly the most immediate obstacle to the success of this technology is its inability to control the application of the treatment to all book materials, and only further experience with the current equipment will clarify its capability. As the treatment was applied to these test materials, the uniform application of adequate alkalinity, especially in the gutter of the book, was not assured. Particle and magnesium salt concentration surveys indicate that deposition does not occur uniformly across pages. While process parameters seem to be adequate for providing overall neutralization of pages in LC blue test books, page-averaged alkaline reserve measurements suggest that even greater particle deposition is needed for these books, for the values attained fall short of the 1.5% alkaline reserve of the Library specification. Performance on the wider variety of book materials in the twenty-five book test batch indicates that the process parameters used may not be conservative enough to deal with more difficult, but nevertheless typical, books.
It is impossible to determine whether the insufficient particle loadings in some books can be rectified with the technology used in August 1993. Some books may simply be much more acidic than the LC blue test books, and may require greater particle loadings. It is also possible that fluid flow through some book structures and/or sticking efficiency of the particles on the paper may be the limitation in particle deposition, in which case it may not be possible to assure adequate treatment by realistic adjustment of the process parameters. Further examination of such difficult book structures (e.g., small, tightly bound or oversewn books whose pages resist fanning) or paper types (smooth, coated, or highly acidic papers) should be helpful in determining whether this problem can be corrected. Only when the process parameters have been adjusted to meet the Library specifications can realistic projections be made about the potential for scaling up to address mass treatment needs.
With the exception of the low particle loading (and consequent nonuniformity of treatment and low average alkaline reserve), the overall performance of this treatment is comparable to that required in the Library specification. The treatments did not measurably affect the physical or appearance properties of the papers in the LC blue test books. The oven-aging tests generally indicated some long-term benefits to the paper as a result of the treatment. Treated papers generally retained their desirable performance properties for 2-4 times longer in the oven, which is comparable to the factor of 3 improvement required in the Library specification. As noted above, initial alkaline reserve values were somewhat lower than the Library specification, but the initial value of the alkaline reserve did not fall during the 30 days of humid oven aging.
In summary, it would appear that the most immediate shortcomings of the process involve the application of the treatment chemicals to book materials, rather than the process technology itself. These shortcomings may be remedied by changes in the process parameters or by modifications of the equipment. At this stage the Bookkeeper process should continue to be viewed as a potentially viable candidate for mass deacidification technology while its performance in treating a variety of library books is evaluated. Rather than focusing so narrowly on the performance on the LC blue test books, particular attention should be paid to developing conservative process parameters to adequately treat "difficult" book papers and structures, so that the capabilities and limitations of the process can be better defined. Further research should also be devoted to supporting the claims of the process chemistry which occurs at ambient conditions.
The physical appearance and condition of collections following mass deacidification are as important to those whose responsibility it is to ensure longevity and access as are chemistry, health and safety, and environmental impact. It was critical, therefore, to include the expertise of a paper conservator in the evaluation process. Guided by the conservator, and the questions submitted in response to an invitation issued to the library and archival communities through the Conservation DistList, the Team decided to focus on the following potential problem areas: posttreatment color changes of paper and binding materials; instability and color changes of the inks; distortion of the textblock or the paper; residual effects such as chemical rings or blemishes; damage to adhesives; damage to labels, pockets, or security tags; distortion or damage to plastics or binders; damage to case bindings, covers, book cloth or leather; and noticeable odors. Twenty-five books, in addition to the mandated LC blue test books, were selected to provide a range of ordinary materials similar to that found in libraries. After they were cut in half, the Team inserted rubber bands, paper clip, and staples, and marked them with a variety of inks, pencil marks, and marker pen colors. Each Team member used a form designed by the Team to record observations resulting from a comparison of the treated with the untreated half. This report is based on the conservator's expertise and the Team evaluation. The questions she poses for PTI result from this work.