HOW THE COMPONENTS OF FORMATION OF FINE PAPER AFFECT PRINTABILITY FOR DIFFERENT PRINTING PROCESSES

Transcription

1 HOW THE COMPONENTS OF FORMATION OF FINE PAPER AFFECT PRINTABILITY FOR DIFFERENT PRINTING PROCESSES Jean-Philippe Bernié, Patrick Hurd*, Patricia Sutton**, W.J. Murray Douglas Department of Chemical Engineering McGill University 3420 University Street Montréal, Québec, Canada H3A 2A7 * Georgia-Pacific Corporation 55 Park Place, 15th floor Atlanta GA pghurd@gapac.com ** OpTest Equipment Inc. 900 Tupper St. Hawkesbury, Ontario, Canada K6A 3S3 psutton@optest.ca ABSTRACT Paper printability depends on formation. This relation has proved difficult to establish because of the difficulty in quantifying formation. Standard formation instruments provide a single-number formation index, blind to the concept of scale of formation. Using a new technique which partitions formation into its components as a function of scale of formation, the present study quantifies the relationship between the components of formation of unprinted paper and subjective solid print quality evaluation. For two sets of fine paper, same grade, printed with different processes, the Harris and Heidelberg presses, we established the specific range of scale of formation which controls print quality, and the importance of this effect. For the Harris press the important formation component was at 2mm scale, with a correlation to print quality of R 2 = For the Heidelberg press the relevant scale range was 3-5mm, with R 2 = 0.82, confirming a previous study. Thus the key formation scale and R 2 level are specific to the printing process. Moreover, as the print quality sensitivity to the key formation component is high, formation of unprinted paper can be an important predictor of print quality and thereby the central element of a control loop for optimizing product quality. INTRODUCTION: ESTABLISHING THE PAPER QUALITY CHAIN Formation measurement involves analysis of paper structure in the plane of the sheet. This analysis can be based upon measurement of local grammage, but generally the preferred measurement is local opacity, which can be easily obtained by light transmission. Conventional commercial instruments generally collapse two-dimensional local opacity maps into a single number such as the "Formation Number" (coefficient of variation of local grammage or of local opacity) or any of a variety of proprietary single-number indices of formation. However it is

2 impossible to describe the complex two-dimensional structure of a sheet of paper with any singlenumber index of formation. Therefore these instruments provide an oversimplified characterization of sheet formation. Paper properties depend on formation, but this relationship is scale of formation dependent. For example, small flocs do not affect printability or strength in the same way as large flocs. Thus it is necessary to document paper formation as a function of scale of formation. Consequently single-number formation index instruments are of very limited utility for predicting or controlling sheet properties or for paper machine control. It has therefore proved very difficult to relate properties quantitatively to formation although it has been long known that formation has a strong impact on paper properties [1-3]. To solve this problem Bernié and Douglas developed the "components of formation - scale of formation" technique, which addresses the need for a meaningful description of sheet formation. To evaluate the quality of the formation over different scales of formation, formation nonuniformity is partitioned into its components over a range of scale of formation from 0.6mm to 37mm. This method has been used in extensive studies with a variety of grades, from tissue to printing papers to linerboard. The overall objective is to use the components of formation as the key variable for papermaking process control. Formation depends on the process parameters and in turn affects end-use product properties. It is therefore necessary to establish two sets of relationships: between papermaking parameters and components of formation, and between components of formation and paper properties. With these relationships it becomes possible to use formation as a predictor of product quality and as the central variable for process control. This linkage has been named the Paper Quality Chain: from papermaking parameters, to components of formation, to paper properties. This Paper Quality Chain has been explored in a series of studies [4-13]. It was shown that paper strength properties and print quality are controlled by formation components over only a specific range of scale of formation, a range dependant on the particular combination of paper grade - paper property [4, 5, 9, 11, 12]. The other part of the Paper Quality Chain, tracing the papermaking parameters which are the source of the various components of formation, has been initiated [7, 13]. In 1996 Shallhorn and Heintze of Domtar studied the effect of formation on the offset print quality of uncoated fine papers [14]. They printed 32 samples of grammage g/m 2 on a commercial Heidelberg press. Print quality visual ranking was established by an all-pair ranking technique [15] with 34 observers. Formation was measured with the M/K instrument, and an R 2 value of was found between print quality ranking and the M/K single-number formation index. Subsequently this work was extended as follows. The formation of these sheets was determined using two other conventional formation test instruments, the NUI and the Micro- Scanner. Values of R 2 of 0.22 and 0.35, respectively, were found for the correlation between print quality and the formation indices of these instruments. On the same samples of paper, Bernié and Douglas [4] determined the components of formation and correlated the results to the previous determination of print quality visual ranking. It was established that the specific scale of formation that controls print quality in this case was the 4-8mm range. For this range of scale the R 2 correlation value between these components of formation and subjective print quality was 0.56, about double that found for the three conventional formation measurement instruments. These results indicate that formation, appropriately determined, has a much stronger impact on print quality than had been indicated based on the use of single-number formation instruments.

3 After these Domtar - McGill studies, carried out with paper printed on a Heidelberg press, we now examine how changes in the printing process affect both the specific range of scale of formation that determines print quality of fine paper and the strength of the formation - print quality relationship. EXPERIMENTAL: COMPONENTS OF FORMATION Two sets of bright offset paper from the same grade, grammage g/m 2, were used, one of 14 sheets, the other of 11 sheets. The sheets came from 8 paper machines located in 5 mills (3 were Georgia-Pacific mills, 2 mills were of other companies). These sheets were printed on commercial presses, the set of 14 sheets on a Harris press, the 11 sheet set on a Heidelberg press. The print quality of these solid prints was evaluated visually. Formation was measured with three methods: the M/K instrument, a visual panel evaluation, and the new McGill PaperPerFect components of formation - scale of formation technique. In the latter case the formation was measured on 280x430mm (11x17 inch) sheets, with 8 fields of view of 66x66mm. Thus the components of formation results represent the average of these 8 determinations. The three alternate determinations of formation were correlated to the visually evaluated print quality. Fig. 1 shows the components of formation of 4 representative samples out of the total of 14 samples from the set for the Harris press. For facility of interpretation the results are expressed normalized, at each value of scale of formation, to the average of the 14 samples. A component of formation of 1.1 means that at this specific scale of formation, the formation nonuniformity of this sheet is 10% less than the 14 sample average, i.e. that the formation of this sheet is 10% better than for the 14 sample average. A component of formation of 0.9 means that at this value of scale, the formation is 10% worse than that of the 14 sample average. Fig. 1 shows that there are substantial differences in formation, and moreover that different sheets have significantly different formation patterns relative to scale of formation. For example, over the 0.6-2mm range of scale sheet A has poor formation at small scale of formation, being about 20% worse than average, but improves gradually as scale of formation increases above 2mm, and has a much better formation than the set average at the three largest values of scale of formation, 14, 22 and 37mm. Thus it is not possible to state that sheet A is of better or worse formation than the set average - it is both! This example demonstrates the necessity of partitioning formation into its components to describe sheet structure characteristics adequately. Sheets E and H are seen on Fig. 1 to be always of better formation than the set average formation, but in quite different ways. At small scales of formation sheet E formation is very much better than the average, up to 40% better in the 2-5mm range of scale. However the quality of formation of sheet E gradually decreases to become, by 37mm scale, indistinguishable to that of the set average. For sheet H it is the opposite: at small scale its formation quality is indistinguishable from the set average, but at larger scales it becomes much better, by up to more than 40%. As for sheet J, its formation is better than the set average at both extremes of scale of formation, but is distinctly worse, by up to 20%, in the intermediate range of formation scale, 3-5mm. Thus, like sheet A, it is impossible to state that sheet J is of better or worse formation than average as it is both, depending on the scale of formation. Thus for these 4 sheets, if the important scale of formation is in the small range, sheet E is the best and sheet A the worst. However if for a key paper property the important scale of formation is in the large range, sheet A is the second best and sheet E is the worst. Fig. 2 shows the components of formation of 4 sheets from the set II, printed on the Heidelberg press, normalized to the set average; again we have a large variability of formation and different patterns of formation behaviour relative to scale of formation. Sheets G and E are

4 better than the set average for all scales of formation, but in very different ways. Sheet G is only slightly better than the set average at extreme scales of formation, while the formation of sheet E is much better than the set average for smaller values of scales. The results in Figs. 1 and 2 demonstrate that the components of formation technique is essential for determining formation characteristics in a meaningful way. RESULTS: FORMATION COMPONENTS - PRINT QUALITY CORRELATION Paper was printed using the 4-color offset process in either of two ways: Harris M-1000B web offset press. Press operation speed was 1200 feet/min. Heidelberg Speedmaster 74 sheet fed offset press. The press was operated at a printing rate of 8000 impressions per hour. Both presses are located at the Rochester Institute of Technology (RIT). Standard 4-color test plates developed by RIT for the purposes of comparative print evaluations were used on all paper printed. Ink density targets were as follows: Web: Black = 1.15; Cyan = 1.00; Magenta = 1.00; Yellow = 0.85 Sheet: Black = 1.00; Cyan = 0.85; Magenta = 0.85; Yellow = 0.70 Solid print quality was determined visually by a panel of 25 evaluators. The panel consisted of paper and printing industry professionals, who are accustomed to making subjective judgements on aesthetic print quality features. Images of a light colored subject against a dark colored background were printed from the test plates. The evaluations were done on pictorial images (a light colored flower against a very dark background, and a basket of fruit (of all different colors), so all 4 colors are combined in different ways. The panel was instructed to judge these images in terms of image sharpness of the primary subject and print mottle of the background. The evaluations were made on a 4-point scale between best and worst print quality. Figs. 3a and 3b show the correlation between M/K formation index and print quality for both types of printing press while Figs 4a and 4b show the corresponding correlations for formation evaluated visually. For the correlations to subjective print quality, the R 2 correlation values with the M/K instrument are 0.23 (Harris press) and 0.44 (Heidelberg) while with formation evaluated visually these R 2 values are 0.21 (Harris press) and 0.50 (Heidelberg). With average R 2 values of about 0.35, these two single-number formation evaluation methods provide too low a level of correlation with print quality to be of practical utility. Correlation between print quality and the PaperPerFect components of formation gives more useful results. For Harris press printing Figs 5a and 5b show the correlation between subjective print quality and two components of formation selected for illustration, at two values of scale of formation, 0.8 and 2mm. On Fig. 5a, 2mm scale of formation, there is significant correlation, with an R 2 correlation value of 0.52 indicating that just over half the variation in print quality comes from changes in the 2mm component of formation. However at 0.8mm scale of formation there is very low correlation, with an R 2 of Thus with the Harris press formation at this very small scale has little effect on printability. It is impressive to note that components of formation at values of scale as close as 0.8 and 2mm can have such substantially different effects on printability. Likewise for Heidelberg press Fig. 5c shows the correlation between subjective print quality and 5mm component of formation. At this scale the correlation is exceptionally high, with an R 2 of For all 10 values of scale of formation the R 2 values for the print quality - components of formation correlation were likewise determined. Figs. 6 and 7 show the R 2 values as a function of scale of formation for both types of press used. For the Harris press the correlation between print

5 quality and components of formation passes through a maximum of R 2 = 0.52 at 2mm scale of formation and drops for scales of formation both above this 2mm value. Thus Harris press print quality is controlled dominantly by formation in the range of scale 1.2-3mm. This R 2 value of 0.52 is more than double that found for the M/K and visual formation evaluations. Thus by partitioning formation into its components and determining which components are important for print quality, we establish that in this case the formation effect is more than twice (R 2 of 0.52 vs 0.23 and 0.21) than indicated by single-number formation index methods. For Heidelberg press printing, Fig. 7 shows that again there is a maximum in the R2 value for the print quality - components of formation correlation as a function of scale of formation. As was found for Harris press printing, with Heidelberg printing the correlation to print quality for the most important component of formation, R 2 = 0.82, is again much higher than the 0.44 and 0.50 values found with formation measured by the M/K instrument and by visual evaluation. Comparison of Figs. 6 and 7 shows that the print quality - formation components correlation for printing with the Heidelberg press differs in three ways from that for the Harris press: with the Heidelberg press the level of R 2 correlations is higher, the maximum R 2 occurs at larger scale of formation and the drop in R 2 values outside the maximum is less pronounced than with Harris press printing. With respect to the latter effect, the larger plateau observed for the Heidelberg press set may derive from the lower variability of formation pattern than for the Harris press set. Use of a larger sample size than in the present study, with just 11 sheets in the Heidelberg press set, would enable a finer resolution and a better identification, inside the 3-22mm range, of the most important range of scale of formation for Heidelberg press printing. The differences in the formation components for the 14 sheet Harris press set and the 11 sheet Heidelberg press set must be viewed relative to the precision of the measurements, i.e. the formation variability within a single sheet. For all sheets 8 independent determinations were made. The coefficient of variation for the 2mm component of the Harris press set and the 5mm component of the Heidelberg press set are, respectively, 2% and 5%. As these 2 and 5mm components of formation have been shown to give the best prediction of subjective print ranking, the sensitivity of these components to print ranking is now examined. The resolution of human evaluation of print quality is a change of 1 in 4 levels of ranking, i.e. between a ranking of 1 and 2 for example. This change in print ranking corresponds on Fig. 5a to a change of 60% in the value of the 2mm formation component, and on Fig. 5c to a change of 65% in the 5mm formation component. These 60% and 65% changes in the key component of formation must be seen relative to the precision of the determination. As the coefficient of variation for 2 and 5mm components are, respectively, 2 and 5%, if the resolution of the method is taken as 2σ, this indicates a resolution of about, respectively, 4 and 10% for these components of formation. Thus the resolution of the formation determination is 6 to 12 times higher than that of visual print ranking. Therefore the components of formation technique constitutes a precise predictor of print quality. Fig. 8 integrates the present results with those of the earlier Heidelberg press study [4] referred to already. This comparison confirms that the important range of formation scale for Heidelberg press printing is, at 3-8mm, somewhat larger compared to 1.2-3mm for the Harris press. The interaction mechanisms between the sheet, the press, and the ink for each printing process have different sensitivity to the various components of formation, depending on the inherent design characteristics of the press and how it is operated. Some differences between the two studies with the Heidelberg press may also be attributed to differences in how the subjective print quality evaluation was carried out. The results from the earlier study have the substantially

6 greater precision that derives from the larger sample size used, 32 sheets vs the 11 and 14 sheets used in the present study. CORRELATION OF PAPERPERFECT TO OTHER FORMATION RESULTS Standard methods of formation measurement, whether instrumental or using visual evaluation, provide some single-number as a formation index. Each instrument has its own algorithm. Some process the standard deviation of the grey level of the image, which provides an index that integrates all the information of the nonuniformity of the sheet, regardless of the scale of formation. Other instruments provide a single-number formation index that is sensitive to some range of scale but insensitive to others. By definition a single-number formation index cannot be highly sensitive to all scales of formation. A single-number index can, on one extreme, be weakly sensitive to all scales (such as the standard deviation), or on the other extreme, be fairly sensitive to some range of scale of formation and insensitive to others. Each single-number method has such a specific signature. It is a fixed characteristic, intrinsic to the instrument, which depends on both the algorithm used for data processing and the parameters of the image acquisition system (pixel size, field of view). Thus by correlating the components of formation to the results by any other technique, it is possible to establish the extent of the sensitivity of that technique to each scale of formation. With the results of the 25 samples of the Harris and Heidelberg press sets, Fig. 9 plots the R 2 values for the components of formation - visual formation correlation, as a function of scale of formation. Fig. 10 shows the comparable results for the M/K instrument. Figs. 9 and 10 show that the M/K instrument and the visual formation evaluation have similar signatures. Both methods show a maximum sensitivity for scale of formation in the 8-14mm range. The sensitivity of both methods drops off sharply both above and below the 8-14mm range of maximum sensitivity. It may be that the algorithm of the M/K instrument was designed to emulate visual evaluation of formation quality. The characteristic signature apparent in Figs. 9 and 10 means that these two single-number methods can detect formation features in the 8 to 14mm range of size with moderately good sensitivity, about 60 to 70% of that when the components of formation are partitioned and determined individually at these scales of formation. However, outside this 8-14mm range, Figs. 9 and 10 show that the sensitivity of these singlenumber methods quickly becomes insignificant. Thus if one of these methods is used for correlating formation to a paper property that depends on the 8-14mm range of scale of formation, then this method would provide moderately good results. But if these methods are used for correlating formation to a property that depends on formation outside the 8-14mm range of scale, then poor to totally unacceptable results would be provided because of the insensitivity of these methods to scales of formation outside the 8-14mm range. This explains why the correlation obtained between print quality and formation determined by these two methods was much lower than that obtained with the components of formation, as documented in Figs Only the scale of formation - components of formation technique, with its ability to analyze and segregate formation over a wide range of scale of formation, is able to identify each component of formation independently and to establish which components of formation are affecting any paper property and the sensitivity of this effect. SUMMARY AND CONCLUSIONS A clear correlation was found between visual print quality of fine paper and formation. For the Harris printing process, the controlling component of formation is at 2mm scale of formation, with an R 2 value of For Heidelberg press printing, the important scale of

7 formation is higher, in the 3-22mm range. The results from the use of the 11 sheets in the present study indicate that for Heidelberg press printing the most important range of scale of formation is 3-5mm. Results from an earlier study [4] using Heidelberg press printing with 32 sheets of fine paper identified 4-8mm as the important range of scale. Over the range of scale of formation found important, the correlation of Heidelberg print quality to formation was high, with R 2 correlation values of 0.56 in the earlier study and 0.82 in the present work. In spite of the differences between the two studies there is a good consistency between the most important range of scale of formation determined for Heidelberg press printing, 4-8mm in the earlier study, 3-5mm in the present one. For the important range of scale of formation the value of the R 2 correlation to print quality is high in all three cases: 0.52 for the Harris press, and 0.56 and 0.82 for the two Heidelberg press studies. Thereby it is established that a high proportion of the variability in print quality, 52% to 82%, comes from specific components of formation. Establishing that such a high percentage of print quality variation comes from formation is another benefit from switching from the use of conventional techniques giving some single-number index of average formation to the scale of formation components of formation method. Also, the key components of formation are determined with high precision, the coefficient of variation for the 2mm and 5mm formation components being in the present study of 2% and 5%, respectively, or 2σ values of 4 to 10%. As one level of visual print ranking corresponds to 6 to 15 times this precision of determination of the relevant formation components, this method offers a very high sensitivity. Based on these results, the strategy for print quality optimization for fine paper of these grades would be to adjust papermachine parameters that improve the important component of formation, 2mm for Harris press printing, 3-8mm for Heidelberg press printing. Thus the findings of the present study are very promising for the prospect of using components of formation of this grade of fine paper to predict the key product quality, printability. REFERENCES 1] Lyne, M.B., and Hazell, R., Formation testing as a mean of monitoring strength uniformity, The Fundamental Properties of Paper Related to its Uses, Transactions of the Cambridge Symposium, ed. BPBIF, 1, p (1976) [2] Norman, B., and Wahren, D., Mass distribution and sheet properties of paper, The Fundamental Properties of Paper Related to its Uses, Transactions of the Cambridge Symposium, ed. BPBIF, 1, p (1976) [3] Kajanto, I.M., Komppa, A., and Ritala, R.K., How formation should be measured and characterized, Nordic Pulp and Paper Research Journal, 3, p (1986) [4] Bernié, J.-Ph. and Douglas, W.J.M., Exploration of the print quality - paper formation relation, Proceedings, TAPPI Process & Product Quality Conference, Jacksonville, p (1997) [5] Bernié, J.-Ph. and Douglas, W.J.M., Paper strength - paper formation relations, 84th CPPA Technical Section Annual Conference, Montréal, preprints p. A330-A332 (1998)

8 [6] Bernié, J.-Ph. and Douglas, W.J.M., Role of scale of formation in monitoring papermachine CD variability of formation, Proceedings, TAPPI Process & Product Quality Conference, Milwaukee, p (1998) [7] Bernié, J.-Ph., Cheung, P. and Douglas, W.J.M., Components of formation for newsprint: CD variability and papermachine parameters, 85th meeting, Preprints, Pulp & Paper Technical Association of Canada, Montréal, p. A332-A335 (1999) [8] Bernié, J.-Ph. and Douglas, W.J.M., Full sheet mapping of components of formation for machine-wide strips of newsprint, Proceedings, TAPPI Engineering/Process & Product Quality Conference, Anaheim, p (1999) [9] Hashemi, S.J., Tawfik, R.P., Bernié, J.-Ph. and Douglas, W.J.M., The partitioning of formation into its components for analysis of paper strength properties, Proceedings, TAPPI Engineering/Process & Product Quality Conference, Anaheim, p (1999) [10] Bernié, J.-Ph. and Douglas, W.J.M., Scale of formation for differentiation of sheet structure between grades of paper, Proceedings, International Paper Physics Conference, San Diego, p (1999) [11] Bernié, J.-Ph., Romanetti J.L., and Douglas, W.J.M., Use of components of formation for predicting print quality and physical properties of newsprint, 86th meeting, Pulp & Paper Technical Association of Canada, Preprints p Montréal (2000) [12] Bernié, J.-Ph., Journeaux, and Douglas, W.J.M., Prediction of print quality using the components of formation of unprinted paper, Proceedings, 2000 International Printing & Graphic Arts Conference, Savannah (2000) [13] Bernié, J.-Ph. and Douglas, W.J.M., Effect of a wet-end additive on the components of formation of tissue, Proceedings, TAPPI Papermakers Conference, Cincinnati (2001) [14] Shallhorn, P.M., and Heintze, H.U., Offset printing and the formation of uncoated fine papers, Proceedings, TAPPI Product Quality Conference (1996) [15] Heintze, H.U. and Rutland, D.F., Visual scale quality correlates with small-scale opacity change, Pulp and Paper Canada, 79 (3): T102 (1978)

9 Relative components of formation A E H J Ligne Scale of formation (mm) Fig. 1: Relative components of formation of four samples from Harris press set Relative components of formation C E G K Scale of formation (mm) Fig. 2: Relative components of formation of four samples from Heidelberg press set

10 4 y = 0.044x R 2 = Print ranking M/K formation index Fig. 3a: M/K formation - visual print ranking correlation: Harris press 4 y = 0.056x R 2 = Print ranking M/K formation index Fig. 3b: M/K formation - visual print ranking correlation: Heidelberg press

11 y = 0.445x R 2 = Print ranking Formation Score Fig. 4a: Visual formation - visual print ranking correlation: Harris press 4 y = 0.627x R 2 = Print ranking Formation Score Fig. 4b: Visual formation - visual print ranking correlation: Heidelberg press

12 4 y = 1.662x R 2 = Print ranking mm relative component of formation Fig. 5a: Visual print ranking - 2mm component of formation correlation: Harris press 4 y = 1.339x R 2 = Print ranking mm relative component of formation Fig. 5b: Visual print ranking - 0.8mm component of formation correlation: Harris press 4 y = 1.812x R 2 = Print ranking mm relative component of formation Fig. 5c: Visual print ranking - 5mm component of formation correlation: Heidelberg press

13 R Scale of formation (mm) Fig. 6: Effect of scale of formation on the visual print ranking - components of formation correlation: Harris press R Scale of formation (mm) Fig. 7: Effect of scale of formation on the visual print ranking - components of formation correlation: Heidelberg press

14 R Present study, Harris press Present study, Heidelberg press 1997 study, Heidelberg press, ref. [ Scale of formation (mm) Fig. 8: Effect of scale of formation on print quality - components of formation correlation: Present and previous studies

15 R Scale of formation (mm) Fig. 9: Effect of scale of formation on the visual formation - components of formation correlation R Scale of formation (mm) Fig. 10: Effect of scale of formation on the M/K formation - components of formation correlation