Printer Friendly

Wood color control during kiln-drying. (Solid Wood Products).

ABSTRACT

The coloration of wood during drying in a laboratory kiln was investigated and modelled using multivariate techniques: principal component analysis and partial least squares (PLS). The wood species included were Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies [L.] H. Karst), and birch (Betula pubescens Ehrh.). Color changes were determined at the unpianed surface and after planing off 1 and 3 mm from the surface using a photoelectric colorimeter. Color parameters presented were lightness, chroma, and hue. PLS models for the color for each species was a useful method and the results indicated possibilities for future batch-color control of kiln-drying. The coloration of the wood was equally intense during the kiln-drying when free water was present (early part) compared to when no free water was present in the wood (late part). The coloration was most intense at the wood surface, however, planing removes the major coloration induced by drying. Color homogeneity was not found to be affected by rais ed temperatures during drying.

In industrial wood processing, artificial wood drying is often required to achieve a fast and good quality result. Many different drying techniques are used, and the basic principles are circulation with hot air and steam, vacuum, microwaves, etc.

Kiln-drying is a process using heat and steam with air circulation; temperatures have generally increased in recent decades, mainly to speed up the process. This increase has lead to wood with somewhat changed properties, such as lower hygroscopicity, increased brittleness, darker color, etc. (13).

Besides drying, heat treatments in various forms have been used to change the properties of wood, such as rot resistance and color. For example, steaming has been traditionally used to enhance the natural color of walnut (1) and black locust (11). Application of heat and steam has been shown to increase the durability of wood against biodeterioration (4,15,16).

The kiln-drying process of today is often governed by schedules that regulate the kiln parameters needed (19). These traditional schedules are mainly based on the theory of diffusion flow of water. The diffusion is in some models considered to control the drying rate independent of the moisture content (MC) in the wood. However, recent studies on drying have shown that diffusion is not the predominant mechanism when free water (capillary water) is present in the wood structure. In this phase, heat and mass transfer control the drying rate (12,19).

Wood exposed to heat and moisture at the same time often shows color changes (6). The occurrence of free water is thought to cause color changes due to hydrolysis and migration of color precursors (2,8). The influence of free water on color changes that develop during kiln-drying has been recently studied (14). However, in the present work, we report on an attempt to divide the drying process into two phases, when free water is present and when free water is not present (below and above the fiber saturation point). These are expressed as the capillary phase and diffusion phase, respectively, in this work. This was done in order to be able to model the coloration by more easily defining the parameters of the kiln climate and to determine the effect of each individual phase on coloration.

MATERIAL AND METHODS

Green boards of Scots pine (Pinus silvestris), Norway spruce (Picea abies [L.] H. Karst) and birch (Betula pubescens Ehrh.) were taken from small-diameter logs from sawmills near Skelleftea, Sweden, in January to April of 2000 (Fig. 1). For each species, 21 to 26 boards 4.5 [+ or -] 1 m long (50 by 100 mm for spruce and pine and 38 by 100 mm for birch) containing both sapwood and heartwood were selected. Each board was cut into sections of 0.7 to 0.8 m, which were marked according to individual full-length board, section, and species. The board sections were divided into 9 groups of 30 pieces and stored outdoors, covered with snow, to prevent drying.

For each species, the experiments were mainly designed to have drying runs that could be comparative with each other and to emphasize the effect of the two phases (capillary and diffusion) on coloration. This was done by distributing the sections from each board and species to be included in all three drying runs and by drying with chosen temperatures for each phase (Table 1). Since the color induced by drying is considered to be strong when free water is present (8,14), lower temperatures were chosen in the capillary phase of drying (Fig. 2 and Table 1). Due to the same consideration, higher temperatures were chosen in the diffusion phase. This drying method is based on novel studies by Moren (12).

The laboratory kiln is based on air circulation, with heating, steaming, and ventilation (Fig. 3). The kiln, made of stainless steel, has dimensions of approximately 0.3 m high and 0.8 m wide across the airflow, and 1.4 m long in the direction of airflow. This made it possible to run 30 samples in 3 layers for each drying run. Stickers approximately 20 mm thick were used at the bottom, in between the layers, and on top, spaced approximately 0.7 m apart. The heating unit has a maximum power of 6 kW and the steam generator has a maximum power of 6 kW. The fan gives an air speed of approximately 2 m/see. No cooling or dehumidification device is included, only two adjustable vent holes with diameters of 0.1 m. A software program was used to govern the control by using signals from thermocouples. The two signals are dry-bulb temperature and wet-bulb temperature.

For control of MC during drying for the entire charge, the weight of a reference board was recorded continuously (Fig. 3). At the start of a drying run, the MC of the reference board was estimated by reference to a slice taken from it, whose MC was calculated by weighing and drying at 103[degrees]C for 16 hours. Thus, by using the estimated MC at start and the actual weight of the reference board, the point of time for transition from capillary phase to diffusion phase was decided, as well as the final point of time for the drying run. The transition time point was set when the MC dropped to 30 percent, approximately the fiber saturation point, and the final time point was when the MC dropped to 8 percent. This was, of course, not a precise method for deciding transition and final time since the MC of the reference board represented the whole batch.

The color measurements were done after the kiln-drying with a tristimulus colorimeter, Minolta CR 310, with a measuring head 50 mm in diameter. The color system chosen was CIEL x C x h, a suitable system for industrial purposes and easy to relate to practical experience of color, lightness, saturation/chroma, and hue (7). Five measurements were performed on each side of each board for unplaned, 1 mm planed, and 3 mm planed. Each batch contained 30 boards. This gives 300 measurements for each planing depth and each batch.

By visual selection, only measurements on clearwood were included in the data, whereas measurements containing knots, dirt, mold, mixture of sapwood and heartwood etc. were excluded. The data were then statistically treated, assuming normal distribution. Average values and the corresponding 95 percent confidence intervals were calculated from the data.

Multivariate analysis, principal component analysis (PCA) and partial least squares (PLS), were done for 3-mm planed boards. The color responses at a planing depth of 3 mm are considered the most important when manufacturing furniture, joinery etc. PCA was done with SIMCA P 7.02 statistical software (Umetrics AB) (20). This is a method that calculates correlation between variables and responses in a data set, i.e., examines the data structure, using the NIPALS algorithm (9). The program presents different parameters to diagnose the data set and two important ones are [R.sup.2] and [Q.sup.2]. [R.sup.2] expresses the explained variance. Close to 0, little variance is explained and close to 1 almost all. [Q.sup.2], based on cross-validation (20), expresses the predicted variance, and is often not acceptable below 0.3 and is satisfactory above 0.7. The difference between [R.sup.2] and [Q.sup.2] should in general be below 0.2 if the analysis is not to be suspected of overfitting the data set and containing unwanted noise. Results are often presented in score and loading plots. The score plots show the observations (i.e., measuring points) projected in two dimensions spanned by the principal components. The first principal component is in the direction were the data has the largest variation, the second in the direction were the second most variation of the data is and so on. The loading plots reveal the relationship between variables and/or responses in two dimensions and they are regarded as complementary to the corresponding score plots (i.e., it is possible to overlap them) (20). In an earlier and similar work, a multivariate analysis PCA was done (14).

Predictive models were made using PLS (3). This method also uses the NIPALS algorithm to investigate correlation. However, PLS in contrast to PCA also makes a linear least square fit between variables and responses and thus generates predictive models. [R.sup.2], [Q.sup.2] and plots are also used similarly in PLS as in PCA. The importance of each variable in the predictive model is presented in a "variable importance plot" (VIP), where values close to and over 1 are considered important and values less than 0.5 are considered to be of low importance.

RESULTS

In Table 2, all color measurements are presented as average values of each drying batch, side of the board, and planing depth. The variation around every average color value can be evaluated by regarding the half of the 95 percent confidence intervals in Table 2 and moreover a graphic example of the variation is given in Figure 4. The multivariate analyses are performed using PCA and PLS, based on average color values. PLS modelling was done separately for each species.

At first, in order to make an overview, a PCA was done on all three species involved (Fig. 5). In the PCA, responses are treated as variables. Eighteen observations (average color values for each species, batch and outer side or inner side) were used, only for 3 mm planed wood. A model with two principal components, t(1) and t(2). was generated for the data with [R.sup.2]x = 0.71 and [Q.sup.2] = 0.49. From Figure 5a, the score plot, Scots pine, Norway spruce, and birch appear as three groups. When the different species show separate clusters, interpretation of the variables in the loading plot shows that there is low reliability and separate models are recommended (20). w x c is the weighted loading in a direction corresponding to the direction in the score-plot. However, generally pine and spruce are lighter (higher lightness), more saturated (higher chroma), and more yellowish (higher hue) than birch. Birch, as can be expected, exhibited longer drying time in the diffusion phase (higher [t.sub.2]) than pine and spruce.

As a result of the PCA, predictive PLS models were created for the three species separately. The three species investigated were also treated at different conditions (Table 1). For these models, three variables and three responses were used: board side (0 = inner side, 1 = outer side), temperature in the capillary phase, temperature in the diffusion phase, lightness, chroma, and hue. The times for the different drying phases and the wet-bulb depression temperatures showed small variation and low precision and were therefore excluded from PLS modelling (Table 1). Six observations for each model/species were used (average color values for each batch and outer side or inner side).

The model for Scots pine was quite weak: [R.sup.2]x = 0.94, [R.sup.2]y = 0.66, and [Q.sup.2] = 0.33. The model is defined by (L* = lightness, C* = chroma (saturation), h = hue, Side = inner side (= 0) or outer side (= 1) of the board, [T.sub.1] = dry-bulb temperature in capillary phase; [T.sub.2] = dry-bulb temperature in diffusion phase):

L* = 90.540 - 0.466 X Side - 0.049 X [T.sub.1] - 0.055 X [T.sub.2]

C* = 13.458 - 0.919 X Side + 0.051 X [T.sub.1] + 0.087 X [T.sub.2]

h = 82.602 - 0.177 X Side - 0.012 X [T.sub.1] - 0.012 X [T.sub.2]

The variation of hue was large, which gave a bad prediction model (Fig. 6), and subsequently the variable Side (Fig. 7) for Scots pine was not successfully modelled, which then affected the total model. Figure 8b shows that hue is weak and that Side is of low dependence to the other variables and responses (almost orthogonal). However, for lightness and chroma, the modelling was successful and the importance of the variables is given in Figure 7.

Norway spruce was modelled in the same way as Scots pine, and this generated a strong model with [R.sup.2]x = 0.997, [R.sup.2]y = 0.93, and [Q.sup.2] = 0.85. The model is defined by:

L* = 98.351 - 1.351 X Side - 0.091 X [T.sub.1] - 0.097 X [T.sub.2]

C* = 6.898 - 0.153 X Side + 0.086 X [T.sub.1] + 0.091 X [T.sub.2]

h = 87.420 - 1.277 X Side - 0.032 X [T.sub.1] - 0.034 X [T.sub.2]

The importance of the variables is given in Figure 7. All three variables show importance to the model.

Birch was also modelled in the same way as Scots pine, and this also generated a strong model with [R.sup.2]x = 0.89, [R.sup.2]y = 0.936, and [Q.sup.2] = 0.85. The model is defined by:

L* = 81.511 - 3.654 X Side - 0.045 X [T.sub.1] - 0.105 X [T.sub.2]

C* = 24.428 - 0.821 X Side - 0.043 X [T.sub.1] - 0.023 X [T.sub.2]

h = 82.548 + 2.658 X Side - 0.074 X [T.sub.1] - 0.113 X [T.sub.2]

The importance of the variables is given in Figure 7.

For all species, the score plots in Figures 8a, 9a, and 10a show well-distributed observations according to variables and responses. Side is mainly spread in the vertical direction (outer side downwards) and the three drying batches in the horizontal direction of the plots (higher drying temperatures to the left). The loading plots reveal that Scots pine and Norway spruce become darker, redder, and more saturated when drying temperatures are raised (Figs. 8b and 9b). [T.sub.1] and [T.sub.2] oppose hue and lightness while they lie close to chroma in the first component. Birch becomes darker, less saturated, and redder when temperatures are increased, [T.sub.1] and [T.sub.2] oppose lightness, chroma, and hue, mostly in the first component. Furthermore, the loading plot also reveals that color differences between outer side and inner side are distinguishable. Scots pine inner side is lighter, more saturated, and more yellow than outer side, but these loadings are weak in the second component (Fig. 8b) and the v ariable Side is of low importance (Fig. 7). Norway spruce inner side is clearly more yellow, lighter, and indicated less saturated than outer side (Fig. 9b). Birch inner side is clearly darker and redder than outer side (Fig. 10b). Birch inner side is also somewhat more saturated than outer side.

The color variation within each drying run and side is presented in Table 2 (as half of the 95% confidence intervals); Figure 4 shows lightness for Norway spruce as one example. The variation around the average values is large when compared to differences between drying runs. The distribution around the average values were acceptably normal (Fig. 4) and similar distributions were obtained for the other color parameters and species.

DISCUSSION

The multivariate methods PCA and PLS were useful for analyzing the data collected and for modelling. The PLS modelling of each species showed stable and strong models, except for pine inner side, which was difficult to model due to the small variation for hue (Table 2, Fig. 6). Note that in this work, average color value was used since quite a large variation was expected (5,14,18) and also found in each unique set of measurements (Fig. 4). The models indicate that it is possible to govern color on an industrial basis.

Both phases, where free water (capillary phase) is present in the wood and not present (diffusion phase), were similarly important for the color change. This is confirmed when comparing with an earlier color study of kiln-dried Scots pine and Norway spruce (18). Many other investigations have noted strong coloration in treatments by the presence of capillary water (14) and steam (1,11). However, in this investigation, the time for the capillary phase was approximately 1 to 2 days, which are short times for inducing strong color changes. It is therefore reasonable to assume that not only hydrolysis (2) but other mechanisms are important, since free water is not necessary for the production of colored compounds. Coloration of the wood occurs throughout the entire drying procedure.

The two sides of the board, "inner side" and "outer side" (Fig. 1), often showed significant color differences (Table 2), and this was also indicated in the PLS models by the variable Side. These color differences are similar to those found in an earlier study (14). For inner side, mainly radial surfaces were measured; for outer side, mainly tangential surfaces were measured. An earlier color study on differences between radial and tangential surfaces of Scots pine (5) showed that color differences similar to those found in this study, between inner side and outer side, can increase if compensations are made. It is reasonable to assume that outer side measurements represent sapwood and inner side measurements represent heartwood. Further measurements with verified sapwood and heartwood are needed for a final evaluation. If there is truly a difference between sapwood and heartwood, it might be considered in future manufacturing of wood products where there can be strong interest in the color of wood.

In general, the largest color changes were found when comparing unplaned and planed wood (Table 2). Comparatively small color changes were found between 1-mm and 3-mm planed wood. The color difference for pine and spruce between unplaned and planed wood increases with increasing dry-bulb temperatures (Table 2), but not for birch, studied at lower temperatures. Measurements on planed and sawn Scots pine showed differences in color (5), but these were small in comparison with the differences found in this work. Thus, the surface of dried wood can be noticeably colored, but planing removes the major color change, even when higher temperatures are used in the drying process.

Significant color changes for 1-mm to 3-mm planed wood were observed (Table 2). Similar observations have been made previously (18). Sometimes a brownish layer just beneath the surface appears when drying wood (8,10,14,17) but no such layer was observed in this work. This shows that planing depth can be of importance if strong qualitative demands are stated for the color. It also raises the question of whether or not an even deeper planing depth than 3 mm will show distinct color changes.

Color homogeneity, the variation around each average value, was not in general found to be dependent on the temperatures used (Table 2, Fig. 4). However, a general dependency of color homogeneity upon temperature was found in an earlier work where the capillary phase was studied (14).

CONCLUSIONS

Multivariate PLS modelling is a powerful tool for batch color control of kilndried wood. During drying, coloration was found equally strong both when free water and no free water was present in the wood.

The author is a Graduate Student, Lulea Univ. of Technology, Div. of Wood Physics, Skeria 3, SE-93187 Skelleftea, Sweden.

[C] Forest Products Society 2002.

LITERATURE CITED

(1.) Charrier, F., B. Charrier, and G. Janin. 1999. Color changes of sapwood walnut (hybrid MJ209xRA) by Tannins (Tara) and Juglone during steaming process. In: Proc. of the 4th Inter. Conf. on the Development of Wood Sci., Wood Technology, and Forestry. Buckingham Chilterns Univ. College, High Wycombe, UK. pp. 116-122.

(2.) Fengel, D. and G. Wegener. 1989. Wood: Chemistry, Ultrastructure, Reactions. Walter de Gruyter Berlin, New York.

(3.) Geladi, P. and B.R. Kowalski. 1986. Partial least-squares regression: A tutorial. Analytica Chemica Acta 185:1-17.

(4.) Girard, P., P. Permadi, A. Mouras, and A. Napoli. 2000. Wood Torrefaction in Europe: An overview and its perspectives (poster abstract). In: Proc. of XXI IUFRO World Congress 2000, August 2000, Kuala Lumpur, Malaysia. 3:323. Forest Res. Inst. Malaysia.

(5.) Hagman, O. 1996. Analys av traytors farg. Sammanstallning av en doktorandkurs. (Color analysis of wooden surfaces. Summary of a Ph.D. Course). Tech. Rept. 1996: 01T. Lulea Univ. of Technology, Lulea, Sweden. (in Swedish).

(6.) Hon, D.N.-S. and N. Shiraishi. 1991. Wood and Cellulosic Chemistry. Marcel Dekker, Inc., New York.

(7.) Hunt, R.W.G. 1995. Measuring Color. 2nd ed. Ellis Horwood series in Appl. Sci. and Industrial Technology. Ellis Horwood Limited, Chichester, UK.

(8.) Kreber, B., M. Fernandez, and A.G. McDonald. 1998. Migration of kiln brown stain precursors during the drying of radiata pine sapwood. Holzforschung 52(4):441-446.

(9.) Lindgren, F. 1994. Third generation PLS. Some elements and applications. Thesis, Umea Univ., Umea, Sweden.

(10.) Millet, M. A. 1952. Chemical brown stain in sugar pine. Forest Prod. J. 2:232-236.

(11.) Molnar, S., I. Peszlen, H.G. Richter, L. Tolvaj, and F. Varga. 1998. Influence of steaming on selected wood properties of black locust (Robinia pseudoacacia L.). Acta Facultatis Ligniensis. pp. 38-45.

(12.) Moren, T. 2000. CT-Scanning of wood during drying: Consequences for the development of adaptive kiln-control systems (abstract). In: Prod. of Iberoamerican Conf. on Forest Prod. Res. and Developments. Laura Reyes Nunez, ed. Univ. of Bio-Bio and Univ. of Concepcion, Concepcion, Chile.

(13.) Sehlstedt-Persson, S.M.B. 1995. High temperature drying of Scots pine. A comparison of HT- and LT-drying. Holz als- Roh- und Werkstoff 53:95-99.

(14.) Sundqvist, B. 200_ Colour response of Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies [L.] H. Karst) and birch (Betula pubescens Ehrh.) subjected to heat treatment in capillary phase. Submitted for pub. to Holz als Roh- und Werkstoff.

(15.) Theander, O., J. Bjurman, and J.B. Boutelje. 1993. Increase in the content of low molecular carbohydrates at lumber surfaces during drying and correlations with nitrogen content, yellowing and mould growth. Wood Sci. Technology 27:381-389.

(16.) Tjeerdsma, B.F., M. Boonstra, A. Pizzi, P. Tekely, and H. Militz. 1998. Characterisation of thermally modified wood: Molecular reasons for wood performance improvement. Holz als Roh- und Werkstoff 56:149-153.

(17.) Viltaniemi, P. and S. Jamsa. 1996. Modification of wood with heat treatment. VTT Pub. 814:57. Technical Research Centre of Finland, Espoo, Finland. (in Finnish).

(18.) Wiberg, P. 1996. Colour changes of Scots pine and Norway spruce. Holz als Roh- und Werkstoff 54:349-354.

(19.) Wiberg, P and T.J. Moren. 1999. Moisture flux determination in wood during drying above fibre saturation point using CT-scanning and digital image processing. Holz als Roh- und Werkstoff 57:137-144.

(20.) Wold, S., K. Esbensen. and P. Geladi. 1987. Principal component analysis. Chemometrics and Intelligent Lab. Systems 2:37-52.

[Graph omitted]
TABLE 1

Experimental parameters (a)

Species Run no. [t.sub.1] [T.sub.1] [DELTA][T.sub.1]
 (hr.) ([degrees]C)

Pine 1 47 48 4
 2 49 65 4
 3 51 82 6
Spruce 4 20 63 3
 5 25 73 4
 6 21 82 4
Birch 7 37 40 4
 8 37 52 4
 9 40 57 3

Species [t.sub.2] [T.sub.2] [DELTA][T.sub.2]
 (hr.) ([degrees]C)

Pine 60 84 18
 37 82 17
 37 103 16
Spruce 38 92 17
 35 100 16
 33 111 19
Birch 58 69 17
 57 68 15
 54 85 17

(a)[t.sub.1] = time elapsed in capillary phase; [T.sub.1] = dry-bulb
temperature in capillary phase; [DELTA][T.sub.1] = wet-bulb temperature
in capillary phase; [t.sub.2] = time elapsed in diffusion phase;
[T.sub.2] = dry-bulb temperature in diffusion phase; [DELTA][T.sub.2] =
wet-bulb temperature in diffusion phase; nine dryng runs.
Table 2

Color results for Scots pine, Norway spruce and birch subjected to kiln
drying (a)

 Run Planing Outer side
Species no. depth/mm L* C* h

Pine 1 0 80.9 (0.2) 25.8 (0.3) 81.9 (0.2)
 2 77.4 (0.3) 27.8 (0.4) 80.6 (0.3)
 3 73.0 (0.3) 29.6 (0.3) 79.3 (0.2)
 1 1 83.9 (0.3) 21.9 (0.2) 81.8 (0.3)
 2 81.3 (0.4) 24.0 (0.2) 80.5 (0.2)
 3 77.4 (0.4) 26.3 (0.2) 78.7 (0.3)
 1 3 83.7 (0.3) 22.3 (0.2) 81.7 (0.3)
 2 82.4 (0.4) 23.3 (0.2) 80.6 (0.3)
 3 79.9 (0.2) 25.5 (0.1) 79.5 (0.1)
Spruce 4 0 78.2 (0.4) 26.6 (0.2) 81.0 (0.2)
 5 74.0 (1.0) 26.6 (0.5) 79.6 (0.3)
 6 70.5 (1.1) 28.3 (0.4) 77.6 (0.3)
 4 1 82.6 (0.2) 21.6 (0.2) 81.5 (0.2)
 5 81.0 (0.3) 22.2 (0.2) 80.3 (0.2)
 6 79.2 (0.2) 23.6 (0.2) 79.7 (0.2)
 4 3 82.3 (0.2) 21.5 (0.1) 81.2 (0.2)
 5 80.8 (0.3) 22.1 (0.1) 80.2 (0.2)
 6 78.7 (0.2) 24.2 (0.2) 79.7 (0.2)
Birch 7 0 74.0 (0.3) 28.0 (0.3) 77.4 (0.3)
 8 70.1 (0.5) 27.5 (0.4) 75.0 (0.4)
 9 70.5 (0.4) 26.5 (0.3) 76.2 (0.3)
 7 1 78.7 (0.4) 18.8 (0.3) 75.3 (0.2)
 8 78.6 (0.3) 18.7 (0.2) 75.5 (0.2)
 9 74.7 (0.3) 20.1 (0.2) 72.8 (0.3)
 7 3 75.4 (0.4) 20.2 (0.2) 74.1 (0.2)
 8 76.1 (0.3) 19.7 (0.1) 73.8 (0.2)
 9 73.6 (0.2) 19.3 (0.2) 71.5 (0.2)

 Inner side
Species L* C* h

Pine 78.5 (0.4) 25.9 (0.3) 78.7 (0.4)
 77.7 (0.4) 26.6 (0.4) 78.4 (0.3)
 73.4 (0.7) 29.4 (0.6) 75.7 (0.6)
 83.4 (0.2) 22.7 (0,2) 80.6 (0.4)
 82.9 (0.3) 23.3 (0.2) 80.6 (0.4)
 81.1 (0.2) 26.1 (0.2) 80.3 (0.4)
 83.1 (0.2) 23.5 (0.2) 80.8 (0.4)
 82.7 (0.3) 23.5 (0.2) 80.0 (0.4)
 81.5 (0.2) 26.9 (0.3) 81.3 (0.4)
Spruce 80.2 (0.3) 24.1 (0.2) 81.9 (0.3)
 79.0 (0.3) 25.5 (0.2) 81.2 (0.2)
 75.4 (0.3) 27.9 (0.2) 78.4 (0.2)
 83.7 (0.2) 20.4 (0.2) 82.1 (0.2)
 83.2 (0.2) 20.9 (0.1) 81.8 (0.2)
 80.8 (0.2) 23.4 (0.2) 81.2 (0.2)
 83.3 (0.3) 20.9 (0.2) 82.1 (0.2)
 82.6 (0.2) 21.6 (0.2) 81.9 (0.2)
 79.9 (0.2) 24.8 (0.3) 81.0 (0.2)
Birch 70.1 (0.5) 29.5 (0.4) 74.9 (0.4)
 67.2 (0.5) 27.2 (0.4) 73.6 (0.3)
 67.6 (0.4) 27.3 (0.3) 74.2 (0.3)
 74.5 (0.4) 21.0 (0.2) 73.2 (0.3)
 75.6 (0.3) 19.9 (0.2) 72.7 (0.2)
 70.6 (0.4) 20.8 (0.2) 69.7 (0.3)
 72.1 (0.4) 21.3 (0.2) 71.9 (0.2)
 73.0 (0.4) 20.4 (0.2) 71.1 (0.3)
 69.4 (0.4) 19.9 (0.2) 68.4 (0.3)

(a)L*=lightness; C*=chroma; h=hue. Boldface values denote a stepwise
significant color change between drying runs. Values in parentheses are
confidence intervals. All values are average values based on 70 to 150
measurements from 20 to 30 boards.
COPYRIGHT 2002 Forest Products Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sundqvist, Bror
Publication:Forest Products Journal
Article Type:Statistical Data Included
Geographic Code:1USA
Date:Feb 1, 2002
Words:4853
Previous Article:Validation of the ROMI-RIP rough mill simulator. (Solid Wood Products).
Next Article:Mold performance of some construction products with and without borates. (Solid Wood Products).
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters