Measuring resin-adhesive spray characteristics using a laser diffraction analyzer.
Droplet size and size distribution of resin-adhesive spray are often considered significant factors influencing resin efficiency in wood strand-based composite production. Most of the previous work on this subject has focused on measuring and describing droplet size and distribution as it appears on the wood. There is a need for advanced measurement technology to directly characterize wood resin-adhesive spray. This work used a laser diffraction analyzer to characterize the wood resin-adhesive spray generated by a spinning disk atomizer. The results demonstrated that the laser diffraction analyzer may be an effective tool for characterizing wood resin-adhesive spray. The real-time history record of spray characteristics clearly showed the variation of droplet size and size distribution over a period of time and aided in understanding online spray dynamics and recognizing the appearance of coarse droplets at the start-up of the spray process. The different ways in which spray data can be presented provides greater convenience for analysis of spray measurements. The pilot experimental set-up, however, needs refinement for longer periods of data collection.
Resin-adhesive droplet size and size distribution is often considered a significant factor influencing the properties of wood strand-based composites. Characterizing resin-adhesive droplet size and distribution has been, and continues to be, an important topic for researchers (Marian 1958, Burrows 1961, Carroll and McVey 1962, Lehmann 1965, Wilson and Krahmer 1976, Kamke et al. 1996, Loxton et al. 2003, Smith 2003). Few previous researchers, however, directly characterized droplet size and size distribution of resin-adhesive spray. One of the reasons for this lack of information is that most of the research on resin-adhesive droplet size focused on the size distribution of adhesive droplets on the wood elements rather than in the adhesive spray. Another limitation is that most investigations of droplet size and size distribution of adhesive spray were done during the 1950s to 1980s, which was more than a quarter century ago. At that time, measurement techniques of resin-adhesives' droplet size and size distribution were not as advanced as those presently available. The physical or mechanical properties were most frequently used to evaluate the effect of the spray method on the composite performance, rather than the spray efficiency (Burrows 1961, Lehmann 1965, Wilson and Krahmer 1976).
Two early researchers who measured droplet size and size distribution of resin-adhesive spray were Marian (1958) and Burrows (1961). Marian (1958) was the first to use an effective measurement technique, high-speed photography, to directly measure the droplet sizes of resin-adhesive spray. This technique obtained the droplets' diameter range from approximately 20 to 150 [micro]m when urea-formaldehyde (UF) resin was sprayed using conventional air atomization. Burrows (1961) attempted to determine the resin-adhesive droplet size, but used indirect methods, including throwing sheets of paper into the spray and then measuring the heat-cured resin-adhesive on the paper. Burrows, however, did not report the exact droplet sizes but instead characterized them as either "fine" or "coarse" droplets.
Therefore, it is timely to address this lack of knowledge of wood resin-adhesive spray characteristics, since a variety of droplet sizing techniques are available and have been successfully used for the characterization of liquid sprays (Schick 2007), including Optical Array Probe; Laser Diffraction Analyzer; Phase Doppler Particle Analyzer; Phase-Doppler Anemometry (Mitschke et al. 1998); Interferometric Laser Imaging for Droplet Sizing (Kawaguchi et al. 2002); Laser Induced Fluorescence (Park et al. 2002); and Shadow Doppler Particle Analyzer. Of the available measurement techniques, laser diffraction is an attractive technology, which provides a flexible and rapid method for the assessment of the delivered droplet size without the need for any external calibration.
Laser diffraction analyzers usually consist of a laser transmitter, a receiver, a sampling chamber, and a computer control and data collection system (Schick 2007). The technique is based on measuring the scattered light intensity caused by the droplets as they pass through the analyzer sampling area. The scattered light is measured by a series of photodetectors placed in the laser receiver at different angles to form the diffraction pattern. Then the diffraction pattern is converted to electrical and digital signals and calculated to the size distribution of the droplets using empirical functions programmed in the computer. The droplet size using this technique can be measured up to 3000/[micro]m. One limitation of using this technique is known as multiple scattering. Multiple scattering occurs when the light is scattered by multiple drops before reaching the detector, and therefore, introduces errors in computing the droplet size distribution. The error may be negligible for low-density sprays, but becomes more significant as the spray density increases.
The main objective of this study was to investigate the applicability of a laser diffraction analyzer for characterizing the droplet size and size distribution of wood resin-adhesive spray generated by a spinning disk atomizer.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The experimental spray unit included a laboratory scale Coil blender as atomization chamber, a spinning disk atomizer (Model EL-4), a Cole Parmer peristaltic pump (Model 77200-60) for accurate flow rate control, a Malvern laser diffraction analyzer (Spraytee) for spray size analysis, and a conventional canister type vacuum cleaner (Dayton 4TR12) controlled with a powered rheostat for pulling the resin-adhesive spray through the sampling cell. The head of the spinning disk atomizer was located at the center of the blender with a tilt angle 45[degrees] to the side wall. The spinning disk edge was inclined at an angle of 45[degrees] to the disk plane so that resin spray formed a conical pattern range projected at near 0[degrees] to 90[degrees] to the front wall of the blender. The opening of the vacuum hose was located on a protective Plexiglas wall and below the spinning disk atomizer head so that the vacuum could effectively collect the atomized spray (Fig. 1).
The Malvern analyzer is a widely used laser diffraction analyzer. The unit used in this study consisted of a laser transmitter, spray collection cell, laser receiver, and data acquisition computer (Fig. l(h)). The light generated by the laser transmitter was at a wavelength of 670 nm. The spray droplets were vacuumed though the collection cell and scattered the laser light passing through them. The measurement range of the Malvern equipment was from 0.25 to 1000 [micro]m. The data acquisition rate was set at 1 Hz (one measurement per second).
The specialized software (Spaytec RTSizer v5.40) was used to record, save, and analyze the real-time process history of the resin-adhesive spray particle size distribution.
At any given time the distribution of droplet diameters may be represented as a snapshot histogram or a cumulative distribution graph. As measurements are taken over a certain period of time, the distributions may vary from one instant to another reflecting the dynamics of the spray output from the atomizer. These changes are recorded in real-time history records (Fig. 2) generated by the spray analysis system. The droplet distributions are represented by nonparametric characteristics Dvl0, Dv50, and Dv90 ([micro]m), which are droplet diameters associated with 10th, 50th, and 90th percentile of the volume distribution as seen on the cumulative distribution diagram in Figure 3. In addition, the system measures the percentage of light transmitted through the spray and recorded by the laser receiver ("Transmission (%)"). The transmission level was used to control the time period of real recording history in this study. The transmission is directly related to the spray concentration and errors caused by multiple scattering. High spray concentration may significantly interfere' with diffraction measurements. Therefore, the transmission level of 90 percent was considered as a criterion below which the recordings were not reliable and then the measurements were stopped.
[FIGURE 3 OMITTED]
As the spray output may vary the droplet characteristics from real-time history records over a period of time, droplet size and distribution could be summarized and presented in standard statistical terms, including maximum (Max), minimum (Min), diameter associated with the average volume (Avg.), and volume standard deviation (SD).
In this study, all of the data collected by the Malvern analyzer were exported to Excel[R] (Microsoft) files on a personal computer. The tables and figures were created in Excel and mirrored the format presented in the Malvern analyzer.
As one of the predominant resin-adhesives used in the wood strand-based composites industry, phenol-formaldehyde (PF) BB-634 resin-adhesive provided by Dynea Company was used in the study. Its apparent viscosity was 0.246 Pa.s measured at 25[degrees]C using a Brookfield DV-I+ Digital Viscometer. The resin-adhesive feed rate was fixed at 100 mL/min; the resin-adhesive spray was generated at room temperature (21[degrees]C); and, the rotary speed of the spinning disk atomizer was set at 12,900 rpm. The airflow velocity through the analyzer was not measured in this study.
To address concerns regarding the potential effect of heat generated by the atomizer disk spinning at high speeds, the temperature of the dry disk (no resin) was measured with a hand-held laser thermometer while it was running at 15,000 rpm. It was determined that although the temperature on the edge of the disk increased slightly, it leveled off to a constant value just a few degrees Celsius above ambient. The increase in temperature was slightly higher near the bearing. It was concluded, however, that constant flow of resin over the surface of the spinning disk would effectively keep the edge of the spinning disk at room temperature. Consequently, no effect on the resin-adhesive viscosity was anticipated.
Results and discussion
Real-time history record
Figure 2 is a representative real-time history record, clearly showing the variation of the characteristic diameters Dv10, Dv50, and Dv90 and the transmission over 2 minutes. The Dv10 and Dv50 remained roughly constant at 5.0 [micro]m and 12.5 [micro]m, respectively. Dv90, however, varied considerably during this period of recording spray data. A noticeable production of coarse droplets between 90 [micro]m and 100 [micro]m in the Dv90 line was observed during the first minute of resin-adhesive atomization. After that first minute, droplet distribution appeared to stabilize, and the Dv90 remained constant at around 20 [micro]m.
The light transmission steadily decreased during the experiment. This might be attributed either to the increasing spray concentration as the atomization process continued or to the increasing receiver lens contamination.
An explanation for the appearance of the coarse spray droplet fraction is unclear; although it might be attributed to the nonuniform resin distribution at the surface of the spinning disk atomizer head at the beginning of the atomization experiment.
Average droplet sizes and size distribution over the measurement period
Tables 1 and 2 and Figure 3 show a sample data collection of droplet size associated with the average droplet size and size distribution of resin-adhesive spray over the measurement time presented in three different ways, which were directly averaged from the real-time history records shown in Figure 2.
In Figure 3, the horizontal axis represents particle diameters on a logarithmic scale. Dots represent droplet volume frequency (left vertical axis) while the line represents the cumulative volume frequency of droplet diameters (right vertical axis). It can be seen that except for the coarse droplets at around 100 [micro]m generated in the first minute of the measurement, the main droplet sizes ranged from 2.5 to 40 [micro]m with a bimodal size distribution within the main droplet size range. The primary droplet size distribution centered at around 15 [micro]m. The secondary droplet size distribution was around 5 [micro]m.
By examining the more detailed cumulative data collection of droplet size distribution of resin-adhesive spray listed in Table 1, it can be seen that the smallest droplets detected were 1.508 to 1.313 [micro]m. Table 1 also indicates that about 90 percent of the spray volume was composed of droplets smaller than 20 [micro]m, which is much smaller than the "94 [micro]m" optimal resin spray droplet size for effective bonding of strand composites, as predicted by Smith (2003).
Table 2 summarizes the averaged spray characteristics of atomized PF resin-adhesive in terms of standard statistical values: Avg., SD, Max, and Min. D [4,3] is the volume mean diameter (which was not shown in the figures). For Dv 10 and Dr50, the Avg. value and SD confirmed that these two variables remained roughly constant over the measurement time period, which was determined from Figure 2. The values of Avg., SD, Max, and Min for Dv90 indicates that Dv90 varied considerably (Fig. 2).
Fitting a mathematical model
The component distributions may be further determined by fitting a mathematical model with three normal distribution components to the measured data. One model is expressed as:
f(x) = [3.summation over (i=0)] [a.sub.i]/[s.sub.i][square root of 2[pi]] x exp (-[(x - [m.sub.i]).sup.2]/2[s.sup.2.sub.i]) 
where the independent variable x is the logarithm of the droplet diameter expressed in microns and [m.sub.i] and [s.sub.i] are the parameters of the normal distributions representing it's mean value and SD, respectively. The term [a.sub.i] is the scale parameters of the normal distributions.
Figure 4 illustrates examples of the individual distribution components and the overall model curves compared with the experimental data shown in Figure 2. Model parameters resulting from the curve fitting at the 95 percent confidence level are listed in Table 3. The goodness-of-fit was above 0.99 in all cases. Related non-parametric characteristics Dv10, Dv50, and Dv90 for the component distributions and the percentage of resin-adhesive volume carried by each component distribution are also listed in Table 3, which provides detailed size distributions for each individual component.
Application limitation of the experimental set-up
The above results and discussion suggest that a laser diffraction analyzer is an effective and convenient tool for characterizing the resin-adhesive spray. But, one drawback was noticed by visual inspection and can also be seen in Figure 2. The period of real-time history for each record was relatively short (2 rain), which was directly attributed to the decreasing light transmission percentage during the measurement process. This light transmission reduction was presumed from the spray concentration, and also the contamination of the analyzer lenses from the resin-adhesive spray mist during the vacuum collection process. The visible droplets clinging to the surface of the lenses was noticed only when measurements stopped and the three components of the analyzer system were disassembled. This resin-adhesive contamination required frequent cleaning of the lenses in the spray collection cell after each experimental measurement.
This contamination also may be contributed to the relatively small space of the spray collection cell through which the resin-adhesive spray was vacuumed. Therefore, further work is needed to optimize the time-history analysis of resin-adhesive spray characteristics on the pilot-scale experimental set-up. A possible solution would be to use a wind tunnel technique, which is widely utilized in the application of pesticides and agricultural sprays (Teske et al. 2005).
In this study, a laser diffraction analyzer was used to characterize the droplet size and size distribution of resin-adhesive spray generated by a spinning disk atomizer. The real-time history of spray droplet characteristics was recorded, which showed the variation over a period of time as well as aiding in understanding online spray dynamics and recognizing the appearance of coarse droplets at the start-up of the atomization process. The various ways in which the spray data characteristics can be presented provide greater convenience for statistical analysis of the spray characterization. The distributions may be further fitted with mathematical models to analyze the component distributions of droplets separately. Overall, laser diffraction analysis appears to be a promising laboratory technique for better characterization of resin-adhesive sprays generated in the blending process. The pilot experimental set-up, however, needs refinement for more accurate and longer periods of data collection.
[FIGURE 4 OMITTED]
The authors acknowledge W. W. Henry Company and the Maine Agricultural and Forestry Experiment Station McIntire-Stennis project ME09615-03, Wood Based Composite Center Fellowship for financial support, and also thank Dynea Oy for providing PF resin-adhesive used this study.
Burrows, C.H. 1961. Some factors affecting resin efficiency in flake board. Forest Prod. J. 11(1):27-33.
Carroll, M. and D.T. McVey. 1962. An analysis of resin efficiency in particleboard. Forest Prod. J. 12(7):305-310.
Kamke, F.A., C.A. Lenth, and H.G. Saunders. 1996. Measurement of resin and wax distribution on wood flakes. Forest Prod. J. 46(6): 63-68.
Kawaguchi, T., Y. Akasaka, and M. Maeda. 2002. Size measurements of droplets and bubbles by advanced interferometric laser imaging technique. Meas. Sci. Technol.13(3):308-316.
Lehmann, W.F. 1965. Improved particleboard through better resin efficiency. Forest Prod. J. 15(4): 155-161.
Loxton, C., A. Thumm, W.J. Grigsby, A.T. Adams, and R.M. Ede. 2003. Resin distribution in medium density fiberboard. Quantification of UF resin distribution on blowline- and dry-blended MDF fiber and panels. Wood and Fiber Sci. 35(3):370-380.
Marian, J.E. 1958. Adhesive and adhesion problems in particleboard production. Forest Prod. J. 8(6):172-176.
Mitschke, M., T. Wriedt, and K. Bauckhage. 1998. Standard PDA for measuring the size of inhomogeneous droplets. Meas. Sci. Technol.9(2): 197-209.
Park, S., H. Cho, I. Yoon, and K. Min. 2002. Measurement of droplet size distribution of gasoline direct injection spray by droplet generator and planar image technique. Meas. Sci. Technol.13(6):859-864.
Schick, R.J. 2007. Spray technology reference guide: Understanding drop size. Bulletin No. 459B. http://service.spray.com/lit/view_lit.asp?code=B459B. Spraying Systems Co., Spray Analysis and Res. Services, Wheaton, Illinois. 35 pp.
Smith, G.D. 2003. A laboratory technique for coating strands with resin droplets of controlled size and spacing. Forest Prod. J. 53(7/8): 70-76.
Teske, M.E., H.W. Thistle, A.J. Hewitt, R.W. Dexter, and J.H. Ghent. 2005. Rotary atomizer drop size distribution database. Trans. ASAE 48(3):917-921.
Wilson, J.B. and R.L. Krahmer. 1976. Particleboard--microscopic observations of resin distribution and board fracture. Forest Prod. J. 26(11):42-45.
Xuelian Zhang *
Lech Muszynski *
Douglas J. Gardner *
The authors are, respectively, Graduate Research Assistant, Advanced Engineered Wood Composites Center, School of Forest Resources, Univ. of Maine, Orono, Maine (Xuelian_Zhang@umit.maine.edu); Assistant Professor, Wood Sci. and Engineering, Oregon State Univ., Corvallis, Oregon (Lech.Muszynski@oregonstate.edu); and Professor, Advanced Engineered Wood Composites Center, School of Forest Resources, The Univ. of Maine, Orono, Maine (Doug_Gardner@umenfa.maine.edu). This is paper 2974 of the Maine Agricultural and Forestry Experiment Station. This paper was received for publication in October 2007. Article No. 10416.
* Forest Products Society Member.
Table 1.--Detailed cumulative data collection of droplet size and size distribution of resin-adhesive spray. [D.sub.upper] [D.sub.tower] Volume Cumulative frequency volume ([micro]m) (%) 1.313 0.250 0 0 1.508 (a) 1.313 0.01 0 1.732 1.508 0.01 0 1.988 1.732 0.02 0 2.283 1.988 0.05 0 2.621 2.283 0.1 0 3.010 2.621 0.22 0 3.456 3.010 0.57 1 3.969 3.456 1.35 2 4.557 3.969 2.65 5 5.232 4.557 4 9 6.008 5.232 4.85 14 6.899 6.008 5.13 19 7.921 6.899 5.29 24 9.096 7.921 5.84 30 10.444 9.096 7.06 37 11.992 10.444 9.11 46 13.770 11.992 11.54 58 15.811 13.770 12.84 71 18.155 15.811 11.79 82 20.847 18.155 8.43 91 23.937 20.847 4.47 95 27.486 23.937 1.61 97 31.560 27.486 0.37 97 36.239 31.560 0.06 97 41.611 36.239 0.01 97 47.780 41.611 0 97 54.863 47.780 0 97 62.996 54.863 0 97 72.335 62.996 0.06 97 83.058 72.335 0.47 98 95.371 83.058 1.1 99 109.509 95.371 0.98 100 1000.00 109.509 0 100 (a) The smallest droplets detected were 1.508 to 1.313 [micro]m. Table 1 also indicates that about 90 percent of the spray volume was composed of droplets smaller than 20 [micro]m. Table 2.--Average droplet size distribution of resin-adhesive spray listed in standard statistical terms. Variable title Avg. SD Max Min Dv10 ([micro]m) 5.39 0.15 5.53 4.74 Dv50 ([micro]m) 12.58 0.37 14.05 11.98 Dv90 ([micro]m) 25.28 16.51 93.62 18.84 Transmission (%) 94.72 1.29 97.58 93.06 D[4,3] ([micro]m) 14.53 3.21 27.14 11.94 Table 3.--Parameters of the fitted models (Eq. (1]), non- parametric characteristics, and volume fractions of the droplet distribution components. Characteristics (b) Model of the component Component parameters (a) distributions ([micro]m) Component 1 [a.sub.1] 0.0574 D1 4.82 [m.sub.1] 2.08 D1 8.02 [s.sub.1] 0.397 DI 13.35 Component 2 [a.sub.2] 0.0757 D2 11.7 [m.sub.2] 2.81 D2 16.6 [s.sub.2] 0.273 D2 23.6 Component 3 [a.sub.3] 0.006 D3 83.5 [m.sub.3] 4.56 D3 95.1 [s.sub.3] 0.102 D3 108.4 Volume shares of component Component distributions (%) Component 1 [a.sub.1] 41 [m.sub.1] [s.sub.1] Component 2 [a.sub.2] 55 [m.sub.2] [s.sub.2] Component 3 [a.sub.3] 4 [m.sub.3] [s.sub.3] (a)[m.sub.i]; and [s.sub.i]; are mean and SDs, respectively, of normal distributions related to logarithm of droplet diameter D expressed in microns, and [a.sub.i]; are the scale parameters of the normal distributions. (b)droplet sizes at 10th, 50th, and 90th percentile of the cumulative distributions in microns.
|Printer friendly Cite/link Email Feedback|
|Author:||Zhang, Xuelian; Muszynski, Lech; Gardner, Douglas J.|
|Publication:||Forest Products Journal|
|Date:||Jan 1, 2009|
|Previous Article:||Electrochemical impedance spectroscopy (EIS) as a tool for measuring corrosion of polymer-coated fasteners used in treated wood.|
|Next Article:||Wood-borne formaldehyde varying with species, wood grade, and cambial age.|