THE PRESS CYCLE AS A PROCESS VARIABLE EN FLARE. EOAE’ED {MQUFACTURMG AND [W EFFECT ON BCARD PROPEMIES Thesis {or the Degree 0* M. 5. MICHIGAN STATE UNIVERSITY James D. McNatt 1961 Irma-:3; L I B R A R Y Michigan State University Th3 PRESS CICLE :3 A PHJCESS VARIAJLE IN FLaKE BOaRD daNUFACTURIIG adD ITS EFEECT ON BOARD PROPERTIES BY James D. TcNatt The problem studied here was the effect of certain raw-material and process variables on the prOperties of a homogeneous wood flake board. Sample boards were made in which the variables were moisture distribution, initial pressure, and overall board density. The study emphasized the effects of these variables on the distribution of the density through the thickness of the finished boards; and in turn, the effect of the density distribution on the elastic properties of the board, especially the modulus of elasticity. also included was a study of the temperature changes at different points in the board during pressing. This was used to explain the drying process of the boards in the heated press. It was found that the three variables had a considerable influence on the density distribution in the finished board and this in turn had an influence on the modulus of elasticity of the boards. The density of the surface layers of the boards was increased by adding moisture to the surface of the mat before pressing, by increasing the initial pres- sure; and the relative density of the face layers was increased by decreasing the overall board density. THE HESS CYCLE as A Pi-iOCESS Vi-JiLiBLE IN FLDKE BOARD HALUFnCTURIRG AND ITS EFFECT ON BOARD PROPERTIES By James D. NcNatt A THESIS Submitted to Michigan State University In partial fulfillment of the renuirements for the degree of IhfiTER OF SCIENCE Department of Forest Products 1961 AC ‘rllCOn'LlE-DGgi. El" 3 The author wishes to express his sincere aporeciation to V Dr. Otto auchsland and Killian '. Cobler: To Dr. Suchsland for defining the problem, for supervising the manufacture of the boards and the testing procedures, and for his criticism offered during the writing of this thesis, and to William E. Cobler for his assistance in the manufacture of the boards and in conducting the tests. Ho Ho ,‘- .9... - TALLE OF CONTENTS 1“ F“ fielcnOIvglledgernents o O O O O O 0 O O O O O O O O O O O O O O O O O ‘l—lo List of Figures . . . . . . . . . . . . . . . . . . . . . . . ii List 01 Tables . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . The Press Cycle. . . . . . . . . . . . . . . . . . . . . . . . 3 Review of Literature . . . . . . . . . . . . . . . . . . . . . 7 The Experiment . . . . . . . . . . . . . . . . . . . . . . . . 16 -. t 9 D ‘ T‘VV. , .‘n ‘ , I. 331.311 0.1. t]: G 4.)). 3‘3"}.‘911‘t a o o o o o o o o o o o o o o o o 10 '3 r. ,vJ-w' 5‘ ‘. ,- _,1. ’ F'1"Ci,’c..]f‘_"vo_1.CII OJ til‘T} lilCtr-eS o o o o o o o o o o o o o u o 0 lb 1..“ 7‘1; v. — .. 1 1 TITO “LETHLOCOJDIBb o o o o o o o o o o o o o o o o o o o o 1.6 hQC it .I_(.) .I Of QLITfQCe n LII—261‘ o o o o o o o o o o o o o o o o 19 Formation of the Boards . . . . . . . . . . . . . . . . . 19 The Testing Procedures . . . . . . . . . . . . . . . . . . . . 21 Cutting the Specimens . . . . . . . . . . . . . . . . . . 21 Tests Conducted . . . . . . . . . . . . . . . . . . . . 21 Results 0f the Experiment . . . . . . . . . . . . . . . . . . 24 DiSCuSSiOII Of tIie IieSIfl—ts. . O O I O O I O O O O O O O O O O O 28 o N Density ‘istribution . . . 1. The effects of moisture distribution . . . . . . 29 2. The effects of initial pressure. . . . . . . . . 29 3. The effects of overall densi Bendinf Strength of the anrds . . . . . . . . . . . . . 3~ _‘ L) The Drying Proces . . . . . . . . . . . . . . . . . . . ‘ .‘ I a+ uIIIaI'ZY o o o o o o O 0 o o o o o o o o o o o 0 o o o o o o o 3 9 Lin-flit {it ik: IIS 0 f I; 138 StUCIy o o o o o o o o o o o o o o o o o o [$0 Liter:3-I:]lre 01:.th O O O O O C D O O C I O O C I C C C C O O O . 41 11. 12. 13. 14. LIST OF FIGUEES "Press CEICIGH for E-II'C‘- TJEI‘Ieer StC‘.CkS o o o o o a o o o o o o o o o o o o "Press Cycle" for Brita Vera Veneer stacks . . . . . . . . . . . . . Effect of Teaperature and Hoisture Content on the ‘C? if Effect of hnw- ateri l and Process Jarisxles on the Todulus of Llasticity of a Three-L yer Flake Board.. LocatiJH Of TilerrTIOCOI-lpleSOOOOOOO......OOOOOOOO...... fethod of Cutting Specimens from Finished Boards.... Density Distribution in Top Half of Board Number 22 (0.005-11301’1 interva]-S)OOOOOOOO0............OOOOOOOOO Density Distribution in Top Half of Board No. 22 (0.015-inCI’I interVaIS).............................. Density Distribution as a Function of Initial Pressure and ioisture Distribution.............................. Density Distribution as a Function of Initial Pressure and Overall Board Density.............................. Effect of Density Distribution on the Hodulus of BlgkstiCity Of tl’Ie BoardSOCOOOOOOOOOOO......00.0.0000... The Effect of Temperature and Toisture Content on the Preportional Limit of American Beech - with additions.. Time-Temperature Curves for tie three Thermocouples in Board. :Iwn.eer 22...........OOCOOIOOO......OOOCOOOOOCOOOO H° Ho Ho Oportional Liiit of nnerican Beech................... f ”oisture Ccntcnt Increase in the Face Haterial ensitv of a Three-Layer Flake Board............ [.1 \h L4 01 T 51131 e "l ”e LL ~0— #— 1. Design of the Experiment.............................l7 2. Results of the Experiment 9 ...-oooooooooo-oooooooooooaofiv iv EITH‘JL' CT I01“? The manufacture of wood particle board is essentially a laminating process in which a batch of small wood particles in the form of flakes, chips, or splinters is mixed with a binder and com- pressed into a thin sheet in a heated press. Flake bOards, one type of particle board, is composed of small, thin, rectsnsular flakes of wood (8). There are two tyoes of variables involved in the manufacture of wood flake board which affect the properties of the finished board: 1. the raw-material variables and 2. the variables of the press cycle. The raw~material variables are flake species, flake geometry, flake moisture content, and glue content (9). The variation in properties would obviously affect the density, compressibility, etc. of flakes produced from the various species. By flake geometry is meant the size and shape of the individual flakes. Host flakes are produced with a thickness within the range of 0.005 inch to 0.015 inch, a width of 1/8 to 3/8 inch, and a length of 1/2 to 1 inch. Flakes are usually cut by means of planer—type knives set into the face of a circular disc. The flake length is controlled by scoring knives projecting from the disc or by precutting the boards to length. The flake thickness is controlled by adjusting the projection of the knives. The width of the flake is the most variable of its dimen— sions. It is random and depends on the type of cutting machine used. Further reduction by hammernilline is generally necessary. 3.; The flake moisture content is usually kept within the range of 6 to 12 per cent. Sometimes, however, the moisture content of the surface-layer flakes is increased to as high as 30 per cent in order to attain desired results in the progerties of t’e finished beard. r"he adhesive used in the manufacture of wood flake anrd is a thermosetting, synthetic resin; for example, urea formaldehyde or phenol formaldehyde. The amount of adhesive added is usually betdeen 6 and 12 per cent based on the dry weight of the flakes. a water solution of the adhesive containing about 50 to o0 per cent solids is sprayed onto the flakes in some type of rotating mixer. The variables of the press cycle are press temoeriture, initial pressure, and press time (9). Conrercial Operations use multiple-onening hydraulic hot presses wtich have high heat capacities. Press tempera- tures in the range of 300 to 3500 F. are necessary for adequate curing of the glue. The oress temperature is kept constant during the pressing time. The initial pressure is that which is applied to the mat of flakes in order to compress it to some predetermined thickness. This thickness is controlled by steel steps which are placed in the press. The time it takes for the press to reach these steps is called the press closing time. The initial pressure may vary from 100 psi in some operations to as high as 500 psi in others. The_purpose of the pressure is to bring about the largest possible contact area between the flakes. The length of the press cycle is defined as the total time that the board is in the press. Usually 8 to 12 minutes are necessary to insure an adequate curing of the glue. The press cycle is discussed in detail below. TEE PRESS CYCEE The relationships between the variables which make up the press cycle are demonstrat€d by a test involving the use of a stack of veneers in place of a mat of flakes. a stack of veneers 0.60 inch in thickness was placed between heated platens wlich were mounted in a testing machine. The temperature of the platens was 3000 F. A predetermined load was applied and held constant until the veneer stack had been compressed to a thickness of 0.50 inch. The machine was then shut off for the remainder of 10 minutes. The changes in pressure and veneer thickness with time were recorded for the 10- minute "press cycle". Figure l and Figure 2 show the result of the test on birch ven- eers and prime vera veneers respectively. The stack of 0.035-inch birch veneers at an initial moisture content of 13.8 per cent was a compressed from a density of 0.665 gm/emJ to a density of 0.790 gm/cm3 The stack of 0.037-inch prima vera veneers at an initial moisture con- tent of 10.3 per cent was compressed from a density of 0.500 gm/cm3 to a density of 0.600 gm/cmB. Figure 1 shows that the higher initial pressure results in a shorter "closing time" and a greater drop off in pressure afterwards. Figure 2 shows that if the initial pressure is not sufficiently high, the veneers cannot be compressed to the desired thickness even though the pressure is maintained for a considerable length of time. A comparison of Figures 1 and 2 would suggest that moisture con- tent has some effect on the compressibility of the veneers. The birch veneers had an initial moisture content 3.5 per cent higher than the 'Ti H. l n are 1. "Press Cycle" for Birch Veneer StaCkS Veneer thickness 0.035 inch Initial moisture content 13.8% m/cm3 Initial density 0.665 C") Final density Z. .l C; wwmzxoik 0 O O O O O 0 w 6 4 2 O 6. . 5. 5 5. 5. 0 O O O 0 O 0 lat: iii: i 1 Jr P . \Ih.bl\l& \\ 600 400 SOOr—erfi . 1' II it w I l L A: 300‘L \ .wa I. an: memmwmn. PRESSTIME - MIN. Figure 2. "Press Cycle" for Primn Vera Veneer Stacks Veneer thickness 0.037 inch Initial moisture content 10.3% Initial density 0.500 gn/cm3 Final density 0.600 gm/cm3 600 Z. I CL mwmzxof... O O O O O O O 8 6 4 2 O 6 5 6 5 5 5. O O O O O O O O 200 I00 .ma I An: mmDmmwma IO PRESSTIME — MIN. prime vera veneers. at the 50 psi level, the "closing tine" f the birch veneers was 45 secones siorter than that of the prime Vera, in Spite of the higher density #%d congressiVe strength of birch. Expressing the resu ts of this study in terms of the manufacture r r‘ 1.. ‘1. (‘r‘ ‘Y.' c. L ",r r '1' 1 . 4"" i" 1‘ r '1 P. = ‘. f flake heard, the press cycle can he S€udrob€Q into two shits. The A .‘ first part of the cycle is the period during which tle flake met is compressed to a desired thickness. The period of time covered by this tie closing time of the press. The remainder U) pert of the press cycle i of the gress cycle comprises the second part. During the first part, the initial pressure is built up and rennins constund until the press closes. The deflection of the sat during this time is a process involv- ing heat and moisture transfer inward from the mat surface. This move— ment of heat and moisture causes a decrease in the compressive strength of successive layers of flakes. The consequence of this successive wea- kening of layers is an uneven density distribution through the thickness of the finished board. The highest density will occur at the point in 'the board thickness where the effect of heat and moisture is greatest. True density distribution in the finished heard is, therefore, determined dUJtinc the first part of the press cycle while the full initial pressure is zacting on the mat. During the second part of the press cycle, the pressurre on the heard is indete*minate and us ally decreases as a func- tion of‘time. REVIEJ 0F LITERnTURE Only one article was found in which the author actually plotted curves of the density distribution through the thickness of the board (7). Considerable work, however, has been done in the area of raw- material and press cycle variables and their effect on the preperties of flake board. also useful information was found in a study of the effect of heat and moisture on the strength of solid wood (2). H. D. Strickler found in a study of Douglas-fir flake board that a density distribution is the direct result of varying the press cycle and moisture content. He found that by increasing the initial pressure, the density of the surface layers was increased and the center density was decreased. an increase in the moisture content of the surface lay- ers had the same effect if the initial pressure was high but for lower pressures the point of maximum density occurred at a point about 1/1 the distance from the surface of the board to the center. When the ini- tial moisture content of the flakes was low and the boards were produced with a low overall density, the density distribution tended to "even out". The mechanical properties of the board illustrate the effect of {Iressure and moisture on the density distribution. The modulus of elas- ‘bicity generally increased with an increase in initial moisture content arxi initial pressure. The effect of moisture distribution is shown by tine fact that the modulus of elasticity increased with an increase in snxrface-layer moisture content up to about 21 per cent moisture content, 'then decreased. Too much moisture adversely affects the gluing process, and therefore decreases the density of the surface layers. Strickler used thermocouples placed in the mats to study the temperature increase during the pressing. Initial pressure, moisture, and moisture distribution all influenced temperature increase. Heat- ing up of the center of the board was accelerated by increasing the initial pressure, increasing the initial overall moisture content, and increasing the surface moisture content. Increasing the overall den- sity of the boards did not have any significant influence on tempera- ture increase. The press cycles used in the study were somewhat artificial since the boards were not pressed to stoos. Therefore, the effect of the closing time could not be evaluated because there was no determinable closing time. Heat transfer was the subject of a study conducted by T. F. 1 Duncan (1). He studied the effects of neat transfer on the properties of 3/h—inch particle board in which he varied the particle geometry, overall board density, press temperature and resin type. He found that the board density did affect the rate of heat transfer. he noted that the center-line temperature of the denser boards was lower than that of the less-dense boards at the same platen temperature. This was attributed to the fact that although heat conductivity is directly related to density, the amount of heat recuired to increase the tem- txzrature of the denser board is greater than that required for a low— density board. He indicated that the leveling off of the time-tempera— ‘ture curves was probably due to either moisture escaping from the board or a temperature gradient being set up in which the flakes lost heat as fast as they absorbed it. Another study on heat transfer was conducted by G. Rackwitz (C). A temperature increase in the board above 1000 C. can best occur if all the moisture in the board is converted to steam. Then, according to Rackwitz, this temperature increase follows the laws of equili- brium moisture content of wood in superheated steam. The center of 1 the board alaays has the lowest temperature and highest moisture con— tent. As the steam pressure builds up in the center of tie board, the temperature increases. Therefore the temperature gradient between the center of the board and the surface decreases, resulting in an in- creased heating time. The heating time necessary to cure the adhesive is expressed by the following formula (6): C, , 2a 1-75 t — n.\l.o7la - 9.3)(55 in which t = heating time in minutes K1 u a temperature factor depending upon the platen temperature (K decreases with increased platen temperature) Ua= initial flake moisture content in per cent 2d= board thickness in mm according to tils formula, the press cycle can be s ortened by de- creasin33the initial moisture content of the flakes or by increasing the platen tcmperature. Shortening the press cycle would increase produc- tion in a connercial Operatihn, bit either of the above-mentioned methods 'would have adverse effects on tie finished board quality. A better me— thod of increasing production would be to add more openings to the press. R. Keylwerth (5) analyzed the imoortance of moisture content on the density distribution in one- and three-lever flake boards. In the : II \.0 production of flake board, the extent to wiicn wood is elasticized depends upon the temperature of the wood, the wood moisture content, and the pressure. This "softening" of the wood is important in flake board because it affects the amount of contact area achieved when the H° mat of flakes s pressed. If it is assumed that wood does not have a pronounced yield point, then the proportional limit can be considered to be the starting point of plastic deformation. E. L. Ellwood (2) studied the effect of temperature on the mechani— cal properties of beech in compression perpendicular to the grain. Keylwerth used the results of this study to Show the relationship be— tween the prOportional limit stress (d’p), the wood moisture content (m), and the wood temperature (T). This relationship is expressed by the formula: (‘9 = '70 log 9—3-9; - 0.45/3 (kg./Cm2) Figure 3, taken from Keylwerth's article, is the graphic representation of this formula. This graph shows that by increasing the moisture content and temperature, the proportional limit of the wood will be lowered. Figure 4, also taken from Keylwerth's article, shows the effect of the moisture content of the flakes in the surface layer on the density of the surface layer, the density of the center of the board, and the overall density of the board. The values used in this graph were taken from a study by F. Fahrni (3) on the effects of high moisture contents in the surface layers. Figure 4 shows that by increasing the moisture content of the sur- face layers, the density of the surface layers is increased without in- creasing the overall density of the board. The graph also indicates lO that there is a limit to the amount of water that can be added to the surface layers and still increase their density. 0. Suchsland (8) studied the effect of raw-material and process variables on the properties of a two-species, three-later flake board. The core consisted of elm flakes and the faces were composed of aspen flakes. The thin, narrow, high-density elm flakes were less compressi- ble than the thin, wide, low-density aspen flakes. This resulted in a sandwich-type construction in which the core density was lower than the face density. Figure 5 shows the effect of the press closing time and the face material moisture content on the modulus of elasticity of tiis board. The two closing times of 0.5 min. and 1.5 min. were obtained by using initial pressures of 500 psi and 250 psi respectively. The graph is plotted in terms of shellinq ratio, core density, and face density. These had to be determined before the effect of the raw-material and process variables on the finished board properties could be analyzed. The shelling ratio (ii) is defined as follows: = thicknc s of the two faces it total thickness of the board From Figure 5 it can be seen that the modulus of elasticity is in- creased by decreasing the closing time or increasing the moisture content of the face material. This increase in modulus of elasticity is due pri- marily to the increase in face density. At the same time, the core den- sity decreases since the overall density of the board remained constant at about 0.700 gm/cmB. The modulus of elasticity of a sandwich construc- tion is dependent almost entirely on the stiffness of the face material. ll 000439 OdOJd mpg) t1 t'tOIzz t: A m v 3330 8.330: coo: as om 0H NH m o o: 11.1.11 I I A o / m. on ou 111111111111 111111111 n / u. 0 1, In em \1 11 11 I / / / / 00 111 1/ 11/ ooa 1\ 'f 1 l/// / _/ //V ONH /. 0 ocean cuuo macs». a. .o doc: ho» a so u«I«A\\ ova an. «Hosea: ..npuoydhcu noun eoosegugpu ...uoop naoauon< no canua annouuuomoun one in advance cusp-«cl can ouuoauoaloa mo vacuum n unshah 3.: an aqua-goo hand." 033- mo 6.5 A3 antenna coon no anon-30 0.2533. n: 3 on on mm 8 ma i .33 .35\ n tell _ _. .. III 1‘ 938 £25\ «9 «flash up .31» 43 ...—03.3 nit-sauna no.6 c.9389... goon-Mac “93.708: a no 5333 on» no 3.0 m6 v.0 5.0 w.o dun-cal noun a: 5 0:93: advance 0.25:9. mo no.3! 3 Pan: (Clo/fl) £3 “qu 13 On the basis of the material discussed in the Introduction and Review of Literature, an experiment was designed to study the effect of the press cycle as a variable on the density distribution and other properties of a homogeneous (one—layer) flake board. 14 figure 5 Effect of run material and process variable. on the modulus of olouticity of o 2 - opooioo 3 - layer flake board Taco density ( {I/cn3 ) 6? /°. 0-,6 N 29 .. 7/ 2/ 7 ‘ 800 A 3 w "‘5 // \ 622’ It ,STH 1////# V ' / / i? (/r /’//// E ' o .3 & er a I /// 700 £3 éf 00 / ’/ 0'61 600 0.4 Sholling ratio <3 0.5 lin. cloning tine Gore/Face weight ratio . 1.5 min. closing time ® 1.50 ----- 6f face X.C. 2.13 ‘_——" 12$ face Moo. 3.16 Graph reproduced from Suchaland'o article (8) 15 ( 1000 psi ) Modulus of elasticity EXE‘ER--ZEEI‘TI‘ *3 so. 51 Design of the Experiment after a study of the results of sane preliminary tests, it was decided to Lake the flake boards at two overall density levels using three different initial pressures. Surface water was added to half the boards made at the lower density level and to all the boards made at the higher density level as shown in Table 1. Pre ”Matias-afjhsfilsl19.55. Aspen flakes were used which were cut with a nominal thickness of 0.910 inch. The flake length was 1/2 inch and the flake width before reducing was 1 inch. The flakes were reduced in width by hammer— mill ng through a B/L-inch screen. They were brought to a uniform moisture content of about 6 per cent in a controlled humidity cabinet. Tine Thermocouples Thermocouples were made from 8-foot lengths of BO-gauge copper wire anti constantan wire. One end of a length of cooper wire was wrapped to- getlaer with one end of a length of constantan wire and the connection was 'then soldered. The length of this soldered connection was approxi- mateily'l/Q-inch. This end of the thermocouple was placed in the mat to measuire the temperature increase. Three thermocouples were placed in eachllnat as it was formed: One at the center of the mat thickness, one nfiihnay'hetween the center of the mat and the surface, and one on the top surnface of‘the mat directly under the top caul as shown in Figure 6. i¥fter the boards were removed from the press the wires leading into them 'were cut off and reused by soldering together the ends as before. New 'thermocouples were prepared when the wires becane too short. 16 nu an «N 3 S S S 8 o 3 S a hoe...- 28A ooh on a.» no» on no» so» on no» no» on a.» v.3! you: gamut on: 9.2 and RS on: 83 on: omen v33 you :3 be no ...-.5 8: 8m 8n 2: 3.3 93......" - ouafloeno on» no nude-n H 0549 17 H530 floovlon' U 3 .353: an: \kj we . no»: be go .o .bv \“MvwvmV\‘ ." k\\w non«1w\n i- w on. o \\\\ .i \ \H \ \ W J seam... \\ .\ \x \ \ u. \ \ 8 1 noflmsoooaquu no noaudooa w ousuuh Additiun of Surface UthE Surface water was added to the various mats by spraying a very fine mist onto the cauls from a small cents ner attached to a spray gun. The water was sprayed onto the tone erf ce of the bottom call before the mat was f:r.ed. "he sane anount we 5 snrayed onto the un- derside of the top caul before it was placed over the mat. Fifteen gra has of water were sprayed onto each caul, but only about half of it n'as retained due to eVanoration. as; i..-engaging;fissile The adhesive used was urea formaldehyde JE-l? diluted to 50 per cent solids content for eg sie rsoraying and better coverage of the flakes. The amount of glue sprayed onto the flakes in a rotary drum mixer was 8 nor cent glue solids, based on the dry weight of the flalw Formation of t-e_§9h gs The moisture content of the flakes was recorded before the adhe- sive was added, after the adhesive was added, and jus t before the mats were placed in the press. In add .tion, the tote l eigl nt of the mat before pressing was recorded. The mats, held between the two aluminum earls, were pressed to 3/3~inch steel stops in a 20 x 20-inch, single-opening hot press. In each case the press clos Ing time was recorded. The press platen ten- perature was 3250 F., and the press cycle was 10 minutes long The initial pressure was 100, 200, 300 and 400 psi. .A continuous time-temperature curve was obtained for each of the tkqema'thermocouoles in the boards. They were connected to an automa- tic.X-Y recorder throurd: a device used to switch fre m one ther ocouole l9 to another. Each of the boards was removed fr m the pres after the 10 minutes, and immediately moisture content samples were taken from the edge and center. 2O TIE TESTING PRWCEDURES Cuttine the Soecinens The following specimens were cut from each board: 2 for moisture content (mentioned above) 1 for density distribution 1 for thermocouple location 7 or 8 for bending tests The method of cutting these is shown in Figure 7. Tests Conducted The moisture content of each board immediately after pressing was determined as described in Formation of the boards. The exact location of the thermocouples in each board was deter- mined by first measuring the thickness of the specimen, then planing it down from the top surface to the thermocouples. The thickness of the remaining part of the specimen was measured at each of the thermo- couple locations. The locations were expressed as a per cent of the board thickness, measured in from the tOp surface of the board. Theo- retically they should be 03, 255, and 50% in all the boards (see Figure 6). The modulus of elasticity for each of the bending specimens was determined according to the ASTM standards for evaluating the proper- ties of fiberboard (10) except that the specimens were one inch wide instead of 3 inches. The tests were done on a Baldwin-Emery SR-A testing machine. Before testing, the specimens were conditioned at 680.F. and 50 per cent relative humidity to bring them to a uniform meisture content of about 8 per cent. 21 Figure 7 Method of cutting specimens from finished boarde 1 Bending specimen 2 Bending specimen 3 Bending specimen h Bending epeeinen M.C. u.c. Thermocouple location (:> Density distribution <:) 6) ® 5 Bending specimen 6 ' Bending specimen 7 Bending specimen 8 Bending specimen The density distribution in the tOp half of each board was deter- mined as follows: 1. The thickness of the density distribution specimen was measured to the nearest 0.001 inch at 6 locations which were marked on the bottom surface as shown in Figure 7. 2. The length and width of the specimens was measured to the nearest 0.01 inch. 3. The specimen was weighed to the nearest 0.01 gram. 4. A nominal thickness of 0.005 inch was taken off the top surface with a table jointer and the specimens were again weighed and the thickness measured at the same six locations . Step 4 was repeated until just over half of the specimen was removed. From the data thus obtained, the overall density of each specimen could be determined as well as the density of each layer removed by the jointer. 23 RESULTS OF TIE 1331131331333;- The results of the experiment are shown in Table 2. Using the data obtained from the density distribution specimens, a graph was constructed for each board showing the density of each of the 0.005- inch layers removed from the tOp half of the specimen. For conven- ience of comparison, the relative density was plotted instead of the actual density. The base density, represented by 1.000 on the graphs, is the overall density of the top half of the specimen. Some of the irregularities in the densities of adjacent layers is due to errors in measuring the thickness of the layers. The accuracy of the measurements (0.001 inch) was not sufficient for the thickness being measured (0.005 inch). The effect of this error was reduced by constructing a graph in which two layers were combined to form one and the density computed accordingly. The graph for board number 22, used as an example, is shown in Figure 8. The effect of the error was fur- ther reduced by constructing another graph in which three layers were combined to form one. This graph for board number 22 is shown in Figure 9. The modulus of elasticity for each board was taken to be the aver- age of the 7 or 8 bending specimens. This modulus of elasticity was adjusted to the target density of the board by the use of regression lines. The "least-square" method was used to determine the equation of the regression lines in which board density was the independent variable and modulus of elasticity was the dependent variable. 24 Table 2 Results of the experiment irate-ma (p81) 100 200 300 000 Gram of dry flakes per board 1050 1150 1050 1150 1050 1150 1050 1150 Surface water added you no yes yes no yes yes no yes yes no yes $111 number 1 19 16 6 20 17 11 21 18 22 21+ 23 10: of flakes before mixing with glue (7!?) 0.5 6.2 6.2 5.5 6.2 6.2 0.5 5.6 7.2 6.0 5.5 6.1+ £0! flakes after mixing with glue (‘5) 12.7 10.4 13.6 13.6 114.5 10.6 11.0 13.“ 11+.“ 14.6 13.9 110.7 lief flakes Just before pressinc (5) 12.0 13.3 12.8 13.1 13.2 12.9 10.2 12.0 13.3 13.6 12.9 13.6 & of board immediately after pressing W) 3.0 0.0 0.0 2.6 3.3 3.1 2.5 1.8 2.1+ 2.6 2.3 3.0 I M closing time (aim) 1.50 n n 0.60 1.00 0.80 0.55 0.55 0.50 0.00. 0.25 0.30 inter-line temperature of board at closing time (°Zl') 223 --- -- 115 198 18“ 135 106 122 --- 85 88 Aofllu; off“ point of center-line temperature (°r) 230 220 231 200 235 251 232 200 262 208 205 260 £et which center-line temp. reached 220°? (31.11.) 1.100 2.90 1.90 1.10 1.1+0 1.10 0.90 1.15 1.00 0.80 1.10 1.10 Much... (inches) 0.360 .006 .397 .359 .362 .360 .350 .355 .351 .308 .350 .309 £156.00 density (gm/c113) 0.578 0.533 0.591 0.590 0.593 0.639 0.568 0.61? 0.666 0.620 0.625 0.681 w of elasticity — Adjusted to target density (1000 psi) 586 062 631 712 659 758 677 703 828 7M 645 78“ "'" Preu did not close 25 "' Press closed immediately -.‘J (m a canola v oneuuouau mafia 8m... cum... . com .6 Sad _ _ .... _ / a _ / > r36 _ / / > g <\ / 184 w lea.” A. 1.03.: as: o Sod v «N 03.3.. 2.3 no m n3 .5 .: Signifis 91.3 m can. f8; Ina; tune}: on: 1.1.3 26 A unload v woonxouna an.o oo~.o om~.o oom.o omn.o H .b _ ._ _ _ _ / / _ ///// ///// _ / /////w w _ _ _ A .1»...an :63 .. 30.0 V Nu genial cheep.ho m no» 01» ad aoauspauauau hagmhofl m ensuuh 00.0 om.o oo.H 0H.H 0N.H on.a 3.3.4 lztfllop 0Atzvtva 2? DISCUSSION OF Th.) RESULTS Density Distribution There are actually two types of density distributions in a flake board: 1. the density distribution across the area of a given layer in the plane of the board and 2. the densities of successive layers through the thickness of the board. The first is controlled by the flake geo- metry and the method used to form the mat. It is determined when the mat is formed and is not changed by the pressing operation. The second type, which will be the subject of this discussion, is not affected by the mat formation. The results of the experiment show hat the density distribution through the thickness of flake board is signifiCantly affected by the raw—material and press-cycle variables. Figures 10 and 11 illustrate the effect of three of the basic variables on this density distribution. Figure 10 illustrates the effect of surface water and initial pressure. Figure 11 also illustrates the effect of initial pressure together with the effects of the overall density of the board. These two figures to— gether include graphs for all 12 boards made and are simplified vers- ions of the type shown in Figures 8 and 9. Each interval is shown in the form of a column. The height of the column represents the density of the "layer", and the width represents its thickness. The thickness is 1/3 of the thickness of the upper half of the board. For reasons of comparison, the thickness is expressed as a per cent. The influences of the variables on the density distribution occur simultaneously; however, they will be discussed as if they occurred one at a time. 28 l. The effects of moisture distribution Increzsing the moisture content of the surface flakes results in an increase in the density of the layers near the surface and a cor— responding decrease in the density of the layers toward the center of the bOard. Figure 3 shows that by increasing th; moisture content of wood, the pro ortional limit is lowered, making it more compressible. If the surface flakes are easily compressed, a large contact area will result. As was stated earlier, the highest density will occur where the combined effect of heat and moisture is greatest. Because the flakes are made more compressible by the addition of the water the press closing time is reduced. his means the greatest combined effect of heat and moisture will occur near the surface of the board. The addition of water to the surface also accele~ates the heat transfer from the surface to the center of the board. The shortened closing time speeds up the heat conduction between flakes and the moisture in the form of steam carries heat inward from the surface. 2. The effects of initial pressure Jith increasing initial pressure, the density near the board sur- face will increase to a maximum an‘ then decrease. Excluding all other variables, the closing time varies directly as the initial pressure. fioisture and heat have time to move inward considerably before the press closes when the initial pressure is low. Consequently the com- pressibility of a large portion of the flakes is decreased. The re- sult is an almost uniform density distribution. If the pressure and/ or moisture content is too low, the density of the surface layers 29 __O V\ -'-‘7 no ooa F"-"’ L--___ sun co: a vadoA no m no» no “and duo-ounuo v coo-Moua» “chin on r on “e no co” on s n co Pr _ — _ F\ W _N _ H WIIJ l TIIJ . . . . . . . _ - . _ _ IIIJ _ _ Tl laud _ v _ _ . _ _ _ _ . _ . r. L _ u u . _ T «on com "I: IL. «on cow Wulb 13! so»: on .r ..IL. "--"1 ........1 5w _ no _ wed rtll «on 00A nodaapauauae unsouno- qua .na..oun Hades.“ no uoaoousu a n. noaaupaup.«e »»«.non "'L downs you.) cannula on ohuuuh no.0 omoo na.o coca no.n ann nu.a ON.H lat-toy solbt 0Atgutog will be lower than that of the center of the board. This is illustra- ted by board number 19 in Figure 10 and by beard number 16 in Figure ll. The combined effect of heat and moisture acts nearer and nearer the surface as the ClOSilg time is s ortened by increasing the pressure. Ex— tremely high pressures close the press almost in ediately, and the full effect of heat and moisture is not attained. The result is that the sur- face layers nill not reach the maximum density. In agreement with the tests on stacks of veneers, high initial pres- sures result in a rapid drop off in pressure after the press has reached 1 the stops. In certain cases the pressure on the bOard in the press Q dropped to zero and the board shrunk away from the platens. This 18 shown by the fact that the thickness of tie finished board was less than the thickness of the stops. Also the temperature of tie board surface dropped momentarily near the midpoint of the press cycle indicating that the thermocouple was not in direct contact with the caul (see Figure 14). The thickness of the finished boards decreased with an increase in initial pressure. Jhen high pressures are used, the mats are compressed to the stops before much if any moisture is lost. Loss in moisture af- ter the press is closed results in a shrinkage of the boards in thick— ness. The more moisture there is in the board when the press is closed, the greater will be the shrinkage. The press did not close when boards number 19 and 16 were made, but it did close when board number 1 was made, although the initial pressure was the same in all three cases. The lack of surface moisture kept the press from closing on board number 19, and the increase in the density of the overall board kept it from closing on board number 16. 31 As was expected, the hi§her initial pressures accelerated the heat transfer because of the shortened closing time (Table 2). 3. The effects of overall density Boards made with the 1 per overall density had a hi her relative density in the surface layers than the boards made with the higher over- all density. It is necessary to use "relative density" here rather than "actual density" because the boards were made at the two diff- erent levels of overall density. The lower density boards had a shorter closing time. The results were higher surface layer densities. The effect of the closing time on density distribution has already been discussed. Bending Strength of the_§gard§ Figure 12 illustrates the fact that the effect of the raw-material and press-cycle variables on the modulus of elasticity of the finished boards is a consequence of their effect on the density dist ibution. A comparison of Figures 10 and 11 with Figure 12 will verify this fact. Both the modulus of elasticity values and the densities of the surface layers increase as the initial pressure increases from 100 to 200 to 300 psi and then decreases as the initial pressure increases from 300 to 400 psi. The one exception in the trend of the modulus of elasti- city Values is board number 11 (see Figure 12). This is explained by the fact that the flakes from which this board was made had a moisture content which was 2 per cent lower than the other boards. This is equi- valent to more than 20 grams of water. \JJ I‘d A snoop no m nu» u« n a. coo-ounuo v unannouaa hound We mm 42 m .m. m» a: cm 5% mm m2 cm a mm b2 I 1'IIIJ WIIJ III]. IIIJ ” j. I . _ F FL [1. u _ T . . . _ «a 3.. «unconrluL We?» 36 had-nod undo» Stuck. ‘- ouinuoun 13:“ no 1030!,“ d an nodular—«.3 has-Ion 4m 1IIJ """l | I —--J l"-1 .....I‘IL I'll in oo~ «3 02 1-:4 .r--.“ as? 8... I36 18 .0 1a... Foo!" Inc!" Ion!" In.“ ION!" A." earn ‘ laptop Jot-t OA'H'I’! 32A 3-3 cud-noun .73“: 8: com com 2: -.l..|..l 13? use... on n37» 8.0 II I l 233 .259. a. «no?» 86 13- .2%: a n85.» 3.0 \G 0E. \ van. awn . unocuoana\ ndunop on» yo huaouaudao mo mdaduoa on» no noaaupauuuav huamnov mo vooumn NH ouluah co: 8“ coo con com com ( I“ 0001 ) lama-tn 3° tau-pox 33 It might be concluded fron the above discussion that the modulus of elasticity is a function of board density'alone. Strictly speaking this is not true because two boards cannot be made which vary in den— sity alone. By adjusting the material or process variables to produce a board'with different density, factors which affect the density dis- tribution and other board properties are also affected. The Drying Process The graoh in Figure 3 has already been used to illustrate the ef- fect of temperature and moisture on the compressibility of the flakes. It can also be used to xplain the drying of the flakes which takes place in the board. For this purpose the graph in Figure 3, taken from Keylwerth's article (5), is reproduced again in Figure 13 with the ad- dition of the equilibrium moisture content curves for wood in super- heated stean at pressures of 1.5 and 2.0 atmospheres. r1hese two curves were taken from an article by H. G. {auman. Figure 14 shows the time-temperature curves for the three thermo— couples in board number 22. The curve for the center-line temperature will be used as an illust‘ative example in the discussion of Figure 13. The center-line temperature for board number 22 after 1/4 minutes in the press was 120 C. After 4.2 minutes in the press it had drOpped to 114° C. At the time the board was removed from the press the center- line temperature had increased to 140.50 C. From Table 2 the moisture content of the flakes used to make this board was 13.6 per cent at the time the mat was put in the press. The moisture content of the board immediately after pressing was 2.6 per cent. 34 “my onovuoo oufivu«0l door as ON on wonn NH 0 a o u.u. m um 3 I I .. . . I I u I _ 1' a nu 1//+///1 o: .6“. ON . no . u / MO. I/ 8 O / \l o / . o I he." - ../ / I.\ colon com III, / ONH //R ..., .o.x coo: hon nosuuo aunun IIIII.n.o:H ova 3 303.3 3.3.2.5»: noun 1333.... A vacuums; Ava) v 10.2.. gravid no avid." nu ouuugh auuoaouonoam 91a no auoaaoo onuanuol and onsuduoniuu no vacuum {a ‘L-—-- ~' in ' Figure 14 Time - temperature curves for the thermocouples in board number 22 three o." nu. com I o .u .- ... m. k ‘ ( .2. 8a \ Fathom-o» :1 ... uoaloo/ / . .IIIIIAY‘ d l sli:i+:\d7 .. v 8:288...» 59:. X To 6.1 lb \ con ogudhomfloa OQIHH-fi\ There are three "limit curves" for wood moisture content shown in Figure 13; One for steam at atmospheric pressure, one for steam at a pressure of 1.5 atmospheres, and one for a steam pressure of 2.0 atmos- pheres. These curves mark the upper limit of the moisture content which wood can attain at any given temperature. For example, at a temperature of 1180 C., the moisture content of wood cannot increase above 8 per cent unless the steam pressure is increased above 1.5 atmosoheres. Figure 13 suggests three possibilities for the drying process which takes place in the press. First the board dries out at atmospheric pressure along some curve which does not reach the limit curve for atmospheric pressure until the temoerature at the center line has reached 1200 C. Then the curve foll- ows this curve up and to the left until the temperature reaches 140.50 c, at which time the press is opened and the board removed. There are two reasons why this is an unlikely possibility. In order for the drying to occur at atmospheric pressure, the board would have to dry out to 4.8 per cent moisture content from 13.6 per cent before the center-line tem- perature reached 1200 C. This temperature was reached in only 1.4 minutes in the example board, and a moisture content drop of almost 9 per cent is not feasible for this short length of time. Also, in order for the tem- perature to drop from 1200 C. to 1140 C., as it actually did, the moisture content would have to increase to 6.2 per cent after having initially drOpoed to 4.8 per cent. This, too, is not likely. The second possibility is tTat the temperature increases to 1200 C. with little or no change in moisture content, and then a moisture con- tent increase occurs to account for the drOp in temperature. The in- crease in temperature to 120° C. at a moisture content of aboat 13.6 37 per cent would result in a steam pressure inside the board of about 1.9 atmospheres or 13.2 psi above normal pressure. This steam pressure is quite possible. The graph, however, shows that at this pressure it would be impossible for a large enough increase in moisture content to occur which would account for the temperature drop of 60 C. The third possibility is probably the most correct. The tempera- ture of the center of the board increased to 1200 C. with little or no change in moisture content, resulting in a steam pressure of 1.9 atmos- pheres. Then as the board beg ns to dry out due to steam escaping f‘ through the edges of the board, the pressure drops to 1.25 atmospheres causing the drop in temperature. as more steam escapes, the tempera- ture begins to rise again because the moisture content has dropped. Eventually the pressure drops to atmospheric conditions. Then the dry- ing curve follows along the limit curve for atmospheric pressure unti the press is opened. according to the graph, the moisture content of the board as it came out of the press at 140.50 0. should be 2.4 per cent. This is only 0.2 per cent lower than the actually measured mois- ture content for the example board. 38 SUflLARY This study shows the effects of moisture distribution, initial pressure, and overall board density on the preperties of the finished board. Sample boards were made at tIo different overall density 1e- vels us n? four levels of initial pressure. Surface water was added nix to 1/2 the boards made at the lower overall density and to all the boards made at the higher density level. Eur fece Water was acded to the flat me by spra3~ in: it onto the inner surfaces of the cauls. Fif- teen grams wera added to each surface. The initial pressures used were 100 psi, 200 psi, and 300 psi. The two target dens i ies were 0.600 and 0.650 gm/cmB. The density distribution through the thickness of each board was determined. Thermocouples were used to measure teaperature changes at different points in the board thickness While the boards Were being pressed. The Way in which the basic Variables influenced the finished board properties Was discussed and also the conse uent effect of the density t7 of the boards. Tt was noted Ho distribution on the modulus of elzs tic C that the density distribution was result of the weakening of succes- {D sive layers of flakes during the closing tine of the press. The dry- ing process in the press was discussed as it was related to the equilibrium moisture content of wood in superheated steam. LI TI' TIJZTi OF T331 LTULY Although this study illustrates the general effects of the Va-riahles in the production of wood flake board, its Value is nece eerily limited by certain factors. 1 lim ' ted num er of sa:n— ple boards were made. a larger number might have pointed out effects not shown in this study. The size of the specimens were limited by the size of the press available. In order to obtain bending specimens of stands d length, the thickness f the boards was limited to 3/8 inch. La r.g er, thicl {er boards would have ex- hibited SOIHe mh t different density distributions. In the laboratory a closer check can be kept on the fl moisture content and size but tie method of felting the mats by hand is not as desirable as the me oh nine -fe1ting process used br industry. 40 10. L TEST-LIT U123 C ITS Duncan, T. F., "fffect of he at Transfer in Multi—pleten Process." HOOd and wood Products, Herbert a. Vance Publisher, Louisville, I? cntucky. Vol. 64, No. a, June 1959: pp. 56-58. Ellwood, Eric L., "Pronerties of 1nerican Beech in Tens ion and Compression Perpendicular to the Grai nand their Relation to Drying". Yale University School of Forestry Bulletin *0. 01, New Haven, Conn.; 1954. Fahrni, F., "Des Jerpr res sen von SpanJlatten bei gefouchteten Oder feuchteren De cksognen." Holz als Roh- und nerkstoff, Springer- Verleag Berlin, (14), No. 1, Jam nary 1O 56- pp 8—10, Kaumcn,’d. G., "Eguilibrium Hoisturs Content Relations and Drying Control in Superheated Steam Drying". Commonnealth Scientific and Industrial Research Organisation, Division of r‘orest Products, ;;? lbourne, Australia; 1956. Key1werth, a., ”deer are Verf iron 'foujhtersn mussenschichten' zur Eerstellung dreischichtiger Iolzspa n1latt n". holzforschung und Holzverwertung, Georg Fromne a 00., Hion, Mt rreich, 1959 pp. 51-57. fiackwitz, G., "2““ K nntnis der Vorgange bei der Verleimung von Polesn’nnen zur S Janholzpl atten in beheizten hydraulichcn Iressen”. Entwicklung und Iicrstellunj von Iolsspanolatton, Deutsche Gesellschaft fur Iolzforschung, Stuttgart, 1955: pp. 11-14. Strichler, i. D., "Effect of Press Cycles and oisture Content on Properties of Dou las- fir I11xebo.re" Forest Products Journal, FOTBSC Irqucts neserrtw Society, hadison, Misconsin, Vol. 9, M. 7; JUILv', 1C5,: Pp. 203-2 5. Suchsland, Otto, 'an ana1,sis ofa “Jo-dpecies, mhree-Lszyer Wood Flake anrd", Quarterly P11 _et“1 n of the ”1ch15an fnricultur l E"crln nt Ptation, East Lansing, 'ichibcn, Vol. 43, Io. 2, hov.l9c0: pp. 375-393. Juchs_and, 0t to, “Ln an; lysis of the Particle joard Process", Quarterly illxein of t e Hiczi3an nnr1c111u1 1 Line or Station, Etlst Lans in3, iichigan, Vol. 42, To. 2, Eov. 195 pp. 350-3 2. "Tent1tive Tethods of Test for Evaluating the Propertie. of Pu1ld1n3 Fiber hoards," 55TH designation D103 7-50T, emerican Society for Testing Materials, Philadel*h_a, 19:11., l/DD: p13. ‘--’1_ ()9. 41 m3