THE EFFECTS OF SODIUM CHLORIDE ON THE DRYING CHARACTERISTICS OF A REFRACTORY HARDWOOD Thcsls for Hm Degree of DH. D. MICHIGAN STATE UNIVERSITY John G. Haygreen 1961 This is to certify that the thesis entitled The Effects of Sodium Chloride on the Drying Characteristics of a Refractory Hardwood presented by John G. Haygreen has been accepted towards fulfillment of the requirements for Pthg degree in ForeSt PrOdUCtS flomd Dr. Otto Suchsland Major professor Date June 1961 0-169 L I B R A R Y Michigan State University I . on “m-..- __.___ ABSTRACT THE EFFECTS OF SODIUM CHLORIDE ON THE DRYING CHARACTERISTICS OF.A REFRACTORY HARDWOOD By John G. Haygreen Chemical seasoning has been used for a number of years to re- duce the amount of seasoning degrade which often occurs in thick re- fractony hardwoods. The reduction of surface-checking which results from such treatments is due to a change in the sorption character- istics of the surface layers of the lumber and to the antishrinkage effect of the chemical. These characteristics also make possible kiln-drying under humidity conditions which would be too severe for untreated material. In this study, the kiln—drying of green chemically-treated lum- ber was investigated to determine what effect the presence of a salt has on the factors which limit the optimum drying conditions. A com- mercial sodium chloride preparation applied to northern red oak was used throughout this study. The behavior of elastic strain, of mois- ture, and of irrecoverable creep during drying under different humidity schedules was studied. The salt concentration gradient and the effect which drying conditions have on this gradient were also investigated. When treated oak was dried under a normal kiln schedule designed for untreated material, the stresses developed were lower than those of untreated stock and decreased with increasing amounts of salt. A correspondingly lower degree of tension set was produced prior to stress-reversal. The rate of stress relief during conditioning was the same for treated and untreated stock. In subsequent tests, the John G. Haygreen initial relative humidity was decreased developing maximum tensile stresses as indicated by set, checks, and tensile strain. The diffusion of salt toward the center of the samples was greatly reduced by de- creasing the initial relative humidity. The development of set was more rapid under the lower initial relative humidity conditions, but the total magnitude of set produced was not changed. A more severe sequence of drying conditions was designed on the basis of the first tests. The treated stock which was dried under this sequence developed a stress pattern similar to that occurring in untreated stock. This indicates that the optimum schedule was ap- proached. This stock was dried without degrade in 55 percent of the time required under the normal schedule. The salt concentration in the center portion of the boards dried under the normal schedule was found to be about one-third as high as the concentration on the sur- face. The movement of salt to the center of the boards was greatLy reduced by using the more severe schedule. If the relative humidity conditions are property chosen, the behavior of stress, moisture, and set in sodium chloride-treated lum- ber will be very similar to that normally encountered in untreated lumber. The proper initial relative humidity can be predicted on the basis of the shrinkage-relative humidity relationship, if known. THE EFFECTS OF SODIUM CHLORIDE ON THE DRYING CHARACTERISTICS OF.A REFRACTORY HARDWOOD By A ,J x 1"," John G.~Haygreen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forest Products 1961 ACKNOWLEDGMENTS The writer desires to express his sincere appreciation to Dr. O. Suchsland for his guidance throughout this investigation. The guidance offered by Dr. A. J. Panshin and Dr. A. E. Wylie is also deeply appreciated. Grateful acknowledgment is extended to Dr. E. A. Behr for his advice and assistance regarding the chemical analyses involved, and to Dr. W. D. Baten for his aid in the design of the statistical analysis. ii TABLE OF CONTENTS AMOWI—‘EII‘INEN’I‘S I O O O O O O O O O O O O O O O O O O 118T OF TABI—IES O O O O O O O O O O O O O O O O O O I—lST OF I LLUS TRATI CNS 0 O O O O O O O O O O O O O O 0 INTRODUCTION . . . . . . . . . . . . . . . . . . . . Chemical Seasoning . . . . . . . . . Drying Stresses and the Optimum Kiln-Schedule Statement of the Problem . . . . . . . . . . REVIEW OF LITERATURE . . . . . . . . . . . . . . . . Drying Stresses . . . . . . . . . . . . . . . Properties of Salt-Treated Wood . . . . . . . Chemical Seasoning . . . . . . . . . . . . . . Preservative and Fire Retardant Salts . . . . Drying of Chemically Treated Lumber . . EXPERIMENTAL PROCEDURE . . . . . . . . . . . . . . . Selection and Preparation of Samples . . . . Experimental Design . . . . . . . . . . . . Determination of Strain . . . . . . . . . . . Determination of Moisture Content . . . . . Determination of Set . . . . . . . . . . Determination of Salt Concentration . . . . RESIH—ITS O O O O O O O O O O O O O O O O O O O O O 0 Behavior of Elastic Strain . . . . . . . . . . . Moisture Content Gradients . . . . . . . . . Indicated Set . . . . . . . . . . . . . . . Salt Concentration . . . . . . . . . . . . . Drying Time and Quality . . . . . . . . . iii Page ii Table of Contents. - Continued. DISCUSSION AND CONCLUSIONS . . . . . . . . . . . Basic Properties of Salt-Treated Lumber as to Drying Stresses . . . . . . . . . . Drying Conditions and Stress Development . Drying Conditions and Salt Distribution . Further Research . . . . . . . . . . . . . SWY O O O O O O O O O O O O O 0 O O O O O O O BIBuwRA-PHY O O 0 O O O O O O O O O O O O O O 0 APPENDIX................... iv Related . Page 66 66 68 . 72 73 75 78 . 82 10. 11. LIST OF TABLES Page Outline of the Drying— Tests and the Levels of Salt Treatment Used in Each . . . . . . . . . . . . . . . 26 Drying Time and Elastic-Strain of Surface Strips when Surface-Checks Appeared . . . . . . . . . . . . bl Percent of Tangential Shrinkage from Green to Seven Percent Moisture Content . . . . . . . . . . . 63 Drying Times and Quality . . . . . . . . . . . . . . . 6h APPENDIX Analysis of variance of the Maximum Tension Strain in the Surface Strips During Test I . . . . . . . . 83 Analysis of variance of the Maximum Tension Strain in the Surface Strips in Tests II, III, and Iv . . . 8h Analysis of variance of the Moisture Content of the Surface Strips after 20 and 25 Days of Test I . . . 85 Analysis of variance of the Moisture Content of the Surface Strips After Two Days in Tests II, III, and IV 0 O O O O O O O O 0 I O O O O O O O O O O O O O O 86 Analysis of variance of Indicated Tension Set in the Surface Strips After Two Days of Drying in Tests I, II, III, and Iv. . . . . . . . . . . . . 87 Analysis of variance of Percent Tangential Shrinkage of the Entire Board . . . . . . . . . . . . . . . . 88 Analysis of variance of the Relative Salt Concentration in the Surface Strips After Different Periods of Bulk— —Piling . . . . . . . . . . . . . . . . . . . 89 LIST OF ILLUSTRATIONS Figure Page 1. Sorption Curves for Sodium Chloride-Treated and Untreated Spruce and Pine . . . . . .-. . . . . . . . 12 2. Shrinkage-vapor Pressure Curve for Sodium Chloride- Treated White Pine . . . . . . . . . . . . . . . . . lb 3. Shrinkage- vapor Pressure Curve for Sodium Chloride- Treated Pine and Spruce . . . . . . . . . . . . . . . 16 A. Location of the Eight Test Strips in a Drying Sample . . 29 5. Method of Determining Elastic Strain and Set . . . . . . 3O 6. Strain of the Surface and Center Strips from Three Treatments During Test I . . . . . . . . . . . . . . 39 7. Strain of the Surface and Center Strips During Tests II 5 III ’ and IV 0 O O O O I O O O O O O O O O O b2 8. Strain of the Surface and Center Strips During TeSt V C O C C . . O . . O O C O C O O O O O O O O O 1m 9. Strain Before and After the Equalizing and Conditioning Periods in Tests I and v . . . . . . . . AS 10. Moisture Content Gradients During Test I . . . . . . . . AI? 11. Moisture Content Gradients After Two Days of Kiln-Drying in Tests II, III, and IV" . . . . . . . . AB 12. Tension and Compression Set in Treated and Untreated Boards During Test I . . . . . . . . . . . SO 13. Indicated Tension and Compression Set in Samples With Three Different Amounts of Salt During Test v . . . . 51 1h. The Change in the Percent of Green to Oven—Dry Tangential Shrinkage of the Surface Strips During Tests I, II, III, and Iv . . . . . . . . . . . 53_ vi List of Illustrations. - Continued. Figure 15. Relative Salt Concentrations After Bulk—Piling for various Periods . . . . . . . . . . . . . . . . 16. Salt Concentration Gradients of the 7S Pound per M Sq. Treatment at various Times During Test I . . . . . 17. Salt Concentration Gradients of the 210 Pounds per M Sq. Ft. Treatment at various Times During TeSt I O I O O O O I O O O O O O 0 O O O O O 18. The Relationships Between the Drying Conditions and the Change in the Salt Concentration Gradient During the First Eight Days of Drying . . . . . . . l9._ The Relative Salt Concentration Gradients at the End of Tests I and v . . . . . . . . . . . . 20. Average Moisture Contents of Entire Samples During Tests I and v . . . . . . . . . . . . 21. Representative Samples Showing the Occurrence of Honeycomb and Surface—Checks in the Dried StOCk from Test V 0 O O O O O O O O O O O O O O O 0 vii Page St Ft. . 56 . S7 58 60 . 61 65 INTRODUCTION Chemicals may be applied to lumber to improve its seasoning characteristics, or to reduce its susceptibility to fire or decay. Some chemicals also improve the dimensional stability of wood, but little commercial use has been made of this fact. Regardless of the purpose for which a chemical is added to wood, the drying character- istics may be altered depending on the nature of the particular salt used. Chemical Seasoning Chemical seasoning refers to the process of treating green wood with a hygroscopic chemical and then air-drying or kiln-drying this material in the conventional manner. The primary reason for using such a process is to reduce the amount of seasoning degrade, mainly surface-checking. A water soluble chemical is used which penetrates the surface layer of the lumber, reducing the vapor pressure in this layer. By reducing the vapor pressure the surface is maintained at a higher moisture content than is normal. The antishrink effect1 of the chemical in conjunction with the higher moisture content tends to reduce or eliminate checking. In the last several years the most widely used commercial pro- duct for chemical seasoning has been a sodium chloride preparation. 1Antishrink effect refers to the reduction in the shrinkage caused by the bulking volume of the chemical within the cell wall structure. 1 2 This preparation has found considerable application in the southern hardwood region where it is used on 6/h and thicker oak, pecan, and beech. A survey conducted among the larger hardwood mills in the south in 1958 indicated that about 35 percent of the mills surveyed used some "salt" for treating their thick refractory hardwoods. Drying Stresses and the Optimum Kiln—Schedule The expense of kiln—drying green hardwood lumber can run up to $50 per MBF for thick stock. Although this expense is a function of a number of factors, it is closely related to the drying time. The total cost of drying, however, also includes the loss in measurement and grade during drying. The optimum kiln-schedule could be defined as one which minimizes the total cost of drying. Although this cost is a function of shrinkage, seasoning degrade, and drying time, the most important factor in minimizing this cost is drying time. This is because the other two factors must be held nearly constant. Shrink- age can only be varied a very small amount, and seasoning degrade must be held as near zero as possible due to the high cost resulting from its occurrence. Therefore, the use of a seasoning agent would reduce the drying cost if the drying time were reduced appreciably while not increasing the shrinkage or degrade. It is very difficult to determine the optimum kiln-schedule which minimizes drying time while eliminating defects. Ideally, schedules should be based upon the internal stresses which develop in the lumber as a result of the moisture gradient. In refractory hard- woods stress development follows the pattern outline below. 1. When drying is begun the surface of the board tends to dry 3 to the EMC in the kiln. A steep moisture gradient thus develops in green lumber. Since the outer portion of the board is below the FSP this portion tends to shrink. 2. The outer portion can shrink only a very small amount due to the reaction of the inner portion of the board. Tension stresses in the shell and compression stresses in the core result. The stress produced at this time is a function of the difference between the nor- mal unrestrained dimension and the actual dimension of the layer. 3. Set occurs as a result of creep below the proportional limit and stress above the proportional limit. It occurs both in the shell and in the core, but is more severe in the shell due to the higher stress. Set produces a change in the normal unrestrained di- mension and thus in the stress. A. As drying continues the interior of the board drops below the ESP and tends to shrink. As a result of surface tension set and core compression set the core exhibits a higher coefficient of shrink~ age than the shell. Therefore, as the core continues to shrink the surface layers go into compression and the core into tension. The initial stresses are thus reversed. 5. The moisture gradient flattens during the final portion of the run producing still higher core tension stresses and shell compres- sion stresses. At the completion of the drying qycle this stress con- dition is called casehardening. The initial drying stresses are a result of the moisture grad- ient. Set begins to develop almost immediately, affecting the stresses both in the shell and in the core. Stresses during most of the run are thus a function of both the moisture gradient and set. At the end h of the drying process when the moisture gradient is flat the residual stresses are a function of set. The relationship between drying stresses and drying defects is important in the design of the optimum kiln—schedule. Surface-checking results from surface stresses which exceed the maximum strength. Checking is thus a direct result of drying stresses. Collapse can result from either drying stresses or from hydrostatic tension. It is believed that hydrostatic tension is the more common cause of collapse. When collapse is caused by drying stresses it could be considered ex- cessive compression set. Honeycomb usually is a result of surface- checks which are extended into the center of the board as the drying progresses. It may also be produced by tension failures in the center portion of the board. This often is the case in portions which have previously collapsed. There are three critical points around which the truly optimum kiln-schedule must be based. These points are: (1) the initial tensile stress on the surface which must be controlled so that surface—checking does not occur, (2) the temperature and relative humidity early in the schedule which must be controlled such that hydrostatic tension does not cause collapse, (3) the temperature and the stress level prior to stress reversal must be controlled so that excessive set in shell and core does not occur, which would result in severe stresses after stress reversal. Such stresses could result in tension-failure type honey— combing. It can be seen that the first portion of the kiln run is the most important in the design of the Optimum schedule for a refractory hardwood. The stress development which takes place after stress- S reversal is determined by the set which is produced prior to stress- reversal. After stress reversal, the relative humidity in the kiln is, therefore, not of primary importance. The temperature must be con- trolled after reversal, however, to avoid lowering the tensile strength below the level of the tensile stresses. If this occurred honeycomb would deveIOp. If the strength, elastic, moisture, and diffusion character— istics of a wood were known a theoretical, rather than an experimental, approach could be used to determine the optimum schedule. Theoretic- ally, derived moisture distribution—time values could be used to as- certain the initial stresses which would develop under proposed initial drying conditions. From the stress-strain-creep relationship the amount of set which would be produced could be estimated. In order to minimize drying time the maximum strength and set values would be approached, but to minimize drying defects these maximum values should not be exceeded. Due to the variable nature of wood a statistical approach would be necessary, and an acceptable limit for seasoning defects would have to be chosen. Statement of the Problem The length of time required to dry lumber which has been treated with salt may differ from that for untreated material. Several inves— tigations have indicated that woods dry faster when treated with sodium chloride, while others indicate the opposite effect. These comparisons were noted when both the treated and the untreated lumber were dried under similar conditions, as would necessarily be the case in air— drying. In kiln—drying, however, the drying conditions can be adjusted to the drying characteristics of the wood being dried. Lumber treated 6 with a chemical seasoning agent generally has a flatter moisture content gradient and thus a less severe stress distribution than has untreated lumber being dried under the same conditions. For this reason, the drying conditions could be made more severe during the drying of treated lumber, possibly resulting in a significantly shorter drying time and correSpondingly a lower drying cost. It is apparent that a better understanding of the effects of salt is necessary if the efficiency of drying salt-treated wood is to be improved. The purpose of this investigation was to evaluate the primary factors which determine the optimum kiln-drying conditions for salt-treated lumber. As with untreated lumber the internal or drying stresses, which are developed by the sequence of drying conditions, are the most important factors. REVIEW or LITERATURE Drying_Stresses The presence of drying stresses and their significance in the proper drying of lumber were probably first recognized by H. D. Tiemann in 1915. At that time he developed the familiar prong test commonly used to determine the presence of casehardening. The first complete study of drying stresses was carried on by Peck (36) around l9hO. He investigated drying stresses in two—inch thick blackgum. Peck used what is now known as the "slicing or strip technique" to estimate the stress condition. Peck cut strips from the face of a section cut from the board. He measured the length of the strips both before and after cutting. The change in the length of the strip was taken as a measure of the stress present. Thus, if a strip were longer after cutting, a compression stress was indicated, and if it were shorter after cutting, then a tensile stress was indi— cated. Peck assumed that the magnitude of the change in length of the strips was an indication of the magnitude of the internal stress. Although this assumption was later (12) (28) (29) shown to be an over- simplification of the case, the information obtained in his study made possible the design of a fast drying schedule for this species. In l9h6 Loughborough and Smith (h?) published the results of a more comprehensive study of the same type as that of Peck. They, however, investigated sweetgum. The general finding of both of these early studies was that the tension strains on the surface build up rapidly after drying begins, and that after a few days this strain 7 8 begins to lessen. It is then possible to reduce slightly the rela— tive humidity in the kiln without risking further surface-checking. An empirical formula was derived for determining when the original relative humidity in a kiln can be safely reduced. It was found that when E from the relationship present MCI— kiln EMC B = original MC - kiln EMC ’ reaches 0.7 then it was safe to reduce the humidity. MC is used as a symbol for moisture content and EMC for equilibrium moisture content. From the knowledge of the behavior of drying stresses gained in these studies the U.S. Forest Products Laboratory published in 1951 a new set of kiln schedules for American woods. The design of these schedules is thoroughly discussed by Rietz (39). These schedules resulted in more rapid drying and reportedly did not increase the occurrence of degrade as compared to the schedules which they replaced. More recently McMillen (27) (28) (29) has employed the strip technique to study the drying of northern red oak. His extensive studies covered the effects of temperature and relative humidity on both the strain and the set. He estimated the set by noting the change in the total shrinkage of the strips. He found that the magnitude of the set was closely related to the severity of the relative humidity, but was not apparently affected by the temperature. A method similar to the strip technique has been used to deter- ming uniaxial residual stresses in steel rails (3). Other methods, employing strain gages, used to determine residual stresses in metals do not appear applicable to the determination of drying stresses. The severe moisture conditions, high temperature, and long-time load— ing involved would make strain gage analysis unreliable. Improvement 9 of strain gage techniques may eventually make such analyses possible. Kuebler (23) recently used a different technique to estimate the drying stresses in veneer. He sliced very thin sections from veneer and calculated the stresses from the curvature of the sliced sections. It appears that he assumed that Young's modulus is constant through the veneer. This technique might also be applicable to a study of stresses at the surface of lumber, but has not been so used as far as the writer has been able to determine. In the "stress" measurement techniques discussed above it is the elastic strain of a section cut from a drying specimen which is measured. If these strains are to be related to the correSponding stresses a complete understanding of the elastic and plastic proper— ties of wood perpendicular-to-the—grain are required. Since most failure theories are discussed in terms of stress rather than strain it would be very helpful, though not necessary, to convert from strain to stress. Several investigators have worked on this problem in the last six years. Ellwood (12) investigated the effects of temperature and moisture content on the perpendicular-to-grain tensile and com- pressive prOperties of beech. He found that tensile and compressive properties were both affected to the same extent by temperature and moisture. Youngs (50) studied exhaustively the mechanical preperties of red oak perpendicular—to-the-grain. His work included tension and compression characteristics under various temperatures and moisture contents. He also investigated the effects of the duration of heating, creep and recovery, and stress relaxation. This study is a significant step toward the determination of drying stresses. However, even with 10 this information the state of stress at a point cannot be directly determined, but only the average stress in the strip. Since the stress gradient is presumably very steep near the surface of a board, at the beginning of the run, the stress at the surface is much higher than the average stress in the slice. In order to overcome this difficulty Youngs and Norris (51) develOped a mathematical method of estimating the perpendicular-to- grain normal and shear stresses which occur at any point in the cross section of a drying board. For this solution the following are required: (1) elastic strain measurements such as those of McMillen, (2) the moisture content and temperature of the slices at the time they are cut, (3) data on perpendicular-to—grain elastic properties under the temperature and moisture content conditions of the slice. This method uses a stress function which is evaluated by the use of simultaneous equations and the principle of least work. It presents rather imposing problems in carrying out the mathematics involved. At the present time Youngs and Norris are trying to program it to a computer. A different subject in the area of drying stresses which has received considerable attention in the last several years is that of collapse. Kauman (20) investigated tension and compression set in eucalyptus and related his findings to the extent of collapse and to the overall shrinkage of the specimens. He concluded that drying stresses are not the most important collapse-inducing force, but that the concept of hydrostatic tensions is a more likely explanation. Kemp (21) working with aspen found that the drying time prior to collapse decreased with an increase in temperature and increased with an increase in relative humidity. 11 The most recent and thorough" study of collapse is that of Ellwood, Eckland, and Zavarin (12). Their conclusions are in general agreement with those of Kauman. The reduction in collapse which they found in samples in which water had been replaced by an organic liquid was consistent with predictions from liquid-tension theory. It is interesting to note that a void is present in the litera— ture in the area of surface-checking as related to surface stresses. This particularly is significant since surface—checking is one of the most common seasoning defects. The difficulties involved in deter- mining such stresses are, presumably, reSponsible for this void. Properties of.Salt-Treated Wood There are two properties of salt-treated wood that are partic- ularly important in regard to the development of drying stresses. These properties are: (1) the sorption characteristics, (2) the shrinkage characteristics. Paten (3h) found that in Scots pine containing three percent sodium chloride the entire sorption and desorption curves lay above those of untreated material. (See Fig. 1.) At about 92 percent rela— tive humidity these curves approached the vertical. He compared these results with a calculated adsorption curve and found that the experi- mental curve lay above the calculated curve from O to 75 percent rela- tive humidity. From 75 to 92 percent the curves agreed quite closely. A study which included proprietary fire retardants and preserv- atives, as well as common salt, was carried on by Kollmann (22). Spruce which was impregnated with a saturated sodium chloride solution exhibited somewhat different sorption characteristics than the pine in Paten's experiment. (See Fig. 1.) One reason for this difference may 50 E 40 .\.. S 30 Is 5 b 20 t b I\ ‘3 IO 6 a o 12 .Sa/f freafea/ 560 1'5 pine (Paint) I [/rn‘rca‘fea’ pine ' \ Sa/f' freafcd spruce \ i I (Kc/{mend ’ \ . I'- I I 7 I A I / I XIII” I- ‘ // ,,~ __ ”.1--_ 0 20 40 60 80 /0'0 REZAT/I/E HUM/D/ TY Fig. 1. Sorption and desorption curves for sodium chloride- treated and untreated Spruce and pine. 13 be a higher salt concentration in the spruce. Kollmann did not deter- mine the salt content of the dry samples. It should be noted that Kollmann found that the sorption curve for the treated material was in close agreement with that of untreated material below 60 percent relative humidity. It has been widely assumed that the entire sorp- tion curve for treated wood is above that for untreated stock. The sorption curves of Minolith and Basilit-3 treated wood were found to be similar to that of sodium chloride—treated material. Peck (35) determined the antishrink effects of various chemi— cals and the relative humidity in equilibrium.with different percent- ages of saturation of aqueous solutions. The antishrink effect of salt was found to be 12.7 percent compared to hh.7 percent for urea. This percentage is based on the reduction in green to oven—dry shrink- age. He found that the relative humidity in equilibrium with a satu— rated common salt solution was 78 percent; that for a 20 percent satu- rated solution was 96 percent. A study which is more directly applicable to an analysis of drying stresses in treated lumber was directed by Stamm (Ah). One portion of his study dealt with the relationship among shrinkage, relative humidity, and salt concentration in white pine. (See Fig. 2.) The vertical lines labeled I through Iv in Fig. 2 indicate the initial relative humidity conditions of the tests in this study as will be discussed later. It should be noted that the curve for the pine treated with a completely saturated salt solution is not greatly dif— ferent from the curve for pine treated with a 1h percent saturated solution. It can be seen from this curve that pine treated with the 2-gram solution would start to shrink when the relative humidity 1h .83 Bee: Roseansfleofleo Savor. no.“ Q53 anemone noaguomwxfiunm .m .mE «£53m. :36.» . WQBWWWQQ QQQ<> MEKVNMQ Q; m 0 Nd k6 v.0 h..0 V6 .0 ~10 /—M H/ H flaw ”—0— - --*—-—--—-——-———q 14/ /t // M on}. «.36 09 tum Ste .28 to owl {SE 69 odd Do? 633 no of «SE 09 «we. Us? .méc m. $.36 62 O\ 39V/WV/é/HS‘ J/E/JJW/WO/I Z 15 dropped below about 9h percent. Pine treated with a saturated solu— tion would not start to shrink until the relative humidity dropped below 76 percent. In order to investigate the shrinkage—relative humidity rela- tionship further, data from Kollmann, Paten, and Stamm were combined in Fig. 3. Strictly speaking, it is not entirely correct to combine this information since different species and testing procedures were involved. It is interesting to note, nevertheless, that the curves from Kollmann and from Paten are quite similar except between 0.7 and 0.8 relative vapor pressure. Chemical Seasoning The use of chemicals to aid in the seasoning of lumber is not new. In the early 1900's in Australia, lumber was treated with cane sugar to reduce checking. This was then known as the Powell Process. Other low grade sugars were used as seasoning agents in this country about that same time. In parts of Europe during the 1800's timbers were sometimes soaked in sea water to reduce Splitting. In the last twenty years a number of chemicals have been suggested and investigated as possible seasoning agents. These include common salt, urea, invert sugar, molasses, diethylene glycol, and urea aldehyde. Of these, sodium chloride preparations and urea preparations have been used on a wide commercial basis. The others have been rejected because of cost, dif- ficulty in application, change in the color of wood, or other diffi- culties. Personnel of the U.S. Forest Products Laboratory have con- ducted a number of investigations (35) (37) (A8) into the use of chemical seasoning agents. 16 roosuam pom ego” Uopmonpuoguoano Evfioom pom Ergo snowmen Conglommxcfiham .m .mHm MQS WWMQQ odors. meats «Not 0.. m6 m6 «6 v.0 .06 v.0 n6 N6 50 so 7 , _ I I I _ _ / i 7 ,1 _ _ r . m _ _ , A T: L : I ,- . I- I I, IIIIL I- N r I_ l x i_ w A , _ _ u r _ t _ _ TI r I l _ II «I. . r l / ,1 _ _ ._ H // . i _ _I- IL 1 : I I r. I i z ..I III I I I It: I w _ W . I I o , u _ «552%.» 25¢ 9.353 1 tan. o>\ , u I./ IN. [H/ ”finish bxmexxex256? A . m 0k. \ T a L. stofiéotob brutuso _ ox msbeo \natmxm .ocm Matteo Lexmv. , \Illnlulullllllllllj \ \ \ \\ llllll uh’ “Pl llllll i \ \ ‘Hd \ \ rl'lll'll-‘lll'Tl'll \ \ 1 Min \ \X'lll llllllllllll v \ N‘NJ \ \ lllllllllllllll s \\ do ‘\\ \ \ \< More max Muncfix m. \.\.\. EWSCS Riflm. Q2? >\\v\m\k.fl m\<\>\\\333- x \ emu... 0.5.56 marsh mousse can commasm m5 Mo 5.93m .m .mE QMC>RC III MOYKQSW I QECAQQ \ was H moms 5 muoruom mcflcosfiocoo cam meanness one not... new 888 Swarm .a a: QMkEmC .I I (7w / 'xv/ e...0/) xv/vous 37.191773 NUVQQSM. I $§<§§ S36 oz m QSNSS am he SSS... 3.. am o s... a. E o T K m... \\ m- > .v\ o A / M 0 ex r./\\ O /\ 2 o I / ,Xe. // [Al/l /4 [III/II III// 3 IIIH N I/IuML N NM 1 .N. RWWK kwmh V . a. ma mb\ fl Ce Nb 8: H 63 m 9.3 no (:10) 2114.91 N 7/3/ 116 no strain measurements were taken between equalizing and conditioning. As would be expected, the conditioning period in test I which was designed for untreated material was too severe for both treatments B and C; therefore, reverse casehardening was produced. The important fact is, however, that the change in stress during equalizing and conditioning was the same for both the untreated and treated stock. This suggests that a normal conditioning period could be easily altered for treated stock by changing the conditioning time rather than the conditioning temperatures. In designing the conditioning period for test v the above find- ings were taken into consideration. An equalizing period of 2h hours was used in both test I and v, but the 18 hour conditioning period of test I was reduced to 12 hours for test v. It was found that this treatment produced complete stress relief for treatment 0. The case— hardening stresses in treatment N were not completely removed, while very slight reverse casehardening was produced in treatment P. Moisture Content Gradients The moisture gradients during test I are shown in Fig. 10. During the first 23 days of drying there was a difference of over 10 percent between the average moisture contents of the surface strips of treatments A and C. After this time the gradients of all treatments ‘were very similar. Samples of treatment C exhibited an essentially flat gradient for the first 20 days of the run. Fig. 11 shows the moisture gradients after two days of drying in tests II, III, and Iv. After two days there was no significant dif- ference between the surface moisture contents of the 150 and the 210 pound treatments. (See Table 8.) However, the MC of the 90 pound .H emu» marge mucofiommm pcopcoo 33.302 hub VKQSW ESQQ IRQWQ Q fl mVAVRV.w\ nvw .X _ saw on N 8 es 9»va Nm. 6w a .na bow Qm an .oH .maa = Q» Rn . no a % \ mXYQ VN Q ON a B ha C) .LNJJ N0) 38/415701” l/I/JJE/Jd C) 9’) C) It) Om Aum 118 .3 new .HHH .HH Base 5 moagéa Mo manor 95 93mm mpcgomnm pcoucoo 83.302 7: .mfim 1%ka 93m; SQQK IR QNQ has ..n\... 0 use . sea, 6 Q 0 ON ON 9v 9» 0w 0w 0% “\x 0% _ MN I.\\ B IIIII KWE QQ\ rIIII IF KWINIHNL QQ\ Kk Gm. \<\3 .2. he do, 3}: 9m - HA .3. E\3 0.3 - Knew 3‘33. - sax. ..m\o OQDX Q k WINK. 0 GM D? On On 03 .1 N31 N03 35/71 SVOW .l NJJHJO’ A9 treatment was significantly lower. These results agree with the fact that the maximum tensile strain of the 90 pound treatment during tests II, III, and Iv was significantly greater than the strains of the 150 and 210 pound treatments. Indicated Set In treatment A, a slight change in the green to oven-dry Shrinkage of the surface and center strips occurred after only one day of drying (See Fig. 12.). very little change occurred in treat- ment B until the 10th day of drying. The tension set in the untreated stock developed more rapidly than did compression set, but the opposite was true for the treated material. The amount of tension set in the treated stock remained very small throughout the kiln run; for the last 16 days of the run it was nearly zero. After the first 16 days, the compression set in the core of treatment B increased rapidly un- til the end of the test when it was about twice as high as the com- pression set in treatment A. The behavior of indicated set can be explained on the basis of the magnitude of the elastic strains up to the time of stress reversal. After stress reversal there does not appear to be a direct relation between set and the magnitude of strain. The conditioning period in test I was found to remove both the tension and compression sets in treatment B. A small amount of com- pression set remained in the core of treatment A after conditioning, although the tension set on the surface was apparently relieved. During test v (Fig. 13), set in both the shell and the core developed rapidly in all three treatments. The final set was approx— imately the same in tests I and v. The tendenoy of the tension set to 50 .H poop weapon mohmon copmoppco vow oopmopp CH pom :ofimmoudaoo new cofimcoh .NH .mfim bEion \<.\\\< k0 WXYQ mm vm om a: .3 n we I w M I _ QMA SMOKES I uuhln Ron since .. 3min I-.-I-TIIIIII I % I «C on SSS R 1.38 . 3Q swat >3 1560 _ _ _ _. h \ Q T ‘}l “P Z "" .79 V/VN/é/HS‘ 7V/UJI/ N “3 z) (.79 tow/we 7WVcVOlV H7 I Hfilflllll 7A 51 .> ammo mousse mama Ho madness pcouommfio omega no“: moaosmm a“ pom cofimmoposoo one cofimcoh .mH .mHm 62:3 >3? to wed mm cm 3 S o v o Mil \ I 7 \ K r/ {I I v 3. I \ | 1 . II a; ‘s. \ o (39 Toma/HQ 7vwo0/v z —— 39 vyN/afls 7wupv Z) \ \ I \ 4 N #x/ \\ mqoo ‘0 UEIVD/UN/ 139 52 be relieved and the compression set to be increased during the final stages of the run was again apparent in test v. The relief of the tension set on the surface can be explained on the basis of the com- pression strains present at the end of the run. However, in the core, increasing compression set was accompanied hy increasing tension strain. These observations appear to be contradictory. Fig. lb illustrates the relationship between the initial rela— tive humidity and the tension set of the surface strips. An analysis of variance (Table 9) indicated that the tension set in tests I and II was significantly less than the set in tests III, Iv, and I-untreated. It can be seen that as the initial relative humidity decreased the amount of tension set increased. Presumably, this is a response to the increasing surface stresses which develop as the relative humidity is decreased. It Should be noted that this increase in surface stresses was not indicated by the elastic strain measurements, but was indicated by the increasing occurrence of surface-checking. Salt Concentration The samples for the different tests were "dry-salted" and then bulk-piled for periods of time varying from 7 to 1h days. At the end of this period, the salt concentration gradient was determined. (See Fig. 15.) The salt gradient did not appear to change appreciably be— tween one and two weeks of bulk-piling. In fact, after 1h days, the surface strips in test I had a slightly higher salt concentration than strips in other tests after shorter periods of piling. There was no statistically significant difference in the relative salt concentration of the surface strips after the different periods of bulk-piling (Table 11). Though all samples were bulk—piled in a room at constant , .3 new a n n E HH H memo» marge weapon Summon mop Mo ommxfinam .. Hmfipcomafi Stage op coonm .Ho £50qu 9.3 3 smudge 0E. .43 .mE NQWK VMWQK \<§\ N. . SH - x a - . Skids t5 93G E - o N- . m. e . v w o m. e a. N o I a. _ a. 31E DE 3: L. . 3E RES: u. / l \KWI N .K/s \\E N a. / T a H /_ \ E a / s a N a \ a. _ .s a \ \. /Z . I \c / \ c Q o KK .%%\ new H mpmmp mo new map pm mpamfiumum dofipmnpcoocoo pawn o>HpmHmn one workman 36E thin Rx“, o _ 3.. 5E \\ _ H SE I III] m 35.. N when l t 93% «Md .m V Sm Q N6 «.6 v.0 ax. RA, .mH .maa fl #II!.-1-|:.4 - .H honking. Gm. XRQQ m: on N Emu IRA an E «Md .3 ha f l _. lllll °9 <3 flDEVV .JC711/M ‘7H16LL 3/2/19 273d 19‘?” V6 \0 C3 ) :17 :144 NO/l V5.1 NJD/VOJ 1 7 VS‘ 3/1/1 V7.75 61 w% m, m. .> cam H. Emma 9256 mmfidemm 93.28 Mo mpcmpcoo 953.39: ommnmfiw .om .mE Mm; QN VN b\<\>QQ \\ b \ \ G V. O \o 0 Q 03 IAI‘H‘iLJjA I All?! tun“) jun loin/AI 62 used was too severe for treatment N, hS pounds of salt, as evidenced by rather serious surface—checks which reappeared when the caseharden— ing stresses were relieved. Untreated stock was found to contain considerable honeycomb (Fig. 21). Treatment 0, 90 pounds of salt, was dried without serious defect; the checks which developed early in the drying cycle were closed and generally did not reappear. No seasoning defects were noted at any time in samples of the 210 pound treatment. The overall shrinkage in the widths of the boards is shown in Table 3. The shrinkage of the untreated stock was significantly great- er than for any of the treated samples (Table 10). The differences between the shrinkage of comparable treatments in test I and V were not statistically significant. Some cupping was evident in all treat- ments during all the tests, but there was no difference in the amount of cupping among the treatments and tests. The samples were not load- ed from above as would be the case in a stickered pile; thus cupping could occur without restraint. 63 TABLE 3. The Percent of Tangential Shrinkage from Green to Seven Percent Moisture Content in Flat—Cut Red Oak Boards Test No. Treatment Pounds of Salt/M Sq. Ft. Percent Shrinkage A O 7.71; I B 75 7.08 c 210 6.20 N AS 6.96 v 0 90 7.20 P 210 6.71 6b TABLE h. Drying Times and Quality of Drying Test No. Time for Drying from Lbs. of Salt Description of Season- 85 to 7 percent MC Per M Sq. Ft. Checks and Honeycomb (days) in Kiln-Dried Samples 0 none I b0 75 none 210 none 0 All h samples contained honeycomb V 22 hS In 2 of h samples—check- in on 50% of area, over 3 h" deep 90 On 2 of h samples-several checks of 2" X 1/32" or less, less than 1/8" deep 210 none 65 TEST 5 9616'. / M so. F1 w may i , Fig. 21. Representative samples showing the occurrence of honeycomb and surface-checks in the dried stock from test V. DISCUSSION AND CONCLUSIONS Basic Properties of Salt-Treated Lumber as Related to Drying Stresses Drying stresses are produced when the moisture content of the outer portion of the board drops below the fiber saturation point. At the start of the drying cycle, these stresses could be considered a function of: (l) the expected normal shrinkage to the moisture con— tent at the point in question, (2) the elastic properties of all por- tions of the board at the corresponding temperature and moisture con- tent. As drying progresses creep occurs accompanied by stress relaxa— tion. This greatly complicates drying stress analysis. By using the strip technique, instantaneous strain measurements can be obtained which are helpful in studying the relationship between stress behavior and drying conditions. If wood is treated with sodium chloride, the drying stresses which would develop under a given set of drying conditions are altered. The stress changes because the amount of shrinkage which would occur in a particular unrestrained layer is changed. In Fig. 2 it can be seen that white pine containing a saturated salt solution would not start to shrink until the relative humidity at 680 F. drops below 76 percent. If such a sample is placed in a kiln at a higher rela- tive humidity no drying takes place and thus no stresses develop. Relative initial tensile stresses at the surfaces of treated and untreated stock, at various initial relative humidity conditions, could be obtained from a family of curves of the type shown in Fig. 2. Several assumptions must be made when making such a comparison: (1) 66 67 the elastic prOperties of sodium chloride-treated wood are the same as those for untreated wood, (2) the surface of the drying boards reaches the EMC shortly after drying begins, (3) the moisture in the surface layers of the board is saturated with salt, or else the salt concentration is known. For example, suppose an initial relative humidity of 80 percent is known to be safe for untreated white Pine. If it is desired to produce the same initial tensile stresses on the surface of white pine treated with a saturated salt solution, the initial relative humidity would be set at hO percent. The initial relative humidity conditions of the tests in this study are indicated in Fig. 2 by the vertical arrows. The initial conditions for test V were the same as for test III. The relative initial tensile stresses for untreated oak in test I and treated oak in tests II, III, and IV are indicated by the length of the vertical arrows. It is assumed here that a similar family of curves is valid for red oak and that the surfaces of the samples contain a saturated salt solution. The results of the tests indicate that the comparison of rela- tive initial stresses discussed above is valid. In test I small sur- face-checks develOped indicating the maximum tensile stress was ap-_ proached. In test II no checking occurred on any of the samples. In test III no checking occurred in the 210 pound treatment, but in the 90 and 150 pound treatments some season checks occurred. In test IV surface-checks were produced in all treatments indicating that the stresses developed were higher than in the untreated samples of test I. The indicated set measured during the first portion of these tests also indicates that surface stresses were lower in test III, but higher in test IV than those of the untreated samples in test I. 68 The most important decision regarding the optimum conditions to be used for drying salt—treated oak is the selection of the initial relative humidity. If that decision is made correctly, then the sub— sequent drying conditions can be determined by following the procedure recommended by the U.S. Forest Products Laboratory. Basically, the FPL recommends that the relative humidity be reduced when one-third of the evaporable water has been removed, and the temperature be increased to about 180° F. when the core of the stock drops below the fiber sat- uration point. If the shrinkage—relative vapor pressure-salt concentration relationship were known for water-borne preservative or fire-retardant treated stock, then this information could be used as a guide in de- termining the initial relative humidity to be used for drying. About the only related information available today is that of Kollmann (22). He investigated only the sorption and not the antishrink characteristics. Drying Conditions and Stress Development Red oak treated with 75 pounds of sodium chloride, when dried under a normal schedule,1 developed maximum elastic strains about one- third lower than the strain in untreated samples. The moisture con- tent of the surface strips at this time was about h percent higher than that of the untreated controls. The shrinkage measurements dur- ing the first few days of the kiln run did not indicate that set occurred in the surface strip. In comparison, a set of 0.5 percent was noted in the untreated material after only two days of drying. 1In this discussion the "normal schedule" is the schedule recommended by the U.S. Forest Products Laboratory for untreated green 6/h red oak. 69 From these observations it can be concluded that the surface tensile stresses developed in the 75 pound treatment are considerably lower than the stresses in untreated material. For the 210 pound treatment, the maximum tensile strains were lower and occurred later than in the 90 pound treatment. Stress reversal occurred at approximately the same time in the treated and untreated stock. After stress reversal, the elastic strains of treated and untreated oak followed the same general pattern. The final casehardening stresses developed were more severe in the case of untreated stock as would be expected. When dried under a normal schedule, the casehardening stresses which developed in the treated material evidently resulted from rather high compression set in the core. In untreated stock casehardening resulted from a combination of tension and compression. It is apparent that as the drying progresses the moisture gradients for the treated and untreated lumber converge. Kollmann (22) found that below 60 percent relative humidity the EMC of treated and untreated stock was essentially the same. The relative_humidity dur- ing test I was dropped below 60 percent on the 23rd day. There was a significant difference between the moisture contents of the surface strips of all three treatments on the 20th day (Table 7). On the 25th day, however, there was no significant difference in the surface moisture contents of treatments A and B. After 29 days of drying, at which time the relative humidity was 31 percent, there was no signifi- cant difference between treatments A, B, or C. It appears that the moisture content of the surface layers of treated boards is not sig- nificantly different from untreated boards after the relative humidity 70 drops below a certain point. This point is dependent upon the salt concentration in the surface strips. A relative humidity of 73 percent used during the conditioning periods in this study was found to produce an equal rate of stress relief in both treated and untreated oak. This was despite the fact that the moisture content of the 210 pound treatment increased 3 per— cent during conditioning, compared to 1 percent in untreated material. If a relative humidity higher than 75 percent is used during condi- tioning, an increase in the rate of stress relief would probably occur due to the rapid increase in the EMC as the humidity increases. The danger of producing reverse-casehardening would then be greater. The length of a conditioning period must be based on a knowledge of the magnitude of casehardening stresses, as is the case in normal kiln practice. The results of this study indicate that behavior of the meas- ured casehardening strain is similar to that of the maximum surface tensile strain. If a given sequence of drying conditions produces a relatively low strain, then lowering the relative humidity conditions will increase the strain. However, if the first sequence produces a relatively high strain, then lowering the relative humidity condi— tions will not produce a significantly greater strain. The set pro- duced would be increased regardless of the stress level. An explanation for the behavior discussed above can be based on the shape of the stress-strain curve for tension and for compression. In both of these curves there is a stress level above the proportional limit at which a very small increase in stress will result in a large strain. In the case of tension a slightly greater stress will then result in tension failure. If the stresses in the board are near this 71 level, irrecoverable strain is produced rather than the elastic strain which is measured by the strip technique. After the stresses reach this value which is beyond the proportional limit, further stress in- creases cause changes in elastic strain which are so small that they cannot be detected by use of the strip technique. From the strain, set, and moisture gradients determined during a normal schedule it is apparent that these drying conditions do not approach the optimum for treated material. By properly selecting a sequence of relative humidity conditions, while using the normal dry bulb temperature sequence, the strain, mois- ture content, and indicated set gradients can be made to approach those normally deve10ped in untreated material. It is probable that the op- timum schedule is then being approached. If the revised sequence of drying conditions is set up as described, very significant reductions in drying time can be realized at no increase in drying degrade. Un- til methods of accurately determining the internal stresses and strength characteristics are perfected, the truly optimum schedule cannot be designed. Elastic strain measurements do not provide an accurate esti- mate of the surface stresses. Set measurements can be used to obtain more reliable indications of initial surface stresses, but this method is very laborious and the results are not known immediately. When the stresses are below the proportional limit, elastic strain measurements in the core may provide an accurate picture of the stress condition. The stress gradient in the core is rather flat throughout the drying qycle. Therefore, the average stress value is a good indication of the stresses at any point. The problem of travel and moisture content shrinkage which occurs during the strain measuring procedure must be 72 met, however, if accurate strain readings are to be obtained. The correction procedure used in this study was helpful in this regard. Drying Conditions and Salt Distribution When using the dry-Spread method of salt application it is im— portant that the lumber be piled carefully so that the faces are in contact. When this is done, the salt concentration on the top and bottom of the boards will be essentially the same. In order to reduce the problems which may result from the use of sodium chloride—treated lumber, the salt concentration in the center portion of the board should be kept as low as possible. As much salt as possible will then be removed during the manufacturing process. For this reason, it would be best to break down the bulk pile and begin kiln-drying imme- diately after the salt has disappeared. These studies indicate that this is not as important as has been thought. There was found to be very little salt movement between 7 and 1h days of bulk-piling. Longer periods of piling, of course, might be detrimental to maintaining a low salt concentration in the core. The final distribution of salt through a board is dependent upon the drying conditions. Sodium chloride-treated lumber dried under a normal schedule was found to have a salt concentration l/h to 1/3 as high in the core as on the surface. It has been claimed that most of the salt will surface off when the treated boards are machined. Surfacing l/Bth of an inch from each face of these boards would remove less than 1/3 of the salt. If the initial relative humidity is decreased the shell of the lumber dries out faster, and the diffusion of salt is retarded. When an initial relative humidity of 55 percent was used the salt concentration 73 in the surface strips decreased only 0.6 percent during the first eight days of drying, as compared to l.h percent during drying under a normal schedule. The final salt concentration in the center portion of stock dried according to the revised schedule was found to be much lower (See Fig. 19) than the stock dried under a normal schedule. The sur- face strips in the first case contained 62 percent of the salt in the board, while in the second case only b2 percent of the salt was found in the surface strips. The relative salt concentration at a given time in a given drying qycle seems to be constant regardless of the amount of salt applied (See Figs. 15 and 19.) There is a very striking similarity in the relative salt concentrations of the 90 and 210 pound treatments at the end of both complete drying cycles. The salt distribution in oak treated with different amounts of sodium chloride can thus be predicted if the effects of the drying conditions on the distribution are known for only one amount of salt. Further Research Further research in the area of the drying of salt-treated lumber should first be directed toward investigations of the shrink- age—salt concentration-relative humidity relationships. With this information for a particular chemical and Species, a safe yet efficient drying schedule could be designed. Before a truly optimum schedule can be designed, however, a better way must be devised for determining the internal stresses than the strip technique used in this investigation. Since it appears, at least in the case of oak, that the initial 7h tensile stresses are of primary concern, new attempts at measuring stresses might be directed toward measuring only initial surface stresses. One possible approach would be to measure the tangential strain of the entire board by the means of a clip-attached strain gage. If the moisture content gradient were determined at the same time, then the theoretical stresses could be calculated on the basis of being analogous to thermal stresses. SUMMARY Clear, straight—grained flat-sawn samples of one species, north- ern red oak (Quercus borealis Michx.), and one thickness were used throughout this investigation. It could be expected, however, that the findings would be valid for other species with similar drying characteristics. The samples in this study were treated with sodium chloride. Other salts would produce different stress, set, and mois— ture gradients at a given drying condition depending upon the sorption and antishrink properties of the particular salt. 1. As would be expected, the surface and core stress pattern was less severe for treated than for untreated stock when both were dried under a normal1 schedule. If reduced relative humidity condi- tions are properly chosen the stress development produced in treated stock will be very similar to that normally encountered in drying untreated lumber. Under these humidity conditions the treated stock will be dried safely in a much shorter time. 2. The optimum sequence of relative humidity conditions can be approached experimentally in a method similar to that used in this ex- periment. The primary limitation of this method, as far as relative humidity conditions are concerned, is that the proximity to the point of failure is not known until the failure is produced. The optimum sequence of temperature conditions must also be determined if the truly optimum schedule is to be found. In this experiment the normal lNormal conditions refer to the sequence of drying conditions recom- mended by the U.S. For. Prod. Lab. 75 76 temperature conditions were used. Near optimum temperature conditions could be obtained experimentally in a manner similar to that used for humidity conditions in this work. This would require extensive test— ing, however, and would appear to be prohibitively time consuming. A theoretical approach is needed, but cannot be obtained until the sur- face failure-stress relationship is better understood and until sur- face stress can be measured. 3. The strip technique of strain measurement was found to be a satisfactory method of comparing the stress pattern develOped by different sequences of drying conditions. It provided useful informa- tion regarding the time of stress reversal, degree of casehardening, and the relief of casehardening. It was not found to be adequate, how- ever, for indicating the stress level which results in surface—check- ing. The elastic strain obtained by the strip measurement method in- creases as the stress increases up to approximately the pr0portional limit. When the stress in a portion of the strip increases beyond the proportional limit the increase in the elastic strain is evidently too small to be detected by the strip technique. b. An initial relative humidity which is safe, yet efficient, can be predicted if shrinkage-relative humidity relationships are known for the particular salt and Species in question. This relation- ship can be used as a check on initial conditions which are determined theoretically from sorption and shrinkage data. It can also be used as a guide when determining experimentally the optimum initial condi- tions. The salt concentration at the surface of the lumber must be known if the prediction is to be accurately made. 5. Casehardening develops in sodium chloride-treated lumber 77 whether dried under a normal or relatively severe sequence of drying conditions. Casehardening in treated lumber results from a combina- tion of the antishrink effect of the salt on the surface and compres- sion set in the core. Casehardening stresses in the treated stock can be effectively removed by using a conditioning period designed for un- treated stock. The rate of stress relief is essentially the same for treated and untreated material at relatively mild humidity conditions. It is not advisable to condition treated lumber at a high relative humidity since excessive moisture pickup may occur. The maximum rela- tive humidity which can safely be used during the conditioning period can be obtained from the sorption curve of the Species and salt in question. The highest relative humidity, at which the sorption curves for treated and untreated wood have the same gradient, should be used for the conditioning period. Below this point the conditioning period will be unnecessarily prolonged. Above this point there is a danger of producing reverse casehardening. ll. 12. 13. lb. 15. BIBLIOGRAPHY American Lumberman. 1959. Arizona Dealers Solve the Lumber Checking Problem. April. Barkas, W. W. 19h9. The Swelling of Wood Under Stress. Dept. of Sci. & Ind. Dev. HMSO. Barrett, Charles. l9hh. A Critical Review of Various Methods of Residual Stress Measurement. Proc. of Soc. of Exper. Stress Analysis. 2(1). Berliner, J. F. l9hl. Seasoning and Treating Southern Lumber with Urea. Southern Lumberman. Dec. Campbell, G. S. 1959. Can Chemical Seasoning Help the Timber Industry? Forest Products Newsletter. CSIRO. Churchill, J. W. l95h. The Effect of Time, Temperature, and Relative Humidity of the Relief of Casehardening Stresses. For. Prod. Jour. h(5). Colgrove, W. H. 1955. Now You Can Cure Your Lumber. Southern Lumberman. June. Colgrove, W. H. 1956. Chemical Seasoning of Lumber. For. Prod. Jour. 6(10). Devine, J. 1960. Production Kiln Drying. DuPont Chemical Company. 1951. DuPont Crystal Urea for Chemical Seasoning of Wood. Information Bulletin. X-l7. Ellwood, E. L. l95h. Properties of American Beech in Tension and Compression Perpendicular to the Grain and Their Relation to Drying. Yale Univ. Bulletin No. 61. Ellwood, E. L., B. A. Ecklund and E. Zavarin. 1960. Collapse in Wood—Exploratory Experiments In Its Prevention. For. Prod. Jour. 10(1). Forest Products Research Laboratory. 1956. The Effect of Pre-Treatment with Urea on the Checking of Ramin During Kiln Drying. Unpublished Report. Aylesbury. Forest Products Research Laboratory. 19h8. Urea Seasoning. Unpublished Report. Aylesbury. French, W. E. 1956. Kiln Drying of Green Oak Lumber Treated with Crystal Urea. Unpublished Masters Thesis. Duke Univ. 78 16. 17. l8. 19. 20. 21. 22. 23. 2A. 25. 26. 27. '28. 29. 30. 31. 79 Freudenthal, Alfred. 19h8. Stress Analysis Beyond the Elastic Range. Proc. of Soc. of Exper. Stress Analysis. 6(2). Grossman, P. V..A. l95h. Stress Relaxation in Wood. Nature (London). 173: h392. Haygreen, J. G. 1960. The Performance of Two Types of Finishes Applied to Salt-Seasoned Lumber. Unpublished Report. Mich. State Univ. Hill, D. F. and A. L. Mottet. l9h5. A Primer of Chemical Season— ing. Chem. Seas. Progress Report #1. West Coast Lumber Assoc. Kauman, W. G. 1958. Influence of drying on Collapse in Euc. regnans. Paper No. 3, Div. of For. Prod. Tech. CSIRO. Kemp, A. E. 1959. Collapse in Aspen During Kiln Drying. For. Prod. Jour. 9(3). Kollmann, F., und A. Schneider. 1959. Sorptionmessungen an mit Salzen impragnierten Holzern. Hols als Roh und Werkstoff. 17(5). Kuebler, Hans. 1960. Drying Stresses and Stress Relief in Thin Sections of Wood. U.S. For. Prod. Lab. Report 216k. Lund, A. B. and M. A. Taras. 1959. Kiln Drying Chemically Treated Scarlet and White Oak Lumber. For. Prod. Jour. 9(11). Loughborough, W. K. 1955. A Forward Look at Chemical Seasoning of Lumber. National Hardwood Magazine. Dec. Marshall, E. D. 1951. Chemical Seasoning of Oak and Gum Lumber. Tech. Report No. 2. For. Prod. Dept. Texas Forest Service. McMillen, J. M. 1955. Drying Stresses in Red Oak. For. Proc. Jour. 5(1). Mc Millen, J. M. 1955. Drying Stresses in Red Oak: Effect of Temperature. For. Prod. Jour. 5(8). McMillen, J. M. 1957. Special Methods of Seasoning Wood-Chemical Seasoning. U.S. For. Prod. Lab. Report 1665—6. McMillen, J. M. 1958. Stresses in Wood During Drying. U.S. For. Prod. Lab. Report 1652. Mitchell, H. L., and H. E. Wahlgren. New Chemical Treatment Curbs Shrink and Swell of Walnut Gunstocks. For. Prod. Jour. 9(12). 32. 33. 3A. 35. 36. 37. 38. 39. A0. A1. A2. A3. AA. A5. A6. A7. 80 Norris, C. B. 1955. Strength of Orthotropic Materials Subjected to Combined Stresses. U.S. For. Prod. Lab. Report 1816. Olinger, H. L. 1955. Kiln Drying Urea—Treated Green Oak. Unpublished Masters Thesis. Duke Univ. Paton, J. M. 1956. The Effect of Common Salt on the Hygroscopi— city of Wood. Unpublished Report. For. Prod. Research Lab. Aylesbury. Peck, E. C. 1956. Hygroscopic and Antishrink values of Chemicals in Relation to Chemical Seasoning of Wood. U.S. For. Prod. Lab. Report 1270. Peck, E. C. 19AO. A New Approach to the Formulation of Hard— wood Kiln Schedules. Southern Lumberman. Dec. Peck, E. C., G. Baker, and R. M. Carter. 19A8. Chemical Treat- ment and Seasoning of Thick Beech Stock. U.S. For. Prod. Lab. Report R 1708. ' Perkitny, Tadeusz. 1960. Die Druckschwankungen in verschieden vorgepressten and damn starr eingeklammerten Holzkorpern. Hols als Roh und Werkstoff. 18(6). Rietz, R. C. 1950. Accelerating the Kiln Drying of Hardwoods. Southern Lumberman. July. Scott, W. W. 1939. Standard Methods of Chemical Analysis. Vol. I. Skaar, C. 1958. Moisture Movement in Beech Below the FSP. For. Prod. Jour. 8(12). Smith, H. H. 19AA. Speeded Up Schedules for Aspen Boxing and Crating Lumber. Wood Products. A9(3). Stamm, A. J. 19A6. Passage of Liquids, vapors, and Dissolved Materials Through Softwoods. U.S. Dept. Agr. Tech. Bulletin 929. Stamm, A. J. 1956. Effect of Inorganic Salts Upon the Swelling and Shrinkage of Wood. U.S. For. Prod. Lab. Report 1156. Stamm, A. J. and W. K. Loughborough. 19A2. variation in Shrink- age and Swelling of Wood. Amer. Soc. Mech. Engineers. Trans. 6A(A). Stevens, W. C. and G. H. Pratt. 1952. Kiln Operators Handbook. HMSO. Tarkow, H. and A. J. Stamm. 1960. Diffusion Through Air-Filled Capillaries. For. Prod. Jour. 10(5). A8. A9. 50. 51. 81 Torgeson, O. W. 1956. Effect of Yard-Piling and Salt Treat— ment on Checking of 5/A Red Oak Lumber. U.S. For. Prod. Lab. Report 1759. U.S. For. Prod. Lab. 19A6. Kiln Certification. ANC Bulletin 21. Youngs, R. L. 1957. The Perpendicular-to-Grain Mechanical Properties of Red Oak as Related to Time, Temperature, and Moisture Content. U.S. For. Prod. Lab. Report 2079. Youngs, R. L. and C. Norris. 1958. A Method of Calculating Internal Stresses in Drying Wood. U.S. For. Prod. Lab. Report 2133. APPEN DIX STATISTICAL TABLES 82 83 TABLE 5. Analysis of Variance of the Maximum Tension Strain in the Surface Strips During Test I Source of variation Degrees of Sum of Mean Square F F 05 Freedom Squares Total 11 15.016 Treatments (A, B, c) 2 12.59A 2.996 32.67% 14.16 Surface (T0p, Bottom) 1 0.880 0.880 A.57 5.32 T‘x S 2 0.761 0.380 2.92 5.1A Within 6 0.781 0.130 Combined EMS 8 1.5A2 0.193 *Indicates significance at the 5% level. From a Studentized Range Test Treatments: A > B > C Summary Treatments Means A 3.75 x 10'3 in./in. B 2.69 x 10'3 in./in. c 1.25 x 10‘3 in./in. 8A TABLE 6. Analysis of variance of the Maximum Tension Strain in the Surface Strips in Tests II, III, and IV' Source of variation Diggzzszf Sgfiagis Mean Square F F-05 Total 35 6.686 Treatments 2 2.212 1.106 9.15% 3.32 (90, 150, 210 lbs.) Tests (II, III, IV) 2 0.911 0.171 1.02% 3.32 Surfaces (Top, Bottom) 1 0.0156 0.0156 0.133 1.17 Tr x Te 1 0.155 0.111 1.16 2.93 Tr x S 2 0.761 0.380 2.87 3.55 Te.x S 2 0.0938 0.0169 0.60 3.55 Te x Tr ng 1 0.802 0.201 2.58 2.93 Within 18 1.106 0.0781 Combined EMS 30 3.517 0.117 *Indicates significance at the 5% level. From a Studentized Range Test Treatments: 90 lbs. >-150 lbs. & 210 lbs., 150 lbs. = 210 lbs. Tests: 11 > III, IV = 111 = II Summary Treatment Means ,, Test Means 90 lbs. 2.71 x lO’inn./in. II 2.56 x 10‘5 in./in. 150 lbs. 2.27 x 10'3 in./in. 111 2.17 x 10'3 in./in. 210 lbs. 2.13 x 10-3 in./in. IV 2.37 x 10-3 in./in. 85 TABLE 7. Analysis of variance of the Moisture Content of the Surface Strips After 20 and 25 Days of Test I 20 DAYS Source of variation Degrees of Sum of Mean Square F F.05 Freedom Squares Total 11 313.9 Treatment (A, B, C) 2 '295.A 1A7.7 70* A.26 Within 9 18.52 2.06 *lndicates significance at the 5% level. From a Studentized Range Test: A < B < C Summary: Treatment Mean A 17.1% MC B 23.2% MC C 29.5% MC 25 DAYS Source of variation Degrees of Sum of Mean Square F F.05 Freedom Squares Total 11 93.3 Treatment (A, B, C) 2 73.6 36.8 16.9% A.26 Within 9 19.7 2.18 From a Studentized Range Test: A, B < C, A = B Summary: Treatment Mean A 11.8% MC B 13.0% MC C 17.5% MC 86 TABLE 8. Analysis of variance of the Moisture Content of the Surface After Two Days in Tests II, III, and Iv Source of variation D:g::::mof Séflaggs Mean Square F F°°5 TOtal 35 1260.A Tests (II, III, IV) 2 970.1 A85 77* 3.31 Treatments 2 9A.99 A7.5 7.5% 3.31 (90, 150, 210 lbs.) T X T h A1.AA 10.36 1.82 3.31 Within 27 153.9 5.70 Combined EMS 31 195.3 6.30 *Indicates significance at the 5% level. From a Studentized Range Test Treatments: 90 lbs. < 150 lbs.& 210 lbs. 210 lbs. = 150 lbs. Tests: II > III > IV Summary Treatment Means Tests Means 90 lbs. 30.3% II 38.1% 150 lbs. 33.8% 111 31.1% 210 le. 33.8% IV’ . 25.7% 87 TABLE 9. Analysis of variance of Indicated Tension Set in the Surface Strips After TWO Days of Drying in Tests I, II, III, and IV Degrees of Sum of Mean Square Source of variation Freedom Squares F F.05 Total 39 13.385 Treatments 1 .621 .621 2.971 A.l3 (90, 210 lbs.) . Tests 1 6.216 1.551 7.135% 2.65 (I,II,III,Iv,I—Untr.) T‘x T 1 .791 .198 .95 2.69 Within 30 6.313 .210 Combined EMS 31 7.107 .209 *Indicates significance at the 5% level. From a Studentized Range Test .Tests: I & II <:I-Untreated & IV & III I = II I - Untreated = IV = III Summary Tests Means I .01% change in % shrinkage II .15% change in % shrinkage III .6A% change in % shrinkage Iv 1.03% change in % shrinkage I-Untr. .85% Change in % shrinkage 88 TABLE 10. Analysis of variance of Percent Tangential Shrinkage of the Entire Board Degrees of Sum of Mean Square F ' F.05 Source of variation Freedom Squares Total 23 7-099 Treatments 5 A.781 .956 7.Al* A.6O Within 18 2.318 .129 *Indicates significance at the 5% level From a Studentized Range Test (1 - 210) <:(1 - Untreated), (v - 15), (I - 90), (v — 90) (I - 210) = (v — 210) (I - Untreated) > all the other treatments Summary Test Lbs. of Salt Mean % Tangential Shrinkage I O 7.65 % I 90 7.05 % I 210 6.20 % v 15 6.96 % v 90 7.21 % V 210 6.71 % 89 TABLE 11. Analysis of Variance of the Relative Salt Concentration in the Surface Strips After Different Periods of Bulk Piling Source of variation Degrees of Sum of Mean Square F F.05 Freedom Squares Total ‘ 23 3399.5 Surfaces 1 3.60 3.60 .02 1.38 (TCp, Bottom) Times 2 313.6 156.8 1.03 3.52 (7, 11, 11 days) Treatments 1 188.7 188.7 1.21 1.38 (90, 210 lbs.) S x Ti 2 53.57 26.8 .11 3.88. Ti x 1r 2 51.91 27.5 .15 3.88 Tr x S l 61.13 61.1 .31 1.75 S x Tr x Ti 2 551.90 275.9 1.52 3.88 Within 12 2172.1 181.0 Combined EMS 19 2893.6 152.3 Summary Times Mean Relative Salt Concentration of Surface Strips 7 days 69-7% 11 days 71.1 % 11 days 77.7 % 7;” y 1181:“:4'3458453