$0111: CHROMATGGRAPHIC STUDIESCFTRE; PREROLR': RRACRRN OFHARnwgogggmm V : ' Thesis for the Degreeiof‘Ph D MECHTGAN STATE UNIVERSITY ROBERT e. HARPER f ;:-. ‘ * ’* ij‘ 13R7 LIBRA 1» Y “ TanVb . . Michigan c A” University ; This is to certify that the thesis entitled SOME CHROMATOGRAPHIC STUDIES OF THE PHENOLIC FRACTION OF HARDWOOD SMOKE presented by Robert G. Harper has been accepted towards fuifillment of the requirements for Ph.D. Food Science degree in 1.426147%, -' I Major profe iDau:OCtOber 5, 1967 0-169 ABSTRACT SOME CHROMATOGRAPHIC STUDIES OF THE PHENOLIC FRACTION OF HARDWOOD SMOKE by Robert G. Harper There is considerable evidence that smoke phenols are primarily re- sponsible for some desirable properties imparted to food products by the application of wood smoke. These investigations were undertaken to make some qualitative studies of an exploratory nature concerning the influence of various factors on the phenolic fraction of hardwood smoke. Procedures for collecting, extracting, and analyzing the phenolic fraction of various samples were developed through systematic experimentation with guidance from published studies of a related nature. Gas-liquid, thin-layer, and paper chromatographic techniques were utilized for the qualitative analyses of free smoke phenols. The use of thin-layer and paper chromatography was complemented by the formation of the characteristically colored Ernitrophenylazo dyes on the deve10ped chromatograms. These dyes permitted the identification of incompletely separated phenols. A combination of the chromatographic methods permitted the identification of the following phenols: phenol, grcresol, catechol, resorcinol, hydroquinone, pyrogallol, guaiacol, 4-methylguaiacol, 4- ethylguaiacol, 4-allylguaiacol, 2,6—dimethoxyphenol, 2,6-dimethoxyh4- methylphenol, 2,6-dimethoxya4-ethylphenol, 2,6-dimethoxyh4-allylphenol, and 3-methoxycatechol. The latter two had not been identified in wood smoke previously. The di— and trihydroxy phenols could not be chromato- graphed by the gas-liquid chromatography procedures used and thus were analyzed by thin-layer and paper chromatographic techniques only. Robert G. Harper Qualitative comparison studies were made of the phenolic fractions of the following: (1) samples of whole smoke and vapor phase smoke produced in a laboratory model smoke generator (equipped with an electrostatic precipitator) and collected by different systems; (2) smoke samples collected by model absorbent systems (water-filled artificial casings) in a commer- cial Smokehouse operated under various conditions of relative humidity and temperature; and (3) three commercial "liquid smoke" solutions. The comparisons were based on chromatographic analyses of the various samples as follows: (1) relative quantities of guaiacol, 2,6-dimethoxya phenol, and 2,6-dimethoxyh4-methylphenol in each sample determined from gas-liquid chromatography peak areas; (2) visual assessment of paper chromatograms with samples applied in approximately equal "total phenol" quantities to detect gross differences in amounts of di- and trihydroxy phenols present; and (3) visual assessment of thin-layer chromatograms to substantiate findings of (l) and (2). The results of the whole smoke vs. vapor phase study indicated that the removal of the particulate phase of smoke had essentially no effect on the qualitative profile of its phenolic fraction. Therefore, it was postulated that the vast majority of smoke phenols were confined to the vapor phase. In this study it was also noted that the samples collected by different systems had considerably different phenolic fractions. The results of the relative humidity study suggested that this para— meter had little, if any, influence on the qualitative nature of the phenolic fraction of smoke samples collected by the model systems. No Robert G. Harper notable difference existed between the samples collected in two different types of artificial casings used in this study. Samples collected inside the casings differed very little from.those obtained by washing the out- side of the casings with water. The results of the temperature study indicated that as the smokehouse temperature was decreased, the relative quantities of the higher boiling phenols decreased. This suggests that the higher boiling phenols may have been condensing out of the vapor phase at the lower temperatures to some extent. The comparison of the phenolic fractions of the three commercial "liquid smokes" revealed no gross differences among them. This suggested that they were prepared similarly. SOME CHROMATOGRAPHIC STUDIES OF THE PHENOLIC FRACTION OF HARDWOOD SMOKE By H . J 101' 9 Robert G1“Harper A THESIS Submitted to ‘Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR.OF PHILOSOPHY Department of Food Science 1967 G LPSLHA 33%(92 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Professor L. J. Bratzler for his thoughtful guidance and encouragement throughout the course of this study and for his assistance in the preparation of this manuscript. He also wishes to thank Dr. B. S. Schweigert, Chairman of the Depart— ment of Food Science, and the members of his guidance committee, Drs. A. M; Pearson, J. F. Price, R. C. Nicholas, and E. J. Benne for their interest and stimulation. To his wife, Betty Jo, and daughter, Sara, the author is especially grateful for their sacrifices and for making it all worthwhile. ii TABLE OF CONTENTS Page INanCTION O 0 O O O O 0 O O O O O O O O O O O O - O O O O O O O O 1 REVEW OF LITERATURE O O O O O O O O O O O I O O O O O O O O O O O 3 WOod smoke research in general (as applicable to food proceSSing) O O O O O O O O O O - O O O O O O O O O O O O O O I 3 Characteristic properties of phenols . . . . . . . . . . . . 3 Origin of phenols of wood smoke . . . . . . . . . . . . . . . 7 Factors affecting the phenol content of smoke . . . . . . . . 12 Type Of WOOd O O O O C O O O O O O O O O O O O O O O O O 12 Conditions of smoke production . . . . . . . . . . . . . 13 Temperature of destruction . . . . . . . . . . . . 13 Temperature of oxidation and air availability . . . l4 Moisture content of wood amd relative humidity of smoke atmosphere . . . . . . . . . . . . . . . . . 15 Friction vs. smoldering-type generators . . . . . . 16 Electrostatic precipitation . . . . . . . . . . . . 16 Contribution of phenols to the desirable properties of smoked fOOdS O O O O O O O O C O O O O O O O O O O O O O 0 O O O O O 18 Organoleptic properties . . . . . . . . . . . . . . . . 18 Antiseptic and germicidal properties . . . . . . . . . . 19 Antioxidant properties . . . . . . . . . . . . . . . . . 20 Desirable "finish" or gloss and color . . . . . . . . . 22 Qualitative studies of phenols in wood smoke . . . . . . . . 23 Studies indicating relative quantities of individual phenols in WOOd smOke C O O C O O O O O O O O O O O O O O O O O O O O 26 EHERDENTAL PRO CEURE O O O O O O O O O O O O O O O O O O O O O 0 27 Sawdust source and condition . . . . . . . . . . . . . . . . 27 Whole smoke vs. vapor phase study . . . . . . . . . . . 27 Relative humidity and temperature studies . . . . . . . 27 iii Smoke production . . . . . . . . . . . . . Whole smoke vs. vapor phase study . . . . Relative humidity and temperature studies Collection of smoke samples . . . . . . . . . . Whole smoke vs. vapor phase study . . . . Relative humidity and temperature studies Estimation of total phenols . . . . . . . . Extraction of phenolic fraction . . . . . . Chromatographic analyses . . . . . . . . . Gas-liquid chromatographic procedure . Thin-layer chromatographic procedure . Paper chromatographic procedure . . . Procedures for comparing the commercial "liquid SOlutionS O O O O O O O O O 0 0 O O O O O O O O RESULTSANDDISCUSSION............... Estimation of total phenols . . . . . . . . . . Separation and identification of smoke phenols Gas-liquid chromatography (GLC) . . . Thin-layer chromatography (TLC) . . . Paper chromatography . . . . . . . . . . . Qualitative comparison studies of the phenolic fractinns of various smoke samples by chromatographic techniques . . . . Whole smoke vs. vapor phase study . . . . . . . . . . Relative humidity and temperature studies Qualitative comparison of the phenolic fractions of three 0 O O I O O 0 commercial "liquid smokes" . . . . . . . . . . . . . . . SWY O O O O O O O O O O O O O O O O O O O 0 iv Page 28 28 29 3O 3O 32 33 36 39 4O 43 45 47 48 48 52 52 56 61 66 66 78 88 92 Page BIBLIOGRAPHY O O I O O O O O O O O O O O O O O O O O O O O O O O 97 APPEmIX 0 O O O O O O O O O 0 O O O O O O O O O O O I O O O O O 10 4 Table II III VI VII VIII IX LIST OF TABLES Page GLC retention times of separated components of the phenolic fraction of the unknown (Unk) smoke sample and of probable corresponding known (Kn) compounds . . . 55 Phenols identified in maple sawdust smoke samples by chromatographic techniques . . . . . . . . . . . . . . . 65 Estimates of total phenols in samples collected for whole smoke vs. vapor phase smoke comparisons . . . . . . . . 67 Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6-DMP), and 2,6-dimethoxyh4-methy1phenol (2,6-DMA4-MP) in samples of whole smoke (WS) and vapor phase (VP) . . 71 General data for smoke samples collected in water-filled artificial casings for the relative humidity study (Temperature-83°C) . . . . . . . . . . . . . . . . . . . 79 Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6-DMP) and 2,6-dimethoxya4-methy1phenol (2,6-DM94-MP) in samples collected in water-filled artificial casings for the relative humidity study (Temperature-83°C) . . . . . 80 Average data from.dup1icate samples obtained by washing the outside of Nojax casings used in the relative humidity study (Temperature-83°C) . . . . . . . . . . . . . . . . 81 General data for smoke samples collected in water-filled Nojax casings for the temperature study (Relative humidity -30%) . . . . . . . . . . . . . . . . . . . . . . . . . 87 Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6—DMP), and 2,6—dimethoxyh4-methylphenol (2,6—DM94eMP) in samples collected in water-filled ijax casings for the temperature study (Relative humidityz30%) . . . . . 87 vi LIST OF FIGURES Figure Page I Hypothetical structure of lignin with probable sites of thermal cracking (Goos, 1952, with modifications) . . . 10 II Scheme for extracting phenols from.the collected smoke samle O O O O O O O O O O O O O O O O O O O O O O O O O 37 III Formation of indophenols vs. time at 38°C by Tucker's method (average of duplicate runs) . . . . . . . . . . . 49 IV Standard curves for phenol, guaiacol, and 2,6-dimethoxyh phen01 O O O O O O O O O O O O O O O O O O O O O O O O O 50 vii Plate II III VI VII VIII IX LIST OF PLATES Composite GLC chromatogram.showing the separation of components in the phenolic fraction of a NaOH-collected whole smoke sample . . . . . . . . . . . . . . . . . . . Thin-layer chromatograms of guaiacol and 2,6-dimethoxyh phenol and their para-substituted derivatives run individually and as mixtures . . . . . . . . . . . . . . TLC chromatogram.of a mixture of known phenols and four phenolic fractions of unknown smoke samples . . . . . . Paper chromatogram.of known mono-, di-, trihydroxy phenols run separately and as a mixture . . . . . . . . Composite GLC chromatogram (Column A) of phenolic fractions of whole and vapor phase smoke collected in 1N NaOH at 2-300 0 o o o o o o o o o o o o o o o o o o 0 Paper chromatogram.of phenolic fractions of whole and vapor phase smoke samples collected by three different methOdS O O O O O O O O O O O O O O O O O O O O O O 0 O TLC chromatogram.of phenolic fractions of whole and vapor phase smoke collected by two different methods . . Composite paper chromatogram.of the phenolic fractions of smoke samples collected in water-filled NOjax casings and those obtained by washing the outside of the casings Paper chromatogram of the phenolic fractions of smoke samples collected in water-filled Clear-Zip casings for the relative humidity study. . . . . . . . . . . . . . . Paper chromatogram.of the phenolic fractions of smoke samples collected in water-filled NOjax casings for the temperature study . . . . . . . . . . . . . . . . . . . Paper chromatogram of the phenolic fractions of three commercial "liquid smoke" solutions . . . . . . . . . . viii Page 54 58 60 64 7O 74 77 83 86 9O 93 INTRODUCTION For many centuries wood smoke has been applied to food products to improve storage stability and to enhance organoleptic properties. Howe ever, only in the past three decades have research efforts been directed towards obtaining information on wood smoke composition, production factors influencing composition, and specific components responsible for the desirable properties imparted to food products. Scientists in this country have devoted relatively little attention to this area of study, but European workers have made numerous and significant contributions. WOod smoke is physically described as being an aerosol composed of a particulate phase susPended in a vapor phase. The smoking process is now known to be essentially one of vapor absorption and not one of particle deposition as was originally assumed. Different studies have shown that food products smoked with vapor phase alone were similar in both storage and organoleptic qualities to those smoked normally. Chemically, wood smoke is very complex and contains many different types of compounds which may vary considerably under different production parameters. The three classes of smoke compounds which are considered to be of significance in smoked foods are the acids, carbonyls, and phenols. The smoke phenols, though present in the lowest concentration, are considered by many scientists to be primarily responsible for the desirable effects of wood smoke on food products. -2- The objectives of this study were: 1. To make qualitative comparisons of the phenolic fractions of whole smoke and vapor phase smoke. 2. To make qualitative comparisons of the phenolic fraction of whole smoke collected by model systems in a commercial smokehouse operated under various conditions of relative humidity and temperature. 3. To make qualitative comparisons of the phenolic fractions of three commercial "1i uid smoke" solutions. q REVIEW OF LITERATURE Wood Smoke Research in General as a licable to food rocessin Excellent reviews by Wilson (1961b) and Draudt (1963) along with a brief summary by Foster (1959) reveal the current status of the knowledge and understanding of both the technology involved in the smoking process and the chemical composition of wood smoke. Since the present study is concerned with the phenolic fraction of wood smoke, this literature review will cover only those publications which relate to phenols - particularly those of wood smoke - or to desirable properties of smoked foods which are contributed by the wood smoke phenols. For the sake of consistency, one of the most prevalent phenolic compounds of wood smoke, which appears in the literature many times named both as 1,3-dimethylpyrogallol ether and 2,6—dimethoxyphenol, will be referred to as 2,6—dimethoxyphenol in this thesis. Characteristic Prgperties of Phenols Phenols are aromatic hydroxy compounds in which the hydroxyl group(s) is (are) attached directly to the aromatic (benzene) ring. According to Geissman (1959), this class of organic compounds bears a formal resem— blance to alcohols and even undergoes some reactions (formation of ethers and esters) in common with alcohols; however, because of the profound effect of the aromatic ring on the pr0perties of the hydroxyl group and vice versa, it is not proper to regard them merely as aromatic alcohols. Compared with corresponding saturated alcohols, phenols are appreciably -3- more acidic, more soluble in water, and have higher boiling points. These differences are illustrated in the data below taken from Roberts and Caserio (1964). These authors stated that the higher boiling point and Property Phenol Cyclohexanol Melting point 43°C 25.5°C Boiling point 181°C 161°C H20 solubility (g/100 g, 20°C) 9.3 3.6 K 1.0 x 10-10 ~10-18 a greater solubility could be eXplained by the stronger acidity of the phenolic hydroxyl group as compared with the alcoholic hydroxyl group. This greater acidity is paralleled by a more strongly polarized bond between the oxygen and hydrogen of the undissociated hydroxyl group (D- -*H+) resulting in the formation of stronger intermolecular hydrogen bonds with either water molecules (increasing solubility) or other phenolic molecules (increasing boiling points). Morrison and Boyd (1959) stated that most phenols are essentially insoluble in water and have acidity characteristics which place them in the following relationship with some other compounds in order of decreas- ing acidity: carboxylic acids ---> carbonic acid ---> phenols ---> water ---> alcohols. They also noted that aqueous hydroxides convert phenols into salts (very water-soluble but insoluble in organic solvents) and that aqueous mineral acids convert the salts back into the free phenols which have Opposite solubility characteristics to the salts. Since most phenols are weaker acids than carbonic acid, they do not dissolve in aqueous bicarbonate solutions as do carboxylic acids. Accord- ing to Morrison and Boyd (1959), the acidity characteristics of phenols and the water solubility of their salts are very useful tools in their analysis and separation. As an example, they stated that slightly-water- soluble substances which dissolve in aqueous hydroxide but not in aqueous bicarbonate must be more acidic than water and less acidic than carboxylic acid; most compounds in that acidity range are phenols. These authors concluded, therefore, that phenols may be separated from non-acidic com- pounds by means of their solubility in hydroxide base and from carboxylic acids by means of their insolubility in bicarbonate. The simplest member of the phenolic family of compounds is phenol which is composed of one hydroxyl group attadhed to an aromatic ring. More complex members may have more than one benzene ring, more than one hydroxyl group, and an almost unlimited variety of substituent groups (including alkyls, alkyl ethers, carboxyls, aldehydes, nitrates, and halogens) attached to the aromatic ring(s). Since the hydroxyl group is ggihp and para,directing (Morrison and Boyd, 1959; Geissman, 1959), most substituents are attached to either the No. 2 and No. 6 carbons (orthg positions) or to the No. 4 carbon (para postion) of the phenol ring which is very susceptible to electr0philic substitution reactions. In this thesis only those phenols with one aromatic ring and composed entirely of carbon, hydrogen, and oxygen will be considered, and therefore, the substituent groups will be comprised mainly of alkyl (saturated hydro- carbon), alkene (unsaturated hydrocarbon), alkyl ether, aldehyde, and carboxylic acid grOUps. As was noted by Morrison and Boyd (1959), several phenols and their ethers are isolated from the essential oils of various plants. They listed as examples - eugenol from cloves, vanillan from vanilla beans, and thymol from either thyme or mint. Indeed, most of the members of the phenolic class of compounds have a characteristic odor. Amerine, Pang- born, and Roessler (1965) reviewed odor classifications in which guaiacol was the standard for "burnt" odor and vanillan the standard for "sweet" or "fragrant" odor. Most phenols are quite easily oxidized (Morrison and Boyd, 1959) and this property may account for their widespread use as antioxidants. Stuckey (1962) stated that most natural and synthetic food-grade antioxi- dants belong to the phenolic class of compounds with the most common ones being hydroquinone, butylated hydroxyanisole (ERA), and butylated hydroxytoluene (BHT). An important chemical property of some phenols is their ability to react with several nitrogen containing compounds to form dyes or colored compounds(Gibbs, 1927; Randerath, 1964; Stanley ggnal., 1965) which are useful in the study of phenols both for total quantitation and for the identification of individual compounds after they have been separated by either paper or thin-layer chromatography. Structural formulae, molecular weights, and other data that are available for the phenols whose identi- fication in wood smoke has been reported in the literature are presented in Appendix I. Origin of Phenols of WOod Smdke Goos (1952) revealed the complexity of wood as a chemical substance by listing over 200 separate compounds that have been identified in the products of its destructive distillation. However, as was mentioned by Draudt (1963) these same compounds may not all exist in smoke since the products of heating wood depend considerably on the conditions under which it is heated. Goos (1952) stated that although the proportions of the three major components of wood-cellulose, hemicellulose, and ligninumay vary depend- ing on the kind of wood and the method of analysis, it may be assumed that approximately 1/2 of wood is cellulose, 1/4 hemicellulose, and 1/4 lignin. These three components are all located in the cell walls of wood. White SE 31. (1959) stated that cellulose is a polysaccharide having a linear configuration made up of repeating units of the disaccharide cellobiose (two glucose units joined by aléLl,4 glucosidic linkage) and is the most abundant organic compound in the world, comprising over 50 percent of all carbon in vegetation. According to Goos (1952), the pyrolysis of cellulose results in acids of the acetic series and minor amounts of furan derivatives and phenols. Hemicellulose, or the non-cellulose polysaccharide component of the wood cell wall (Browning, 1952), consists largely of pentosans in hard- wood. Goos (1952) noted that pentosans are the least heat stable of the wood components and yield higher quantities of acids of the acetic series than either cellulose or lignin. Furfural might be eXpected to be a major distillation product of pentosans but has been found in quite low amounts, probably due to its extensive reactivity to form other compounds (Goos, 1952). Lignin is generally considered to be the precursor of the vast major- ity of phenols formed by the thermal decomposition of wood (Tilgner ggual., 1962d). Because the structure of lignin has not been determined (Brauns and Brauns, 1960), a good definition of lignin is difficult. Based on the existing knowledge of lignin chemistry, Brauns (1952) defined lignin as the incrusting material of plant tissue which is built up mainly, if not entirely, of phenylprOpane building stones and which contains the major portion of the methoxyl groups in plants. Lignin is unhydrolyzable by acid, easily oxidizable, soluble in hot alkali and bisulfite, and condenses readily with phenols and thio compounds. Brauns and Brauns (1960) noted that lignin yields vanillan and syringaldehyde when oxidized with nitro- benzene. According to Goos (1952), the characteristic or key compounds obtained from the thermal decomposition of lignin, either isolated or in wood, are the phenolic ethers typified by guaiacol in the case of softwoods and both guaiacol and 2,6-dimethoxyphenol in hardwoods. Discussions of the various proposed structures for the ”building blocks" that make up the obviously complex structure of lignin have been made by Brauns (1952) and Brauns and Brauns (1960) along with theories that have been offered in support of the various preposed structures. Goos (1952) presented a hypothetical structdre or "approximate picture" of the various groups that make up the complex lignin molecule which was taken from the work of Brauns (1948). This hypothetical structure is presented in Figure I with a few clarifying modifications based on statements made by Goos (1952) concerning the possible ways in which the lignin molecule might "crack" or thermally decompose and also based on evidence presented by Brauns (1948) concerning the apparent differences in the composition of lignin from softwood and from hardwood. Goos (1952) stated that although quite an array of phenols and phenolic ethers have been identified in the pyrolysis products of both lignin and wood, there is a scarcity of good quantitative estimates of the proportions of each due to the great difficulty in separating and analyzing the mixture. However, he thought it was safe to say that the preponder- ance of phenolic material found consisted of guaiacol and of 2,6-dimethoxy- phenol and their derivatives. The substituent groups are largely methyl, ethyl, propyl, vinyl, and propenyl and it is especially notable that the side chains do not exceed 3 carbon atoms and are nearly always in the para position to the hydroxyl (Goos, 1952). Goos (1952) used the "approximate” picture of the lignin molecule to speculate about the presence of the typical phenolic compounds both qualitatively and quantitatively, and stated that for this purpose it was not necessary to assume the correctness of the structure nor to attempt to evaluate it in terms of other pr0posed structures. In the hypothetical lignin structure (Figure I) the aromatic rings (I through V) will be relatively stable to heat and the scission of bonds would likely occur at the locations marked by the dotted lines A, B, C and D -10- . .Amcofiuwo E u «mma .mooov wcwxoouo Hashes» mo mouflm «Hammond £uH3 awawfla mo ounuusuum HWWMWMSuMQMM .H ousmfim ammo Om Axe -11- (Goos, 1952). The oxygen atoms attached to aromatic rings would remain attached because this linkage is relatively strong. If cleavage did occur as indicated by the dotted lines, guaiacol rings would be detached with incomplete or unstable three-carbon side chains para to the hydroxyl group for aromatic rings I through IV. The side chain could break in various ways, and with sufficient hydrogen might form methyl, ethyl, pr0pyl, vinyl or allyl groups and the oxygen would be eliminated as water. If hydrogenation did not occur, an aldehyde fragment could form which might react to form a high boiling tar. Aromatic ring V might be expected to yield phenol and cresols. These possibilities are applicable to lig- nin from softwoods. In lignin from hardwoods where there are an average of two syringyl groups and two or three vanillyl groups per building unit (Brauns, 1948), 2,6-dimethoxyphenol derivatives might be formed. A syringyl group differs from a vanillyl group by having an additional methoxy group in the 6 position (HO-l, CH30-2,6) and this can occur in only two rings of the assumed lignin structure, namely rings I and IV at the positions marked (x) and (y). The detachment of these dimethoxyphenol rings would form 2,6 dimethoxyphenol or its.2fl£§ side chain derivatives in a manner similar to the formation of guaiacol and its derivatives from vanillyl groups. Results of studies by Creighton ggJal. (1944) and Bailey (1947) support the findings of Brauns (1948) in reference to the basic differ- ences in the makeup of lignin from softwood and from hardwood and the type of phenolic compounds which arise from each. Since lignin has never -12- been pyrolyzed to a single predominant compound analogous to the conver- sion of cellulose to levoglucosan, the theory that the lignin building group is composed of various related simple units rather than repeated identical units appears to be true (Goos, 1952). Factors Affecting the Phenol Content of Smoke Type of Weed While the lignin content of softwoods is generally higher than that of hardwoods (Browning and Isenberg, 1952), the methoxyl content of hard- woods is as high or higher than that of softwoods. These authors stated that though this may seem to be a contradiction, since lignin contains most of the methoxyl groups in wood, it can be explained by the fact that the methoxyl content of hardwood lignin is about 40% higher than that of softwood lignin. 0n the basis of this, hardwoods might be expected to yield more 2,6 dimethoxyphenol derivatives than guaiacol derivatives as was postulated by Goos (1952) based on evidence given by Brauns (1948). In softwoods the lignin content is generally lower in the heartwood than in the sapwood but this generalization is apparently not possible in the case of hardwoods (Browning and Isenberg, 1952). Freeman and Peterson (1941) reported results which indicated the lignin content varied only slightly between the heartwood and sapwood of the two hardwood species, beech and maple, with the higher content being in the heartwood. -13- Conditions of Smoke Production The composition of a wood smoke aerosol is known to be influenced by the temperature of destruction, the temperature of oxidation, and the rate of oxidation (Wilson, 1961a). Temperature of destruction. Hawley (1952) stated that the thermal decomposition of wood becomes exothermic at about 280°C. According to Kuriyama (1962), the thermal decomposition of wood occurs in four stages, based on temperature, and each of the stages is distinctly exothermic. The first stage of thermal decomposition occurs after the wood has been completely dried and has reached a temperature of ZOO-260°C. In this stage the hemicelluloses are broken down to yield acids, methanol, and the gases 002 and CO. Above 260°C the second stage begins and cellulose is broken down, yielding large amounts of pyroligneous liquor (containing acids and aldehydes) and gas. In addition, the side chains of lignin may be broken down in this stage. When the temperature has reached about 310°C the pyrolysis of cellulose is completed but there is very little breakdown of lignin. The breakdown of lignin occurs during the third stage at temperatures between 310 and 500°C. In this stage the largest quantity of wood tar is formed along with the gases CO, CH4, and H2, and the vast majority of phenols. In the final stage, which occurs at temperatures above 500°C, only gas (mainly H2) is produced and it is very likely that aromatic condensation is taking place. Kuriyama (1962) concluded that the major breakdown of the wood components occurred at 250°C for hemicelluloses, 300°C for celluloses, and 400°C for lignin f- - . , _, ._ --,_______-_-,. "1-111..“— -14- since these temperatures coincided well with the phenomena of maximum temperature elevations due to exothermic heat and maximum increases in the generation of distillates and gas. Tilgner a; 31. (l962d), utilizing a two-stage generator in which the temperatures of the destruction and the oxidation chambers could be controlled separately, found a much higher phenol content in smoke pro- duced at a destruction temperature of 400°C than in smoke produced at 300 or 350°C. Porter ggual. (1965) produced smoke by shaking sawdust onto a hotplate and reported the total phenol content at hotplate tempera- tures of 322, 355, and 386°C to be about double that produced at 252 or 287°C. Temperature of oxidation and air availability. Miler (1962) postu- lated that there are three separate zones in the process of smoke formation. In the inner zone, or ”foam zone”, the wood undergoes thermal decomposition in an essentially oxygen-free atmosphere. In the middle zone, or "fizz zone", the volatile decomposition products are escaping from the surface of the heated wood. In the outer zone, or ”diffusion zone", the volatile products are combusted in the presence of oxygen. He further stated that the temperature in the outer zone where combustion is taking place may reach as high as 700-lOOO°C and that some of this heat is transmitted back to the inner zone, thus accelerating the destruction process. According to Draudt (1963), the smoke aerosol is probably formed somewhere outside the combustion zone where the temperature is nearer 300°C. -15- Noting that the supply of oxygen present during the oxidation stage of smoke production may have considerable effect on the resulting products, Tilgner a; a1. (1962d) studied the oxidation process using different temperatures and supplying different quantities of air to the oxidation chamber of the two-stage generator. These workers devised a measure of air quantity termed ”air surplus factor” which denoted the ratio of the amount of air supplied to that amount which was theoretically calculated as being required for the complete combustion of a given quantity of wood. They concluded from their studies that an oxidation temperature of 200°C and an air supply corresponding to an "air surplus factor" of 8 resulted in the production of the greatest quantity of phenols in the smoke. Moisture content of wood and relative humidity of smoke atmosphere. Jahnsen (1961) found a decrease in the phenolic content of hickory-sawdust smoke as the moisture content was increased from 10 to 50%. He stated that this suggested a dilution effect. TilgnerEEEHal. (1962f) reported that the smoke produced from beech sawdust in a smoldering-type generator contained smaller quantities of phenols as the moisture content of the sawdust was increased from 6.9 to 52.5%. These workers also reported that the amount of steam—volatile phenolic compounds was slightly higher in herring processed with smoke from ”dry" sawdust compared to those processed with smoke from "wet" sawdust. Simon 25.2l: (1966) found that artificial casings filled with water collected less total phenols as the relative humidity in the smokehouse -l6- was increased through the range normally employed in commercial practice. This decrease with increasing humidity was to a lesser extent, however, than in the case of acids and carbonyls. Friction vs. smoldering-type generators. Husaini and CooPer (1957) reported that the phenol content of smoke produced by a friction-type generator (a metal drum or disc rotating against the end grain of a block of wood) was higher than that of smoke produced by a smoldering sawdust generator. They obtained 3 times more steam-volatile phenols and 9 times more non-steam-volatile phenols with the friction generator. Tilgner ggual. (1962a) found 1.2 times more total phenols in beechwood smoke that was produced by a friction generator as compared to smoke from a smolder- ing—type generator. Electrostatic precipitation. According to Sikorski (1962) many investigations have been made in reference to the possible applications of a high voltage electrostatic field for precipitating dispersed parti- cles from aerosols such as common industrial clouds, smokes, and dusts. He mentioned that, in addition to the well known cleaning effect of the corona discharge, quite a different effect has been observed, i.e., the catalytic influence of the high voltage electrostatic field on chemical reactions of many compounds, especially organic compounds. Pettet and Lane (1940), in discussing the problems of smoke sampling for analytical purposes, seemed opposed to the electrostatic precipitation method be- cause of the possibility of chemical changes in the smoke constituents in the high voltage electrostatic field. -17- Sikorski (1962) used a smoldering-type laboratory generator equipped with an electrostatic precipitator to study the effect of the high voltage electrostatic field on the components of smoke from beech sawdust. The smoke produced in the generator was divided into two equal streams with half of it passing through the electrostatic precipitator and the other half going directly into the collection vessels and serving as a control. He reported that the total condensates (vapor and particulate phases combined) of the smoke processed in the corona discharging field contained 2.35, 1.94, and 1.25 fold more phenols, carbonyls, and acids, respectively, than the condensates of the whole smoke used as a control. From this he concluded that the corona discharging field caused chemical changes in the smoke, but that the changes were desirable due to the increase in phenols, acids, and carbonyls, and therefore the processing of smoke in a high voltage electrostatic field increased its usefulness for curing purposes. Foster and Simpson (l961a,b) studied the differences in the vapor and particulate phases of smoke with the use of an electrostatic precipitator and concluded that the smoking process for foods is essentially a vapor absorption phenomenon and that contributions from the particulate phase are almost negligible. They reported that there were no significant differences in the appearance, flavor, and keeping qualities of normally- smoked and vapor-smoked fish. Porter (1963) conducted organoleptic studies on food products smoked with either whole smoke or the vapor phase resulting from electrostatic precipitation. His results indicated -l8- that the vapor-smoked cheese had a higher smoke intensity than the whole- smoked, however, the test panel showed no preference for either type over the other. In the case of bacon treated similarly, there was no signi— ficant difference in either smoke intensity or acceptability as evaluated by the panel. Contributions of Smoke Phenols to the Desirable Properties of Smoked Foods Among the many desirable effects which Jensen (1954) lists as being gained from the prOper smoking of meats are: (l) the imparting of desir- able organoleptic properties, (2) the application of antioxidants to the fat, (3) the impregnation of the outside portions of the meat with smoke constituents that serve as antiseptics and germicides, and (4) the pro- duction of a desirable "finish" or gloss on the skin and/or flesh sides of meat. Organoleptic Properties Tilgner at al. (1962b) compared the sensory qualities of phenolic and acid fractions of beechwood smoke by test panel evaluations of taste and odor and concluded that phenols were mainly responsible for the typi- cal "cured" smoky flavor of the test solutions while acids seemed to play a smaller role. Khrko and Kelman (1962) noted in single experiments where summary fractions of smoke components had been added to sausage meat that only the phenolic fraction produced odors which approached "smoked" aroma. They further concluded that (1) of all the smoke compounds only a few were responsible for the smoky aroma, namely those of the methylguaiacol type with medium.boiling points, and (2) that low and high boiling phenols -19- contributed little since they have either no odor or their odor is much different from.that of smoke. Fiddler at al. (1966) stated that preliminary investigations indi- cated the essential smoke odor was in the phenolic portion of their whole smoke condensate from hickory sawdust. Wasserman (1966) combined the three major phenolic components -- guaiacol, 4—methylguaiacol, and 2,6 dimethoxyphenol —- in the same proportion as they occurred in the whole smoke condensate analyzed by Fiddler 23 al. (1966), and observed the resulting mixture to have an odor only slightly reminiscent of the whole smoke condensate. Antiseptic and Germicidal Properties At the present time the official method for testing and comparing germicides is one devised by the Food and Drug Administration of the United States Department of Agriculture in which the disinfecting effi- ciency of a chemical is compared with that of phenol under the same standard conditions (Krueger, 1953). From the results of this test the "phenol coefficient" of the chemical is obtained, and while it gives little information as to the value of a germicide under conditions of actual use, it does serve as a useful yardstick in evaluating germicidal chemicals. Shewan (1949) listed phenol coefficients for various fractions of wood smoke components as determined for E. typhi bacteria. Some of the coefficients from his list are: formaldehyde, 1.05; phenol and cresol, 2.8; cresols and guaiacol, 4.5; guaiacol and creosol, 7.9; creosol and pyrogallol ethers, 9.0; and pyrogallol ether homologues, 6.6. -20... It was reported (Anonymous, 1956) that different microorganisms are inhibited to markedly different extents by different constituents of smoke, and examples cited were that most of the growth inhibition of a strain of Achromdbacter was due to 2,6-dimethoxyphenol and its derivatives, while the major activity of wood smoke against Bacillus cereus was due to hydroquinone. Preliminary studies (Anonymous, 1954) indicated that the greater portion of the bactericidal activity of wood smoke was due to three distinct groups of phenols and the most potent group consisted of 2,6-dimethoxyphenol and its papayalkyl derivatives. Other studies relating to the bactericidal or germicidal pr0perties of wood smoke as imparted to foodstuffs have been reported by Hess (1928); White 33 al. (1942); Jensen (1943); Gibbons a; a1. (1954); Hedrick REHEA- (1960); Kochanowski (1962); and WOlkowskaja and Lapszin (1962). While these studies did not specifically indicate that phenols were responsible, all were in agreement that smoke has a definite preserving effect on foods through the control of microbial growth and hence, spoilage. Antioxidant Properties Kurko (1959) reported on a preliminary study in which fractions of smoke containing phenols, acids, and neutral compounds (alcohols, alde- hydes, ketones), respectively, were added to melted pork fat. The mixtures were heated at 110°C for periods of O to 5 hours after which the degree of rancidity was evaluated by determining their peroxide values. He found that only the phenolic compounds were strong antioxidants while the acids were slightly effective and the neutral compounds had a negative -21- effect. He further stated that the most active antioxidants of the phenol fraction were contained in the portion which had the highest boiling points and were characterized by average methoxy contents of 27.9% and average molecular weights of 162. He investigated this fraction by paper chromatography and found it contained methoxy derivatives of pyrogallol. Kurko (1966) reported a study involving the comparison of the anti- oxidant properties of BHT and BHA with many individual phenolic substances most of which had been isolated from.smoke or smoked meats by the author or other research workers. He stated that: (l) phenol, cresols, xylenols, thymol, guaiacol, and guaiacol derivatives were poor antioxidants; (2) phloroglucinol, pyrogallol methyl ethers, and pyrogallol homologues were only slightly better; (3) pyrocatechol and hydroquinone were strong anti- oxidants being equal to or more potent than BHA; and (4) the strongest antioxidants were 3—methylpyrocatechol, 4—methylpyrocatechol, and pyro- gallol, all of which were considerably more effective than BHT. He then compared these results with those of paper chromatography studies and concluded that the substances which determined the inhibiting properties of the high boiling phenolic fraction of smoke belonged to compounds of the pyrogallol, pyrogallol alkyl derivative, and pyrocatechol types in addition to pyrogallol methyl ethers and pyrogallol homologues. It was reported (Anonymous, 1958) that for smoke deposited in the traditional manner almost all of the antioxidant activity was due to the smoke fractions containing guaiacol, 4—methylguaiacol (creosol), 4- ethylguaiacol, and 2,6-dimethoxyphenol. Estimations of these individual -22.. compounds and antioxidant testing of model solutions showed they were responsible for the antioxidant activity and that 4-methylguaiacol made the largest single contribution. ‘Watts and Faulkner (1954) investigated the antioxidant effect of four commercial liquid smokes in model aqueous lard systems. One of the liquid smokes differed greatly in odor compared to the rest, had no anti— oxidant effect, and gave a negative test for phenols. The other three revealed an antioxidant activity when tested individually and also exhibited a synergistic action with ascorbic acid. They further stated that the smoke solution which had the strongest antioxidant effect on the artifi- cial system exhibited a very pronounced antioxidant effect when tested on frozen meat. They concluded that the antioxidant pr0perties of the smoke solutions were related to their phenolic content. Additional studies involving the antioxidant effects of wood smoke on food products were reported by Lea (1933), White (1944), and Banks (1950). These studies did not implicate phenols specifically as the antioxidant agents, however, all supported the general findings of the studies reported above that wood smoke does prevent the deterioration of foodstuffs to a marked extent by retarding the process of fat oxidation. Desirable "Finish" or Gloss and Color Jensen (1954) stated that the desirable "finish" or gloss of smoked meats was due to resins (resulting from the condensation of aldehydes and phenols in the smoke) in combination with a thin film of grease on the surface of the meat. Draudt (1963) reported that in.work associated -23.. with the electrostatic deposition of smoke on bacon he had observed that the bacon was a light golden color after heavy smoking, but became very dark brown after heating at 125-135°F for several hours. He stated that this practical observation was in line with the idea that smoked color development depended on the components of wood smoke resins. Ziemba (1962) observed that the formation of color and gloss in smoked fish occurred in two stages, the first of which was characterized by the Maillard reaction involving proteins, and the second which resulted from the condensation and polymerization of aldehydes and phenols present in the smoke. Qualitative Studies of Phenols in Wood Smoke Pettet and Lane (1940) reacted the lowest boiling portion of their phenolic extract with dicyclohexylamine to form a "double compound” from which phenol was released by action of caustic soda. They identified phenol on the basis of comparisons between the individual and mixed melt- ing points of the unknown and standard compounds. They further noted that more complex phenols were present but unidentified. Commins and Lindsey (1956) studied the phenolic fraction of wood smoke by converting the phenols to their methyl ethers. They refluxed the phenolic fraction with dimethyl sulfate for 3 hours to form the ethers which were then separated by column chromatography and identified Spectr0photometrically by their ultra-violet absorption characteristics. By this procedure they identified phenol, gfcresol,.m7cresol,upfcresol, resorcinol, catechol, hydroquinone, l-naphthol, and 2-naphthol. They also noted that other -24- methyl ether containing phenols were present in the original extraction mixture, especially guaiacol, but methods had not been worked out to determine them separately from the phenolic methyl ethers that were syn— thesized. Jahnsen (1961) studied the chemical composition of smoke produced from hickory sawdust containing 50% moisture in an all pyrex generator. The residue portion (remaining after "nitrogen stripping" at low temper- ature) of the total condensate was resolved by means of gas-liquid chromatography utilizing a 9 ft. polyethylene glycol succinate column with temperature programming. With this method he identified the follow- ing phenols: guaiacol, 2,6-dimethylphenol, the three cresols, 3,4- and 3,5-dimethy1 phenol, thymol and 2,6-dimethoxyphenol. He also subjected the residue to lead acetate precipitation and phenols which were identi- fied following precipitation and regeneration were guaiacol, 4-methyl- guaiacol, 2,6-dimethoxyphenol, catechol, pyrogallol, and phloroglucinol. Phenols that did not precipitate with lead acetate but were identified by paper chromatography were the three cresols, 2,6-, 3,4-, and 3,5-dimethy1- phenol, hydroquinone, thymol, and vanillan. He also identified the phenolic ethers anisole and veratrole in the residue of the condensate. Simpson and Campbell (1962) noted that phenolic compounds had been shown to be re3ponsib1e for the antioxygenic preperties of smoke as well as most of the bactericidal activity against fish Spoiling micro-organisms, and briefly discussed methods used to determine the individual compounds which produced these preserving effects. They stated that during the -25- past ten years they had worked out methods for separating, identifying and estimating the phenolic constituents of wood smoke which involved preliminary countercurrent analysis of the crude mixture followed by paper chromatography of the individual fractions. They separated the volatile phenols by chromatographing the azo-dye derivatives. By those methods they found the principal phenols of smoke to be alkyl derivatives of guaiacol, 2,6-dimethoxyphenol, catechol, phenol, and hydroquine, which contained 3 carbon atoms in the side chain that was usually located 222i to the hydroxyl but occasionally was in the ggghg position. They also identified vanillan, but grouped it with the aldehydes rather than the phenols. Hollenbeck and Marinelli (1963) identified 2,6 dimethoxyphenol in an aqueous solution of the water-soluble vapor components of wood smoke (sold under the trademark name of CharSol as liquid smoke). Their method of identification consisted of paper chromatography in which the chromatogram was sprayed with diazetized pfnitroaniline following deve10p- ment to form the blue azo dye of the phenolic compounds. Kurko (1966) studied the phenolic fractions responsible for antioxidant properties of smoke by paper chromatography and reported that the fraction which inhi- bited oxidation the most contained pyrogallol methyl ethers and pyrogallol homologues (methyl-, ethyl-, and propyl-pyrogallol) in addition to small amounts of phenols with unknown chemical composition. Fiddler E£.El- (1966) conducted a study of the phenolic composition of hickory sawdust smoke by utilizing the techniques of gas-liquid chroma— tography (6 ft., carbowax ZOM-terephthalic acid column with temperature programming) and infra-red spectroscopy. They identified the following -26.. phenols by both methods: phenol, guaiacol, 4-methy1guaiacol, 4-ethyl- guaiacol, 4-pr0py1 guaiacol, 4-viny1guaiacol, 2,6-dimethoxyphenol, 2,6- dimethoxy-4-methylphenol, 2,6-dimethoxy-4-ethylphenol, 2,6-dimethoxy-4- pr0py1phenol, and vanillan. The phenols, mrcresol and 4-ally1guaiacol, were identified by gas-liquid chromatography alone, as was the phenolic ether, veratrole. Studies Indicating Relative Quantities of Individual Phenols in Wood Smoke Jahnsen (1961) concluded from gas-liquid and paper chromatography studies that the phenolic compounds present in hickory sawdust smoke in the greatest quantities were guaiacol and 2,6-dimethoxyphenol. Hollenbeck and Marinelli (1963) reported that their paper chromato- graphic studies of the aqueous solution of the vapor components of wood smoke showed the phenolic fraction to contain a preponderance of 2,6- dimethoxyphenol. Wesserman (1966) reported that the phenolic fraction of the whole smoke condensate which had been analyzed by Fiddler EEHEA: (1966) con- tained as major components 4—methyguaiacol, guaiacol and 2,6-dimethoxy- phenol in a ratio of l : 3 : 4.6 based on the peak areas from gas chroma- tograms. EXPERIMENTAL PROCEDURE This investigation was divided into two main areas for study. The first part was undertaken to determine if there were major qualitative differences in the phenolic fractions of whole smoke and vapor phase smoke both of which were produced in a laboratory model smoke generator. The second phase involved comparisons of the phenolic fractions of smoke produced in a conventional smokehouse Operated under various conditions of temperature and relative humidity. In addition, the phenolic contents of three commercial "liquid smoke" solutions were compared qualitatively. Sawdust Source and Condition Whole Smoke vs. Vapor Phase Study The sawdust utilized in this study was predominantly from hard maple wood (agg; saccharum). The particulate size of the sawdust was such that it could be sifted through a 1/4 inch but not a 1/8 inch mesh screen. The moisture content of the sawdust was determined prior to each run by heating duplicate samples in a drying oven for 10 hrs at 110°C, reweigh- ing the samples while hot, and calculating the weight loss as percent moisture in the original sample. The moisture content was in the range of 8-10% for all runs. Relative Humidipy and Temperature Studies The sawdust utilized in these studies was that portion of the total sawdust which was not used in the whole smoke vs. vapor phase study due -27- ... : 3H“.- -28.. to its extreme particle sizes. water was added to the sawdust until it appeared to be saturated. The moisture content was determined in the manner described earlier and ranged from 48-54%. Smoke Production Whole Smoke vs. Vapor Phase Study The smoke for this study was produced by the laboratory model smoke generator described by Porter (1963) and Porter et a1. (1965). All runs were made in duplicate and were of either 2 or 5 hours duration depending on the method of sample collection. The sawdust flow rate onto the hot- plate was adjusted to approximately 1.5 grams per minute. The hotplate temperature was measured every 15 minutes with a Leeds and Northrup, multi-range, null-current potentiometer. The potentiometer was connected by leads to a chromel-alumel thermocouple located where the sawdust dropped onto the surface of the hotplate. The average hotplate tempera- ture for every run was in the range of 510-520°C. The temperature of the smoke was measured in the generator chamber at a point just preceding the electrostatic precipitator. It was also measured in the 12 inch x 1/4 inch (ID) piece of Tygon tubing (Mathieson Scientific, Inc.), connecting the generator to the first collection vessel, at a point immediately preceding the vessel. These temperature measure- ments were obtained with a Minneapolis Honeywell, 12-point, recording potentiometer connected by leads to copper-constantan thermocouples located at the two positions. The average smoke temperature in the chamber -29- for every run was in the range of 115-120°C. The average smoke tempera- ture in the tubing for each run where the collection was made in an ice- water bath (2-3°C) was in the range of 52-57°C, and for every run where collection was made in a heated water bath (38°C) the average smoke temperature in the tubing was in the range of 90-98°C. The electrostatic precipitator was turned on only for those runs in which the vapor phase was to be collected, and it was removed and thoroughly cleaned prior to each run in which it was to be used. Relative Humidity and Temperature Studies The smoke for these studies was produced in a Mepaco Tipper Model, smoldering-type, smoke generation unit which was attached to a commercial, air—conditioned smokehouse. The smokehouse was equipped with a tempera- ture and humidity control system.(Minneapolis Honeywell) which utilized wet- and dry-bulb temperature measuring devices. All runs were made in duplicate and were of 2-hours duration; and while the sawdust feed to the generator was not readily adjustable, or closely controllable, it was possible to maintain a heavy, dense smoke in the smokehouse by close attention to the generator throughout each run. It was noted that approxi- mately 13-14 kg of the dampened sawdust were required for each run. The air supplied to the generator by a blower was measured with an air flow meter (Air-Meter, Hastings Instrument Co., Inc., Hampton, Virginia) and was maintained at a setting for every run which resulted in a flow rate of approximately 0.5 cubic meter per minute. The air flow into the smokehouse through the supply duct was also measured with -30.. the above meter and it was determined that the air (smoke) in the smoke- house (14.8 cubic meters) changed approximately 5 times per minute. For the relative humidity studies, the smokehouse temperature was maintained at 83°C and runs were made with relative humidities in the smokehouse of 7, 20, 40, and 60%. The relative humidity was maintained at 30% for the temperature studies and runs were made with smokehouse temperatures of 27, 38, 66, and 93°C. The various relative humidity levels were selected because they approximate the range of relative humidities encountered in commercial practice (Simon 95_§;,, 1966), while the various levels of temperature were used because they cover the range of attainable temperatures for the smokehouse utilized in these studies. Collection of Smoke Samples The water used as a collection medium in these studies had been pre- viously distilled and deionized. Whole Smoke vs. Vapor Phase Study Two different systems were used to collect the smoke samples for this study. The first system was adapted from the collection procedure used by Barber ggdal. (1964) and involved the drawing of the smoke aerosol through two 500 ml gas-type impingers which were connected in series to the laboratory smoke generator. The impingers contained 400 ml of either 1N NaOH or water. When 1N NaOH was used as the collection medium, the impingers were submerged in an insulated ice-water bath where the tempera- ture of the collection solution was maintained at 2-3°C throughout the -31- run. When water was used in the impingers, the temperature was maintained at either 2-3°C with the ice-water bath, or at 38°C with a heated water bath. The second impinger was connected to a trap packed with glass wool (to remove tars), and this trap in turn was connected to a 1/4 inch (ID) vacuum line. The vacuum was adjusted to produce an air (smoke) flow rate through the system of 4.5-5.5 1pm. This was the maximum airflow which could be used without drawing some of the collection solution out of the impingers. Following each run, an aliquot of the sample was removed for pH and "total phenol" determinations, and the remainder was transferred to a dark glass bottle and stored at 2°C until analyzed. The second collection system was adapted from the procedure reported by Simon EEHEA~ (1966) and involved the use of water-filled, artificial frankfurter casing (Nojax casing, size 24, Union Carbide Corp., Food Products Division). The casing was cut into 1.5 meter lengths and washed in 30°C water for 1/2 hour to remove the glycerine. After washing, the casing was tied at one end, filled with 500 ml of water, and then tied at the other end. This resulted in a filled casing with dimensions of 1.45 meters in length and 2.1 cm in diameter. The water-filled casing was then placed on two hooks, each attached one-fourth of the distance from an end, and susPended in the generator chamber at a point immediately following the electrostatic precipitator. The temperature at this location was in the range of 102-105°C. After each run the samples were cooled to room temperature and the percent of water lost by evaporation was determined by comparing the final volume to the original. An aliquot was -32- removed for pH and ”total phenol" determinations, and the remainder was transferred to dark bottles and stored at 2°C until analyzed. A control run was made to determine if the water-filled casing, itself, contributed anything to the phenolic content of the smoke samples. The control sample was obtained by filling a segment of Nojax casing with 250 m1 of water and hanging it in a drying oven at 98°C for 2 hours; it was analyzed in the same manner as the other artificial casing samples. Relative Humidipy and Temperature Studies The smoke samples for these studies were collected only by the use of water-filled artificial casings. In the temperature study the Nojax casing described above was cut into 3~meter lengths and after washing was filled with 1 liter of water. The resulting full casing had dimen- sions of 2.90 meters in length and 2.1 cm in diameter. It was attached to three hooks (two were located 1/6 distance from the ends and the third was attached in the middle) and placed in the smokehouse at a location near the back at tOp-center. After collection the sample was treated in the same manner as the Nojax sample for the whole smoke vs. vapor phase study. Both the Nojax casing and a bologna casing (ClearoZip, fibrous, cellulose casing #5 manufactured by Tee Pak, Inc., Chicago, Illinois) were used for collecting smoke samples in the relatiVe humidity study. The Nojax casing was utilized exactly as in the temperature study. The bologna casing was filled with 2 liters of water resulting in a full casing with dimensions of 36 cm in length and 8.6 cm in diameter. The -33- casing was attached to a hook and placed in the smokehouse at the same location as the Nojax casing. Following collection, the sample was treated in the same manner as described previously for Nojax samples, and it provided a means of comparing the effect of casing type on the quali- tative nature of the smoke phenols absorbed by the water. In addition to the samples whose collectionSwere described above, another sample was obtained from both the temperature and relative humidity runs where Nojax casing was used. This sample resulted from the washing of the outside of the casing (the ends were closed again after the con- tents were emptied) in 200 ml of water at 23-25°C for 2 hours. The purpose of this sample was to determine if the casing was selectively permeable to smoke phenols, and it was analyzed in the same manner as the other smoke samples. Estimation of Total Phenols The colorimetric method of Tucker (1942) was modified to provide an estimate of the total quantity of phenols in each collected solution. The basis for his procedure was the indophenol test which was described in detail by Gibbs (1927) and involves the condensation of phenols with quinonechlorimide compounds to produce the blue-colored indOphenol dyes. Modifications made in Tucker's method were based on preliminary studies utilizing solutions of reagent grade phenol, guaiacol, and 2,6 dimethoxy- phenol individually and as a mixture. These preliminary studies were undertaken to (1) determine the time- temperature requirements for completion of the color reaction, (2) determine -34_ the wavelength at which the Optical density was maximum for the indophenols of phenol, guaiacol, 2,6-dimethoxyphenol, and a representative sample of the collected smoke solutions; and (3) to determine the range of concen- tration (mg/100 ml) of the known compounds for which the relationship between Optical density and concentration was linear, or obeyed the Beer- Lambert Law. The estimation of total phenols in each sample solution was made within 2-6 hours following its collection. The pH of the solution was determined by use Of a Corning MOdel 12 pH meter, and then adjusted to 7.0 with either 1N HCl (NaOH collection solution ) or 1N NaOH (water collection solution). The volume of acid or base required for the pH adjustment was recorded for use later in computing the estimate of total phenols. The indOphenol reagent used in this procedure was prepared as follows: a 0.1% (w/v) solution of 2,6-dichloroquinonechlorimide in absolute alcohol served as the stock solution which was stable for 2-3 weeks when refrigerated in a dark, stOppered flask; the ind0phenol reagent was prepared immediately before its use by diluting the stock solution 1:15 with deionized distilled water; a precipitate formed in the reagent if it was not used in 15-20 minutes. The colorimetric procedure was carried out on the sample solution in triplicate as follows: a 5 m1 aliquot of the sample was pipetted into a 15 x 180 mm test tube, and this was followed in order by the addition of 5 ml Of 0.5% sodium borate buffer (Na2B4O7.1O H20) and 1 ml of the indophenol reagent; the tube was then stOppered and the contents thoroughly -35- mixed by shaking; next, the tube was placed in a controlled-temperature cabinet at 38°C for 2 hours to permit completion of the color reaction; following this, the indophenol dye was extracted from the aqueous solu- tion with 15 ml of n-butanol in a small separatory funnel; the butanol- dye layer was transferred into a 25 m1 graduated test tube and the volume increased to 20 ml with n-butanol; then 2 ml of n-butanol saturated with ammonia was added, and after the solution was mixed by gentle shaking, its Optical density was read against a reagent blank at 635 mu on a VBausch and Lomb Spectronic 20 spectroPhotometer. Similarly, deionized distilled water solutions of standard phenol in concentrations of 0.0 (reagent blank), 0.25, 0.50, 0.75, and 1.0 mg per 100 ml were used to derive a standard curve. The estimate of total phenols (mg/100 ml) in the collected solution was then Obtained by comparing the Optical density Of the sample with the standard curve and taking into account the dilution made by the pH adjustment initially. The total phenol estimates were utilized in the paper and thin-layer chromatography studies where the extracted phenolic fractions of the different samples were compared qualitatively. The qualitative compari- sons were based on a visual examination of the chromatogram patterns, and it was necessary to have some basis for representing the samples on the chromatogram with approximately equal quantities of total phenols. Preliminary investigations utilizing a Bausch and Lomb Spectronic 505 SpectrOPhOtometer revealed the indOphenols of phenol, guaiacol, 2,6- dimethoxyphenol, and a representative smoke sample to each have an absorption —36- Spectrum with a maximum optical density in the wavelength range of 625- 645 mu. Thus, any one of the three known compounds could have served as the reference for the total phenol estimates in these studies. Standard curves for guaiacol and 2,6-dimethoxyphenol were derived only for the sake of comparison, though estimates based on either of them might have been more accurate than those on phenol since the literature indicated that both guaiacol and 2,6—dimethoxyphenol to be more abundant than phenol in wood smoke. Extraction of Phenolic Fraction The method used for extracting the phenolic fraction from the collected samples was developed from portions of the procedures reported by Grouse a; a1. (1963) and Barber gpflal. (1964) as well as information concerning the properties of phenols which was presented by Morrison and Boyd (1959) and Roberts and Coserio (1964). A flow chart of the extrac- tion procedure for water-collected samples is presented in Figure II. The NaOH-collected samples were extracted by the same procedure with the first three steps omitted. The extraction was made from a 100 ml aliquot of the collected sample. The procedure was carried out in a cold room (2-3°C) except for the steps involving the concentration of ether solution by aspirator vacuum. The water-collected samples had an initial pH of 2.9-3.5 and thus the phenols were in their "free" state and ready for extraction. In the NaOH-collected samples, the initial pH was 12.2-12.3 and the phenols had to be released from their salts prior to extraction with the organic solvent. In this case the pH was adjusted to 4.0 to be -37- Smoke Sample Collected in water (l) extract with 2 volumes of ether in separatory funnel Aqueous layer Ether layer (discard) (2) concentrate under asPirator vacuum to 1/2 volume (3) extract with equal volume of 1N NaOH Aqueous layer Ether layer (discard) (4) acidify to pH 4.0/1N HCl (5) extract with 2 volumes ether ALueous layer Ether layer (discard) (6) repeat step (2) (7) extract with equal volume of saturated NaHCO3.solution in separatory funnel gqueous layer Ether layer (8) wash with equal volume of ether Aqueous layer Combined ether layers (discard) (9) dry over MgSO4 (anhydr.) for 2 hrs. (10) concentrate to desired final volume under aspirator vacuum and with nitrogen. Sample for Chromatpgraphic Analysis Figure 11. Scheme for extracting phenols from the collected smoke sample. -38- comparable to those of the water-collected samples. Diethyl ether (hence- forth referred to simply as ether) was selected as the organic extraction solvent because: (a) the majority of phenols reportedly present in smoke are more soluble in ether than in other organic solvents (Handbook of Chemistry and Physics); (b) its low boiling point provided for rapid condensation of the sample during the extraction procedure with probably little loss of phenols by evaporation; (c) it provided a good solvent for the gas-liquid chromatographic analysis where it eluted from the column prior to all phenols, and was not easily contaminated with water which adversely affects the ionization detector; and (d) it provided a good sol- vent for spotting phenols on paper and thin-layer chromatograms because it dried rapidly preventing undue Spreading of the Spots. The concentra- tion steps (2), (6), (10) utilizing a Rinco Rotating High Vacuum-Type Evaporator connected to an aspirator were carried out with the sample in a boiling flask submerged in a water bath at 18°C and the rate of ether evaporation being approximately 10 ml per minute. The NaOH in step 3 was used to extract the acidic components (phenols and carboxylic acids) leaving the neutral compounds in the ether which was discarded. NaHCO3 in step 7 extracted the strongly acidic carboxylic acids leaving the weakly acidic phenols in the ether which was saved and combined with the ether used to wash the NaHCO3 extract in step 8. The combined ether solutions were dried over anhydrous MgSO4 and concentrated to a volume of approximately 15 ml in a boiling flask. The condensed ether solution was transferred to a 25 m1 graduated test tube and the boiling flask -39- rinsed with 10 ml of anhydrous ether which was then added to the test tube. The contents of test tube were concentrated to 2 ml and transferred to a 4 ml graduated vial. The test tube was rinsed with 2‘ml of anhydrous ether which was added to the vial and a final volume of l to 2 ml was achieved by evaporation with a gentle stream of nitrogen. The vial was then stOppered and the sample stored in the dark under refrigeration until analyzed. The sample represented the total phenolic fraction with the phenols in the free state, and it was considered to be the most appro- priate sample for the chromatographic analyses. Any attempts to divide the extract into steam—volatile and non-steam volatile fractions, or to convert the free phenols to derivatives prior to analysis might have re- sulted in the complete loss of some minor components. The extraction procedure was tested for efficiency with a mixture of known phenols both by the estimation of total phenols and by gas-liquid chromatography of the mixture before and after extraction. These tests revealed that 75-80% Of the phenols were recovered from the original solution and that the procedure was not selective for any of the phenols in the mixture. The extraction procedure was also tested with an unknown sample by the estimation of total phenols and the Optical density of the final solution was about 80% of the original. Chromatographic Analyses The ability to separate and identify individual phenol compounds was required before qualitative comparison studies of the phenolic fractions of various smoke samples could be made. The literature was very limited _40_ in the area of studies specifically related to the qualitative analysis of wood smoke phenols, however, there were some reports dealing with chromatographic procedures for analyzing the phenolic contents of other unknown mixtures. The phenols previously identified in wood smoke com- prise a group with a wide range Of physical and chemical pr0perties, and yet some of the compounds are very similar in their characteristics. Either of these facts considered individually would make separation a difficult problem, and taken together they make the separation and subse- quent identification of wood smoke phenols by any single method virtually impossible. Gas-liquid chromatography (GLC) was considered to be the best available technique for analyzing smoke phenols, and therefore, major emphasis was placed on the use of this technique. In addition, thin-layer chromatography (TLC) and paper chromatography were utilized to complement the GLC technique. All standard phenol compounds utilized were of the highest purity grade commercially available. Attempts to further purify them were not made because impurities were not detected by any of the three chromatographic techniques employed. The phenols were all supplied in colored glass bottles and they were stored in the dark at room temperature with the covers tightly clOsed. A complete list of the standard phenol compounds used and the sources from which they were obtained is presented in Appendix II. Gas-Liquid Chromatographic Procedure A Barber-Colman Mbdel 20 gas chromatograph equipped with a radium ionization detector and a Barber-Colman recorder was used. Crouse gp_al. -41- (1963) reported that various workers had found that no single column pack- ing would resolve all C6 - C9 phenols. ‘With this in mind, several different types of column packing were systematically tested under various conditions of column temperature and argon carrier gas flow rate in order to determine which combination produced the best resolution with the least "peak-tailing” of mixtures of known phenols. As a result of this preliminary testing, three different columns designated A, B, and C were chosen for the analyses. A 6 ft section of 1/4 inch aluminum tubing was used for each column. The columns were packed with the aid of an electric vibrating needle and the ends of the column were stOpped with glass wool. The column packings were obtained or prepared as follows: column A, a 10% Carbowax 20M on Diaport W.A.W. (60/80 mesh) packing Obtained from the F & M Scientific Corp. was used as received; column B, a 10% (w/w) mannitol (Difco Laboratories, Detroit, Mich.) on 60/80 mesh aciddwashed Chromasorb W (Johns-Manville Products) packing was prepared by mixing 5 gms of manni- tol, 45 gms of Chromosorb W, and 150 ml of methanol in a boiling flask, and then evaporating the methanol very slowly (2 hrs) by using the Rinco rotating evaporator previously described; and column 0, a 5% (w/w) manni- tol on Chromosorb W packing was prepared in a similar manner as the packing for column B. After packing, the columns were pr0perly placed in the chromatograph and conditioned for 20 hrs at a temperature 5°C above and with a carrier gas flow rate equal to those conditions employed in the analytical runs with each column. In all runs the column pressure was set at 20 psi, the cell voltage at 1250, and the sensitivity (gain) at 10. The columns were utilized with different conditions of column temperature, -42- flash heater (preheater) temperature, detector cell temperature, and carrier gas flow rate. Column A under the conditions used gave a fairly good separation of the majority of known phenols used but did not resolve some of the low boiling phenols and the highest boiling phenols (di- and tri-hydroxy) were not eluted. Column B was used under two different sets of conditions; with a column temperature of 160°C the resolution was similar to that of column A, but when the column temperature was 100°C the low boiling phenols were separated much better. Column 0 was used in an effort to separate the high boiling phenols which were not eluted from the other two columns, but all of them.did not elute from it. An attempt to form and chromatograph their silyl ethers according to the methods of Freedman (1964) proved unsuccessful. All samples were dissolved in ether and were injected into the flash heater with a 10 ul Hamilton Microsyringe. The various known phenols were divided into groups for chromatographing so that all would be eluted from.the column in use, and their retention times were compared with the retention times of peaks in the unknown samples to arrive at tentative identifications. In addition, small quantities of the knowns which were suspected of being in the unknown samples based on similar retention times were added to the unknown mixture to note whether the corresponding unknown peaks increased in size. The chromatograms of the various unknown samples were utilized in making qualitative comparisons once the identity of their components had been verified by one or both of the other chromatographic techniques used. -43- These comparisons were based on the relative areas of peaks corresponding to selected phenol compounds present in major quantities. The peak areas were determined by the general method of multiplying the peak height by the peak width at half-height, both measurements being in the same units. Thin-layer Chromatographic Procedures Realizing that GLC retention times alone were not sufficient criteria for the positive identification of phenols, attempts were made to confirm the GLC findings by TLC techniques. Crump (1964) reported a method in which the p-nitr0pheny1azo dyes of simple alkyl phenols were satisfactorily separated by TLC. However,in.this study it was desired to chromatograph the free phenols so they could be extracted from the thin-layer chromato- gram following separation and rechromatographed by GLC. In order to do this, a modification of Crump's method was utilized. The diazotiud p—nitroaniline spray solution used in this procedure was prepared as follows: (1) a .015% (w/v) solution of the nitro amine in 1N HCl was prepared as a stock solution and could be stored for several months under refrigeration, (2) a 10% (w/v) aqueous solution of sodium nitrite solution was prepared and could be stored for a month in a dark bottle under refrigeration, and (3) 50 ml of the amine hydrochloride stock solution was diazotifled at 0°C with 2 ml of the nitrite solution just prior to use for spraying chromatograms. The diazotiZation reaction was characterized by a complete loss of the yellow color of the stock solution and the resulting spray solution had to be used within 2 or 3 hours after preparation. The thin layer plates utilized were prepared as follows: -44— (l) a basic adsorbent was prepared by slurrying 25 g of Merck Silica Gel G with 50 ml of .5N NaOH in a stopped flask; and (2) five 20 x 20 cm glass plates were immediately layered with a 250 u thickness Of the slurry by use of a Desaga adjustable applicator and mounting board. The thin-layer plates were then allowed to air-dry in place after which they were placed in a drying rack for subsequent activation and storage. They were then activated for 1 hour at 100°C in a drying oven and then either used imme- diately or stored in a dessicator containing a layer of Drierite. A devel- oping solvent composed Of chloroformracetone (5:1) was determined to give the best separation of several solvents tested. Development was carried out in a 8 1/2 x 4 l/2 x 8 1/2 inch Desaga developing tank with a ground, fitted lid. The tank contained 100 m1 of the developing solvent and was sealed with a small amount Of joint sealer. The known and unknown samples dissolved in ether were spotted on the activated plate with a 10 ul Hamilton syringe guided by a Camag spotting guide. The plate was then placed in the develOping tank and allowed to develop until the solvent had migrated 15 cm above the point Of sample application. The development occurred at room temperature (23-25°C) under subdued light and required approximately 1 l/4 hours. Following development, the plates were removed from the tank and allowed to dry. They were then sprayed with the solution of diazotized.penitroaniline solution, allowed to dry, and then sprayed with a 10% (w/v) aqueous solution of Na 00 23‘ appeared as characteristically colored bands. The characteristic colors After drying, the individual p—nitrOphenylazo dyes . n . . u . : r . u- i v 1 . u I . -45- were of considerable aid in the identification procedure since the separa- tion of the bands was incomplete. Identification of the bands in the unknown samples was based on a comparison of their location and color with those of the bands in a known mixture deve10ped on the same plate. A direct link between the GLC and TLC findings was made as follows: a preparative-type run of a representative unknown was made by TLC; after development only the edge was Sprayed to locate the position of the various bands; sections of the layer containing selected bands were scraped from the plate and extracted with anhydrous ether; the ether solutions were concentrated and chromatographed by GLC to confirm the identity of the susPected phenols. For the qualitative comparisons, each of the samples was represented by approximately the same quantity of total phenols based on the estimates of the total phenolic content of the original collected solutions. Paper Chromatographic Procedure Because the resolution of the high boiling phenols had not been satisfactorily obtained by either GLC or TLC, attempts were made to achieve separation by paper chromatographic techniques. Impregnated papers had been reported by various workers to give the best separation of phenols, but all attempts to use papers impregnated in the laboratory proved to be only mildly successful. There was the problem of obtaining a uniform concentration of the impregnating material throughout the sheets and, in addition, most of the methods required a conditioning of the paper prior to use. This latter requirement resulted in a "blooming" ~46- of several of the phenol spots during develOpment. Clark (1964) reported a method Of separating phenols on a commercially prepared ion-exchange paper. After some modification, a portion of his procedure proved to be satisfactory for the separation of the high-boiling polyhydric phenols, therefore it was used. Reeve Angel Grade SB-2 Amberlite ion-exchange resin-loaded paper was utilized. The resin, Amberlite IRA-400, is a strong base of the Cl“ form. The solvent system employed was Clark's "solvent B", consisting of butanolawater-acetic acid (6:2:1, v/v). The chromatograms were deve10ped in a cylindrical chromatography tank (10 1/2 inch diameter x 16 inch height) equipped with an air-tight lid and containing 300 ml of the developing solvent. All chromatographic runs were made at room temperature with ascending solvent flow in the machine direction of 11 x 18 inch strips of the resin-loaded paper. Known and unknown mixtures were Spotted on the paper one inch apart on a line 1 1/2 inches from the bottom edge. After the sample spots were dried, the paper was rolled into a cylindrical form with the sample applications to the inside. The two vertical edges (11 inch dimension) were stapled so they were held about 2 mm apart. The paper was then placed in the tank and allowed to deve10p until the solvent front had migrated 15 cm above the sample application line. This required approximately 2 1/2 hours, and trial runs of 4 1/2 hours resulted in a solvent front only 4 cm.higher with no appreciable improvement in the resolution. After development, the chromatograms were dried in a circu- lating air oven at 70°C for 3-5 minutes. The phenols were detected by a light spray of the diazotized penitroaniline solution (described in -47- the TLC procedure) immediately followed by the 10% Na2003 solution. The chromatograms were then placed on sheets of thick filter paper and allowed to air dry. The pfnitrOPhenyl azo dyes of the individual phenols were characteristically colored and this aided in their identification. Some of the spots were not completely separated, however, their positions and colors were easily compared in the known and unknown samples located adjacently on the chromatogram. For the qualitative comparison studies, the unknown samples were represented by approximately equal quantities of total phenols, based on the estimates obtained from the original collected samples. The samples to be compared were spotted adjacent to each other along with a known mixture. Procedures for Comparing the Commercial "Liquid Smoke" Solutions Three commerical "liquid smoke" solutions, designated "A", "B", and "C", were diluted 1 to 20 with deionized distilled water. The pH measure- ments and total phenol estimations of the diluted solutions were made by the procedures previously described. A 100 ml aliquot of each of the diluted solutions was then subjected to the extraction procedure outlined in Figure II. Following extraction and Concentration, the phenolic fractions were chromatographically analyzed andthen qualitatively compared by use of the resulting chromatograms. RESULTS AND DISCUSSION Estimation of Total Phenols The indophenol method (Gibbs, 1927) Of quantitating phenols was described by Stanley at al. (1964) as being too laborious for routine analysis because it required constant temperature and up to 24 hours for completion of the color reaction. The results of a preliminary study to determine the time-temperature requirements for completion of the indo- phenol reaction in the method of Tucker (1942) are summarized in Figure III. These results indicated that the formation of the indophenols of phenol, guaiacol, 2,6-dimethoxyphenol, and an unknown smoke sample was completed in 2 hours when carried out at a temperature of 38°C. The rate of the reaction is also influenced by pH, and according to Wild (1953), the optimum condition lies between pH 9.4 and 9.8. Gibbs (1927) reported that over 120 minutes were required for the completion of the reaction at a pH of 9.0, while at a pH of 9.5 the time was reduced to about 85 minutes. The borate—buffered reaction mixture in Tucker's method was Observed to have a pH of 9.2 i .1, and thus, its completion time requirement of 2 hours agrees quite closely with the findings Of 'Gibbs. Standard curves for phenol, guaiacol, and 2,6-dimethoxyphenol are presented in Figure IV. A linear relationship between optical density and concentration is observable in each of the curves throughout the concentration range of .25 to 1.0 mg/100 ml. It is Obvious that for any given sample, the estimate based on the 2,6—dimethoxyphenol curve would -49- —-O- phenol (.25 g/100 ml) -£3- guaiacol (.25 g/100 ml) .30.. . .{J— 2,6-dimethoxyphenol (.25 g/100 ml) +unknown sample of collected smoke .25.. /O/G W .20“- a to. 2 co 1:3 A "a 015 - Q Q M .3 +1 3.10- (’:1"”‘,,,-__c3- °I_“ ‘t-E}-—- .05— ! l I I 2 3 4 ‘5‘ Development Time, hours H- Figure III. Formation of indOphenols vs. time at 38°C by Tucker's method. (average of duplicate runs) n -. _ -.-.. .-- «J- “EJ- J‘... . “L: __.-.'_>__‘... _. ._ .{)-. . ' us“- -- -———-—'-":_—_]“ .. -'E_. I _ - -- \ 3 E '- - " .- l?‘ 3 ‘ 5" - ~._...¢' ._ \I -50- L1— —O--phenol . *A-guaiacol 1°C_— -—43—-2,6-dimethoxyphenol .9 _. 08 — . .7 —- a . no «3 <0 .6 __ 33’ .3 I. C: 05 — ‘ :3 “R o .4 *- ‘L‘ A 8‘ 03 - - 0 “’ 02 "' t I A. I .1 " I I I I I I I I I I I I .l .2 .3 .4 5 6 7 .8 1.0 1.1 m1 of solution Mg/lOO Figure IV. Standard curves for phenol, guaiacol, and 2,6-dimethoxyphenol. I”? (D -51- be almost twice as large as that based on guaiacol, and nearly four-fold greater than that based on phenol. Simon at al. (1966) used 2,6—dimethoxye phenol as the reference compound and measured the indophenol absorbance (Optical density) at 580 mu in their studies of smoke samples collected in artificial casings. Any one of the three knowns could have been used for the purpose Of estimating the relative quantities of total phenols in the various collected samples, but phenol was selected because a wider range Of Optical densities was covered by its standard curve. The estimates Of total phenols in the samples collected for a parti- cular study are included in the resubs and discussion for that study. These estimates were very valuable in "equalizing" the various samples, on the basis of total phenol content, for their application to thin-layer and paper chromatograms. It was impossible to control the rate Of smoke production closely in either the laboratory model generator or the commer- cial generator, and thus, considerable differences in the total phenol content of the samples collected for a particular comparison were reason- able and expected. The samples collected in the commercial smokehouse were variable, in addition, because of the different amounts of water lost from the casings during the collection runs made under various con— ditions Of temperature and relative humidity. The only known way Of compensating for these differences was by attempting to represent all samples in a comparison by the same total quantity of phenols. This reasoning was validated by the assumption that all smoke produced in each generator had a unique phenolic quality which could be altered in the -52- final samples only by the conditions of collection and not by the quantity collected. Direct comparisons Of samples produced by different generators were not considered to be legitimate. Separation and Identification of Smoke Phenols Gas-liquid chromatography (GLC) The GLC separation of the components in the phenolic extract of a smoke sample is illustrated in the composite chromatogram (Plate I). The retention times of the unknown sample peaks and of probable corresponding known phenols are presented in Table I. On the basis of GLC retention times, twelve phenols were tentatively identified in the unknown sample. A minor peak (NO. 1) corresponded to the methyl ether of guaiacol, veratrole, while other minor peaks did not coincide with any of the known phenols used. The major peaks in the unknown sample corresponded to guaiacol and 2,6-dimethoxyphenol and their pa§_-substituted derivaives. In addition to the comparison of retention times, small quantities of corresponding knowns were injected with the unknown sample. In each case, the addition of the known phenol increased the size Of the corresponding unknown peak but did not disturb its symmetry. The use of three different columns proved to be cambersome, but it appeared to be the best alternative since the chromatograph was not equipped for temperature programming. Peaks Of compounds requiring more than 25 minutes for elution from any Of the columns were extremely broad and flat, therefore, they were of little value in the identification procedure. —53- Plate I Column characteristics: Column A 10% Carbowax - 20M on Diaport W.A.W. (60/80 mesh) Column B (B1) 10% Mannitol on acid—washed Chromosorb W (60/80 mesh) Column C 5% Mannitol on acid—washed Chromosorb W (60/80 mesh) All columns - 6 ft x 1/4 in. aluminum tubing Column operating conditions: Column A 9 91 9 Sample injected (ul) 1 .5 1 2 Column pressure (psi) 20 20 20 20 Argon flow rate (ml/min) 87 83 78 82 Preheater temperature (°C) 250 200 225 250 Column temperature (°C) 230 100 160 160 Cell temperature (°C) 240 200 210 225 Cell voltage 1250 1250 1250 1250 Gain (sensitivity) 10 10 10 10 Recorder chart speed (in./min) .5 .5 .5 .5 —54— fiplvent (ether) RECORDER RESPONSE Column C 10 11 J 13 O '4 8L 1'2 Ib ab 24 Retention Time, minutes Plate I. Composite GLC chromatogram showing the separation of components in the phenolic fraction of a NaOH—collected whole smoke sample. -55... .H 3on ea weapon 3 oeoamohhoo phoned: Home n.ma «.ma - - - - - - Hoeooeoo ma o.oa e.os o.oH w.oa - - - - Hoaooeeosxoaooe-m NH m.w o.m e.w m.w - - e.om m.om HooonaasHHo-e-sxoerosee-o.m Ha H.o a.m s.s o.s - - m.oH H.oa Hooooaasero-e-sxooeoseo-o.N OH o.e o.e H.n N.n - - w.ma a.mH Hoaoaaaseeoe-e-sxoeeoaae-o.m a m.m N.N m.m H.m - - o.HH a.oe Hoooeasxooeosee-o.m w - - H.N H.m - - m.w s.w HoooeoamasHHo-e s - - e.H e.H o.em H.em H.o a.o Hoooeoswasoeo-e o - - - -- m.eH m.ee m.nv o.n HoooeoaMHsaeos-e n - - - - m.o m.o m.mw H.m Houono-o a - - - - a.n a.n m.nw H.n Hoaoee m - - - - e.n m.n o.e a.m Hoooeosu m 1.- ..- .... In a. 3v m .d m .m w .m oHofivng H are as so: as are on so: on Seeeoooe oseeoeeoe o.oz o oaseoo Hm reason m reason a reason room awesome emcee”? coerced-ban .monsOQEOO A35 855” wogeommoSHoo canonoem mo pom mama-mm oxoam EEC goo; was, Mo convene.“ 0.228an 05. mo mesooomeoo newshomom Mo woe-.3 cowvooeoh 0.5 .H manom- -56- Thin-layer chromatography (TLC) The success Of TLC techniques in the identification of smoke phenols depended heavily on the characteristic colors Of their p—nitr0phenylazo dyes since complete separation of the phenols was not attained. Plate II contains TLC chromatograms of the compounds (guaiacol, 2,6-dimethoxy- phenol and their‘papa—substituted derivatives) which were tentatively identified as the major peaks in the GLC chromatograms of an unknown sample. It is obvious from.Plate II that their separation was incomplete; however, by utilizing their characteristic dye colors they could be iden- tified as four distinct bands in the following ascending order when chromatographed together: 2,6-dimethoxyphenol (dark purple), 2,6- dimethoxyphenol papa derivatives (cream), guaiacol (rose pink), and guaiacol papa derivatives (cream). These four may be observed at the top of the known mixture (C) and the unknown samples (A, B, C) of the chromatogram.of Plate III. Other bands (spots) in unknown samples of Plate III correSponded by color and position to the following phenols in the known mixture: catechol and/or 3-methoxycatechol (A, B, D, E), pyrogallol (B, E), and resorcinol (D, E). The latter two appeared to be present in relatively minor quantities. The bands corresponding to guaiacol, 2,6-dimethoxyphenol, and their papa derivatives were extracted from.a preparative-type TLC chromatogram and re-chromatographed by GLC using Column A. The resulting GLC chromato- gram.contained only peaks with retention times corresponding to guaiacol, 4-methy1guaiacol, 4-ethy1guaiacol, 4-allylguaiacol, 2,6-dimethoxyphenol, -57.. Plate II. Thin—layer chromatograms of guaiacol and 2,6-dimethoxyphenol and their para—substituted derivatives run individually and as mixtures. II-a. Guaiacol and derivatives (10 ug of each) £22323 Identity RF value Color of azo dye A Guaiacol .74 rose pink B 4—methy1guaiacol .76 cream 0 4-ethylguaiacol .77 cream D 4—allylguaiacol .76 cream E mbdme(A,B,C,m II—b. 2,6—dimethoxyphenol and derivatives (10 ug of each) ppjppp Identity RF value Color after spraying A 2,6—dimethoxyphenol .70 dark purple B 4-methy1—2,6-dimethoxyphenol .71 cream C 4—ethyl-2,6-dimethoxyphenol .72 cream D 4—allyl-2,6-dimethoxyphenol .72 cream E mixture (A, B, C, D) Plate II. _ 58- 7 I. q... . egg”. " _-.r II—a II—b .oomOHINOH so mmeflmoo Nowoz ea covenaaoo oxoem Sacra Seem Samson neosxob - m .somoaaafl awash one op nopcsosow zaovesonoa one woflpoooooo morons comma one he Hams one so ooefimomoo mooqcmeSm hhsov.eohm mamaom qsosxob - Q emcee 1 w Mo mo>eeo>esoo mmmm a mafia once as. Hooowosw w sense 1 o Mo mo>fi9o>fieoo mmmm b flea-a shoe 8. Hoooeaoaofiofio its me zoom Pamea om. oqooflsoosozn m oHom mm. Hosflohomos a ”W anew Hsoosoto ma. HorooemOhxonpoE1m .Ho:ooeoo m.m more no. Hoes-owes? H wmemohmm sopmo soaoo mmwmmJJm Nmmmmmmm dmm "mBOHHom mm poufisopooeoto Amoco Mo w: OHV mHooosm ssoox mo chopxflz.- o comm so hope: Se ooeooaaoo cream oHo:s_Eosm oaasom secure: - m com-m we sown: SH oopooaaoo oonm odo:3.fiosm panama secsRSD - < .HHH owcam -60.. .moHanm cream osoexno mo moonwoosm Oflfiocosa snow one wHoconm stock we ohsexfia a Mo sosmoemthno one .HHH oPMHm _51- 2,6-dimethoxyh4-methylphenol, 2,6-dimethoxyh4~ethylphenol, and 2,6- dimethoxy-4—allylphenol. The band corre8ponding to both catechol and 3- methoxycatechol was treated in a similar manner using Column C. The resulting GLC chromatogram.contained only two peaks and their retention times confirmed their identity as catechol and 3—methoxycatechol. Phenol and pfcresol were not distinguishable in the unknown sample on the TLC chromatogram. However, they were present in barely high enough concentra- tions to be extractable and yielded very minor peaks when re-chromatographed by GLC on Column B Operated at 100°C. Phenol chromatographed by TLC had an'R.-va1ue similar to 2,6-dimethoxyphenol and was orange in color after f being sprayed. Q—cresol had an Rf-value Similar to the guaiacol papa derivatives and was pale purple in color. The use of TLC resulted in the confirmation of the identity of ten of the twelve compounds tentatively identified by GLC and in the tentative identification of pyrogallol and resorcinol. The TLC Rf— values were reproducible only when great care was taken in activating the plates equally and changing the solvent regularly. The colors of the penitrophenylazo dyes were quite stable in dry, dark storage for a considerable period of time. Paper chromatography The use of paper chromatographic techniques for the separation and identification of the higher boiling smoke phenols (di- and trihydroxy) proved to be more successful than either GLC or TLC methods. Plate IV illustrates the separation of high boiling known phenols by paper chroma- tography. As in the case of TLC, the characteristic color tones Of the -62— p-nitrophenylaZO dyes greatly facilitated the identification Of phenols in the unknown samples. In addition to the knowns shown in Plate IV, other phenols previously identified by GLC and TLC were characterized as follows by paper chromatography: guaiacol and 2,6-dimethoxyphenol both had Rf-values very similar to phenol, but guaiacol was dark purple in color after spraying and 2,6-dimethoxyphenol was bluish green; the papa derivatives of both guaiacol and 2,6-dimethoxyphenols appeared as a yellowish green streak just above the parent compounds. Based on a comparison of their positions and dye color tones to those of phenols in the known mixture, the following phenols were identified in a whole smoke sample collected in a waterbfilled NOjax casing at 93°C: pyrogallol, resorcinol, hydroquinone, catechol, 3—methoxycatechol, 2,6-dimethoxy- phenol, and the papa derivatives of 2,6-dimethoxyphenol, as a group. The characteristic dark purple color Of the guaiacol dye was identifiable in a smoke sample collected in NOjax casing at 279C. The papa derivatives of guaiacol were observed as a greenish yellow streak just above guaiacol. The paper chromatography Rf-values were highly reproducible and the p-nitrophenylazo dye colors were very stable for long periods Of time. Table II contains a summary of the identification of phenols in smoke samples by GLC, TLC, and paper chromatography. Only two (3-methoxycatechol and 2,6-dimethoxyh4-allylphenol) of the fifteen phenols listed in Table II had not previously been identified in wood smoke according to the litera- ture. Some of the chromatograms indicated that other phenols were present, but they were not identified either because they were present in too low concentrations, or because none Of the Obtained standard compounds (Appendix In I: Amlmv oSSPKHS H oHow mo. Hoewosamohoasm m eoaaoh mama om. HoHHdMthm w Geese mm. Hoofiosomos m MW new wswfifi mm. ocooflsvosozt m zoom smflsan ow. Homeowoo m knew as. HorooeoozxonwoSIm o #oHOH> made on. Hodosm m I: an Am-mv manexfle 4 he as. ,8 .880 13 g aotba "macaaom we oosfisoeoosono who Asooo mo ma oav adenomeoo ozone one .>H owofim —64- 6.59% .m no one haopohomom ass 30:23 backwash use .46 .1882 855“ Mo gmovdfionto hon-mm .bH 3on a a a. w . .C. GI for-Ana r‘ 0". —65— II) definitely compared to them. In both the TLC and paper chromatograms there were unidentified compounds between the catechol and 2,6-dimethoxya phenol bands which were suspected as being either quinones or derivatives of catechol based on their yellowish color tones which were observable even prior to spraying on the TLC chromatograms. Table II. Phenols identified in maple sawdust smoke samples by chroma— tographic techniques. Phenolic compound Methoda(s) of identification phenol GLC (1 column), TLCE o-cresol GLC (1 column), TLC guaiacol GLC (2 columns), TLg, Paper 4-methy1guaiacol GLC (1 column), TLC b 4—ethy1guaiacol GLC (2 columns), TLCb 4—allylguaiacol GLC (2 columns), TLC 2,6-dimethoxyphenol GLC (3 columns), ’1‘ch Paper 2,6-dimethoxy—4-methylphenol GLC (3 columns), 7ch 2,6-dimethoxy-4-ethylphenol GLC (3 columns), 73ch 2,6—dimethoxy-4-allylphenol GLC (3 columns), T catechol GLC (1 column), TLC , Paper 3-methoxycatechol GLC (1 column), TLC , Paper resorcinol TLC, Paper hydroquinone Paper pyrogallol TLC, Paper aTLC and paper chromatography utilized the characteristic color tones of the p-nitrOphenylazo dyes in addition to position (R f—value) for bidentification. bSamples were chromatographed by preparative TLC, extracted, and re- chromatographed by GLC. -66- Qpalitative Comparison Studies of the Phenolic Fractions of Various Smoke Samples by Chromatogpaphic Techniques Based on the results Of their use in the separation and identifi- cation of smoke phenols, the chromatographic techniques were utilized as follows in making the various qualitative comparisons: (l) differ- ences in the relative quantities of guaiacol, 2,6-dimethoxyphenol, and 2,6-dimethoxyb4-methylphenol (compounds with largest individual peaks) were assessed by their relative peak areas on GLC chromatograms (Column A); (2) gross differences in the relative quantities of the high boiling phenols were detected by a visual assessment of paper chromatograms with known and unknown samples run adjacently and sprayed to form the p-nitrophenylazo dyes; and (3) confirmation of the findings of the above techniques was supplied by a visual assessment of TLC chromatograms in the study of whole vs. vapor phase smoke. Whole Smoke vs Vapor Phase Study This study was conducted to qualitatively compare the phenolic fractions of whole smoke and vapor phase smoke both of which were pro- duced in a laboratory model smoke generator. The vapor phase smoke resulted from.the removal of the particulate phase of whole smoke by use of an electrostatic precipitator. Smoke samples for the comparisons were collected by the different systems and at the different temperatures in order to partially compensate for any differences in the qualitative profile Of the phenolic fraction caused by collection conditions. The total phenol estimates for the samples collected in this study are presented in Table III. -57- Table III. Estimates of total phenols in samples collected for whole smoke vs. vapor phase smoke comparisons. Total phenolsa (as phenol), nglOO m1 collection solution Whole smoke Vapor phase smoke Collection system.and temperature Run I Run II Run I Run II 1N NaOH in impingers at 2-3°C 4.2 4.6 4.3 4.1 H20 in impingers at 2—3°C 5.0 4.7 4.8 5.1 H20 in impingers at 38°C 3.7 3.5 3.3 3.4 HZO-filled Nojax casings at 102-105°C 2.9 2.7 3.1 2.9 aEstimates for all samples collected in the gas impingers are for contents of lst impinger only. In the case of the samples which were collected in the gas-type impingers, it was found that the estimate of total phenols for the 2nd impinger was only 10-30% of the estimate for the first impinger. Since preliminary chromatographic analyses revealed there was no advantage in qualitatively analyzing the contents of both impingers, only the contents of the first impinger were extracted in this study to avoid dilution of the collected sample. It is apparent from the estimates shown in Table III that there was considerable variation in the total phenol content of the various samples even though all runs were of equal duration and the sawdust fed to the hotplate was regulated as uniformly as possible. In some cases, the difference between duplicate runs was greater than the difference between the two types Of smoke. This precluded any possibility of making valid comparisons between the total phenol contents of the two types of smoke. -68- Plate V is a composite GLC chromatogram of phenolic fractions from whole and vapor phase smoke samples collected in 1N NaOH at 2-3°C. The relative quantities of guaiacol, 2,6-dimethoxyphenol, and 2,6-dimethoxye 4—methylphenol (peak NOs. 2, 8, and 9) are listed in Table IV. Each value listed for a compound corresponded to the percentage contribution made by its peak area to the combined peak area for all three compounds in the chromatogram of a particular sample. Peak area and quantity injected (weight basis) was proportional for known samples of each of the three phenols. Thus, the peak area percentages were translated to quantity percentages in the table. The data in Table IV indicate there was very little difference in the relative amounts of guaiacol, 2,6-dimethoxyphenol, and 2,6-dimethoxy -4-methylphenol between whole smoke and vapor phase smoke collected in the same manner. The slight increase in guaiacol and decrease in the other two phenols in the vapor phase smoke was probably due to a slight removal of the heavier molecules by the precipitator. It is evident from Table IV that altering the collection procedure altered the relative quan- tities of these three phenols in the samples. It appears that the NaOH may have been slightly selective for guaiacol over the other two phenols when compared to water at the same temperature (2-3°C). The difference, if any, was very small and might be explained on the basis of guaiacol having a greater solubility than the other two phenols in NaOH. Increasing the temperature of the collection medium (H20) resulted in a considerably lower relative amount of guaiacol. This might have ~69— HocosaHsHHs-e-sxocecase-o.m Ha accosaascso-e-sxoceoeee-o.N oH Hococaasseos-e-sxocoosso-o.m a Hococosxocroaeo-c.m w Hoocficsmazfiao-d b HoooeoSMHmseo-m m HoomHoSMHmconId A m HomosO-m M a Hococm M m Hoooecsm m Ao>fiwoPSopv oHosvceo> H hvfiwoooH .oz moom eosmoeceOSSO ea wxoom mo 59HPSooH Am :ESHOU mo macawfioooo moflPoSomo one mooewoOHmfioomm hog H madam oomv .> owefim .o.m-m we mocz 2H SH povooHHoo oonm omega homo> one oHoss mo mGOHpOohm OHHoeosm mo A< GESHOOV thmopoEOMSO 0am oprogaoo .> owch om 0m 0 1 I 1 TAP, all! H among home> é m mm no n. m as . H m_ m H LII-II) d u d d c 100 HH - mm H an m macaw OHo:z Y Asocoov scooaom -71- Table IV. Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6- DMP), and 2,6-dimethoxy-4-methylphenol (2,6-DMé4-MP) in samples of whole smoke (WS) and vapor’phase (VP). Relative quantities (percentages) Type of Collection system.and temperature smoke G 2,6-DMP 2,6-DMé4-MP 1N NaOH in impingers (2-3°c) ws 43 32 24 VP 50 E 28 20 H20 in impingers (2-3°C) ws 39 34 26 VP 42 31 26 H20 in impingers (38°C) WS 26 39 33 'VP 29 38 32 H 0~filled Nojax casings WS 3 63 32 (102-10 5° c) VP 6 6 2 31 aAverages of samples from.dup1icate collections. resulted from.the increased temperature serving to reduce the precipitation of the higher boiling compounds in the collection train prior to reaching the collection medium. A less likely explanation would be that some of the guaiacol was lost by oxidation at the elevated temperature due to the turbulent mixing of the collection medium.with air (smoke) in the impingers. The relative amounts of the three phenols were radically altered in the samples collected in the water-filled NOjax casings. Here the alteration may have been due to a combination of temperature and casing-screening effects. The low relative concentration of guaiacol was probably charac- teristic of the smoke, as produced, since little condensation of the higher boiling compounds would have occurred prior to collection. However, the large increase in the relative amount of 2,6-dimethoxyphenol compared to -72- 2,6-dimethoxy—4—methylphenol is difficult to explain unless the casing was a greater barrier to the larger molecule under the particular condi- tions existing. A probe for the possible selectivity of casings was made in later studies of relative humidity and temperature effects. Plate VI is a paper chromatogram of whole and vapor phase smoke samples, collected by three different methods, together with mixtures of known phenols. No gross qualitative differences are evident in the chroma- tographic patterns of whole smoke and vapor phase smoke samples collected in the same manner. Considerable differences are obvious, however, in the patterns of samples collected differently. The patterns of samples A and B (H20 at 2-3°C) contain rather prominent spots of pyrogallol, hydroquinone, catechol, and 3-methoxycatechol while the patterns of samples D and E (1N NaOH at 2-3°C) contain no spots to indicate the presence of the high boiling phenols. This does not mean that samples D and E were void of high boiling phenols; rather it indicates that, if present, they were in much lower concentrations than in A and B. The reason for this difference was probably due to the fact that the high boiling (di- and tri-hydroxy) phenols were quite susceptible to electr0philic substitution reactions in the strong alkaline conditions of the collection medium (1N NaOH) and were converted to more complex products during the 5 hr run. The patterns for G and H (H20~filled NOjax at 102-105°C) indicate a very high concentration of catechol and 3-methoxycatechol, but relatively minor amounts of hydroquinone and pyrogallol. The dark spot corresponding in location to 2,6—dimethoxyphenol (7) in the known mixture appeared as sea-NS so some... .aHsz scam-one cm 83.38 Sean ones. can: oomOH-NOH Hm awSHmco Ndwoz ooHHHm-omm SH oopooHHoo oHaScm emcee homob I {I‘- I (D o we deem I E:- oom-m Ho mocz 2H SH ooHooHHoo oHQEow oMofiw oHonz I m com-m Ho mocz 2H SH USHOOHHOO oHaeom omega somc> I a SOHHoz SmHSoohw mm.-ob. Any mo mo>HHo>HSoo mmmm w coosm end-3 cs. Hococgooeofio-on s knew SS. Hosoovoohxo£HoEIm m mm hosw SwHSHn ow. Ho£oovoo m _ 88. eon-.3 mm . successes-so e Sacha mm. HoSHoSomos m BOHHoH oHcm om. HOHHowoezm m oHow mo. HoSHoSHwOSOHSm H ohm one go HoHoo mm HHHHSooH 4mm "meoHHom we oonHSoHooscto Acoco mo ma OHV mHOSoSQ SsOSM Mo oSSHtz I o com-m Ho 0mm SH popooHHoo oHasom oxoem oHosa - m com-m Ho 0mm SH oOHooHHoo oHaSow omenm somo> 1 < .mSOHHSHom oowooHHoo SH mHOSoSm HoHOH Mo moHoSHHmo So women moHHHHSosv Hosoo zHoHoSHNosmme SH ooHHmmo moHaSom SeonSb .H> oHch -74— .mooSHoE HSoho%MHo OOSSH an ooHooHHoo onaEdm oMofim omega homob oSo oHoss Mo wSOHvodhm SHHoSosm ho EoSMOHoEOSSO Sodom = a u. m... o m < .H> oHSHm -75.. follows in the unknown samples: A and B---bluish purple indicating both guaiacol and 2,6—dimethoxyphenol; D and E—--strongly purple indicating a preponderance of guaiacol; and, G and H—--strong1y bluish green indicating a preponderance of 2,6-dimethoxyphenol. Plate VII is a TLC chromatogram.cf samples of whole and vapor phase smoke collected by two different methods along with a mixture of known phenols. Again, there are no major qualitative differences evident in the patterns of the whole smoke and vapor phase samples collected by the same method. This chromatogram serves to support the data in Table IV by showing a preponderance of 2,6-dimethoxyphenol over guaiacol in the samples (A and B) collected in the waterafilled NOjax casings at 102-105°C, while in the samples (D and E) collected in 1N NaOH at 2-3°C, the proportions of these two phenols were essentially reversed. The results of these chromatographic analyses strongly indicate that there are no major qualitative differences in the phenolic fractions of whole smoke and vapor phase smoke. Thus, it may be postulated that any desirable properties possessed by smoked foods and attributable to the smoke phenols can be imparted to foods in the same degree by the vapor phase alone as by the whole smoke. These studies further indicate that the methods used for collecting smoke samples can have a decided effect on the relative proportions of individual phenols in the collected sample. com-m Ho mooz 2H SH oopooHHoo OHmEdm mecca Homc> 1 m Dom-m Ho momz 2H SH ooHooHHoo oHaSom cream oHoSB - n Smoke 1 Amv mo wo>HHc>HSoo mmmm a xSHS once as. HoooHoSw w Goose - Amy Mo mo>HHo>Hsoo mmmm h ceased mess os. HocoeaaxoeeosHo-o-m o _ anew HSMHH mm. oSoSHsoosozS m % oHow mm . HoSHohov-os S knew HoooScSO mH. HorooHcohxotpoE-m .HosooHoo m.m Hess mo. HOHHomoshm H use one to cease mm “mascooH 4mm "mSOHHom we poanoHocsono Anode mo w: OHV mHoSonm Seon Mo oHSHNHS I o e.mOH-NOH es nuance stoz assess-one as ooeooaaoo season cacao oases - m oomOHINOH Ho wSHmoo Ndwoz ooHHHMuomm SH ooHOOHHoo ngaom omega Soao> I a .mSOHHSHow UoHOOHHoo SH mHoSocm HoHOH Ho moHSSHHmo So oomcn moHHHHSosw Hosoo zHoHofiHMOHmmd SH ooHHmme moHaSom SachSD .HH> oveHm -77- .mooSHoS HSopoHMHo 03H an ooHooHHoo oxoaw omega homm> oSo oHoss mo mSOHHooSM OHHoSoSm Mo SSSMOHMEOSSS oqe < .HH> oHSHm -78- Relative Humidity and Temperature Studies These studies were made to determine if relative humidity and temper- ature had an appreciable effect on qualitative profile of the phenolic fractions of smoke samples collected in water-filled artificial casings. A practical application was envisioned for these studies since (1) smoke- house temperature and relative humidity are the most closely controlled conditions in commercial smoking Operations, and (2) water-filled arti- ficial casings provide a reasonable model system.of encased comminuted meat products commonly smoked. The influence Of casing could not be fully evaluated. However, attempts were made to obtain a partial assessment of casing effect. The control sample from a water-filled NOjax casing heated in a drying oven at 98°C for 2 hours was subjected to chromatographic analyses and no phenolic compounds were detected. This indicated that the cellulose casing did not contribute to the phenolic content of unknown samples, however, it was impossible to simulate the possible effects of harsh smoke compounds in conjunction with heat. A discussion of the results of the relative humidity study will be made first, and it will be followed by the results of the. temperature study. General data for the samples collected in the humidity study are pre- sented in Table V. The data in Table V reveal that the ClearbZip casing» lost consider- ably less water by evaporation than the NOjax casing. This was probably due to two factors: (1) the Clear»Zip casing was thicker and would be a _79- Table V. General data for smoke samples collected in water-filled artificial casings for the relative humidity study (Temperature 83°CL Relative TOtal phenols (as phenol) humidity Water loss (%) Ins/100 m1 Casing type (%) Run I Run II Run I Run II (Frankfurter) NOjax 7 70 67 5.7 5.4 Nojax 20 53 54 5.6 5.5 NOjax 40 33 31 4.4 4.6 NOjax 60 22 22 3.4 3.7 (Bologna) Clear-Zip 7 19 18 1.9 2.1 Clear-Zip 20 14 16 1.3 1.4 ClearaZip 40 7 8 1.2 1.1 Clear-Zip 60 6 5 1.2 1.0 stronger barrier to the passage of water, and (2) the water-filled Clear— Zip casing had a surface to volume ratio only one-fourth as great as that of NOjax for any given length segment (neglecting end area which would be negligible). Also from.Table V it appears that the water-filled NOjax casings were more efficient in collecting phenols although it is diffi- cult to make comparisons because of the wide variations in the amount of water lost. If there was a greater uptake of phenols by the NOjax casings, it was probably due to a combination of greater surface—to— volume ratio and thinner wall. The relative quantities of guaiacol, 2,6-dimethoxyphenol, and 2,6- dimethoxye4-methylphenol in the samples collected for the relative humidity study are listed in Table VI. 3 The data in Table VI indicate that there was essentially no differ- ence in the relative quantities of guaiacol, 2,6-dimethoxyphenol, and -80- Table VI. Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6- DMP) and 2,6-dimcthoxys4-methylphenol (2,6-DMA4AMP) in samples collected in water-filled artificial casings for the relative humidity study (Temperature 83°C), Relative humidity Relative quantitiesa (percentages) Casing type (%) G Zip-DMP 216-DMH44MP (Frankfurter ) NOjax 7 5 52 42 NOjax 20 3 53 43 NOjax 4o 2 56 4o NOjax 60 3 55 41 (Bologna) ClearbZip 7 7 52 40 Clear-Zip 20 5 53 41 Clear-Zip 40 2 55 42 Clear-Zip 60 4 57 38 aAverages of samples from duplicate collections. 2,6—dimethoxye4—methylphenol in samples collected at different smokehouse relative humidities. It is also obvious from Table VI that the relative amounts of these phenols were not measurably altered by collection in different types of artificial casings. The data for the samples Obtained by washing the outside of the Nojax casings are presented in Table VII. These samples were taken to determine if the casing was selectively screening the smoke phenols. Assuming that none of the smoke phenols adhere to the casing more strongly than others, this sample should be the most representative of the smoke as produced. The data in Table VII indicate that the relative quantities of 2,6- dimethoxyphenol compared to 2,6-dimethoxyphenol-4-methylphenol were slightly lower than in the samples collected in the water inside the casings (Table VI). This slight difference tended to indicate that the -81- casing was more impermeable to the larger molecule. Table VII. Average data from.dup1icate samples Obtained by washing the outside of NOjax casings used in the relative humidity study (Temperature 83°C). Relative humidity Estimate of total phenols Relative quapfities (percentages) (%) (as-phenolll ms/lOO m1 5&7 2,6-EMF 2,6-DMH4—MP 7 0.74 l 50 ' 48 20 0.63 1 49 49 40 0.50 3 49 47 60 0.52 2 48 48 aGuaiacol b2,6-dimethoxyphenol C2,6-dimethoxy-4—methylphenol Plate VIII is a composite paper chromatogram.cf the phenolic fractions of samples collected in water-filled Nojax casings and of those Obtained by washing the outside of the same casings. The patterns in this composite chromatogram revealed no major differences in the qualitative profiles of the samples collected at the different relative humidities whether inside (B, C, D, E) or outside (F, G, H, I) of the Nojax casings. A comparison of (B—E) with (F—I) showed no pronounced differences in the relative proportions of individual phenols in the samples that passed through the casings during collection and those that were deposited on the outside of the casings. Each of the unknown samples in Plate VIII contained pyrogallol, resorcinol, hydroquine, catechol, and 3-methoxye catechol, in apparently similar proportions. The dark spot corresponding to 2,6-dimethoxyphenol in the known sample was a bright bluish green in every sample indicating a great preponderance of 2,6-dimethoxyphenol over guaiacol. l H as Som Home wSHmeo HHS-mes Sam Hoe-Havoc onSow “av Sam poms wSHmoo MSHSmes.SoSm poSHeHno oHaSam I :1: Auv Sow poms MSHmco MSHSmos_Sosm ooSHoHno oHaSsm - w Amy Som poms mSHmoo mSHSis.Eosm ooSHoHno OHmaom l E:- mm &b He MSHmoo Nowoz ooHHHMISoHos SH pOHooHHoo ngacw I Fr: mm Rom Ho wSHmeo Nowoz poHHHm-Soeos SH ooHooHHoo oHSSom - a mm Rom Ho MSHmoo Ncnoz ooHHHm-SoHos SH ooHOOHHoo OHmEdm - o E- e8 so use-8 oncz poise-noses s... eoooofioo see-8m - m SOHHoh SmHSoosw mm.-Sb. Abv Mo mo>HHo>HSoo mmmm w 9_- soosm sod-S E. ScsooSocesSe-on s R- _ hoSm om. HonoowoohxosHoEIm m anew SwHSHQ OS. HonooHoo m as. Ewes mm. cusses-scope- a. SSOSQ mm. 3 HoSHoSomoS m BOHHoh OHmm om. HOHHmmohzm m Rom no. Sofiosamouoaca S WNW can Ho SOHoo mm HHmHmmmM .nflm uwSOHHom mo ooNHSoHoeSoSo Aroma mo ma OHV mHoSoSQ Seon mo oSSHaHS - S mSOHHSHOm ooHOOHHoo SH mHoSosm HoHOH Ho moHoEHHwo So women moHHHHSoSo Hcsoo aHonSHXOSSSd SH ooHHmmm moHoemm SsonSb .HHH> oHon .mwSHmoo OSH Ho oonHSo SSH MSHnwcs an ooSHmHno omoSH USS wwSHmoo waoz ooHHHm-Sowms SH ooHooHHoo moHQEmw oxoam Mo mSOHHOoSm OHHOSSSQ SSH Mo Eosmomeosno Somme oHHmanoo u - .HHH> oHon - m a o a c ---S ---m -83_ -84- Plate IX is a paper chromatogram of the samples collected in the water—filled Clear-Zip casings for the relative humidity study. As with the samples collected in water-filled NOjax casings, there were no major differences in the qualitative nature of the phenolic fractions collected under the different levels of relative humidity. The chromatographic patterns of samples collected in both types of casing were very similar in all respects. The results of the relative humidity study indicate that the quali- tative profile of the phenolic fractions of smoke samples collected in water-filled artificial casings VWHS not markedly affected by either the relative humidity or the type of casing when the collections were made at a temperature of 83°C. General data for samples collected in the temperature study are presented in Table VIII. Only NOjax casings were utilized in this study since more concentrated samples were obtained with them.than with Clear- Zip casings. The samples Obtained by washing the outside of the casings were qualitatively similar to those collected inside the casings. Therefore, the effect of the casing on the qualitative nature of the collected phenolic fractions was considered to be negligible in this study also. The relative quantities of guaiacol, 2,6-dimethoxyphenol, and 2,6- dimethoxy-4-methylphenol in the samples collected for the temperature study are listed in Table IX. Plate IX. -85- Unknown samples applied in approximately equal quantities based on estimates of total phenols in collected solutions. A - Mixture of known phenols (10 ug of each) characterized as follows: 110.- 1 Identity 3: phloroglucinol .08 pyrogallol .20 resorcinol .28 hydroquinone .32 catechol .40 3-methoxycatechol .44 2,6-dimethoxyphenol .70 papa derivatives of (7) .74H.93 Color of azo dyp gold pale yellow brown light tan bluish grey grey bluish green greenish yellow Sample collected in water-filled Clear-Zip casing at 60% RH Sample collected in water-filled Clear-Zip casing at 40% RH Sample collected in water-filled Clear-Zip casing at 20% RH Sample collected in water-filled Clear-Zip casing at 7% RH -86— A n c n - Plate IX. Paper chromatogram of the pehnolic fractions of smoke samples collected in water-filled Clear-Zip casings for the relative humidity study. -87— Table VIII. General data for smoke samples collected in water-filled NOjax casings for the temperature study (Relative humidity - 30%) o _ T6tal phenols (as phenol), Temperature . Water loss (%) r mg/100 ml of collection (°C) Run I Run II Run I Run II 27 7 6 3.2 3.4 38 14 12 3.7 3.8 66 27 29 4.5 4.8 93 7O 67 5.9 6.2 Table IX. Relative quantities of guaiacol (G), 2,6-dimethoxyphenol (2,6-DMP), and 2,6-dimethoxye4-methylphenol (2,6-DMH44MP) in samples collected in water—filled NOjax casings for the temperature study (Relative humidity — 30%). Temperature Relative quantitiesa (percentages) of collection (°C) G 2,6-DMP 2,6-DMH4—HP 27 41 36 22 38 17 57 25 66 6 59 33 93 } 5 63 3o aAverage of samples from.dup1icate collections. The data in Table IX indicate quite strongly that the collection temperature altered the relative quantities of guaiacol and the higher -88- boiling 2,6-dimethoxyphenol in the phenolic fraction collected. The high relative amount of guaiacol at the lower temperatures was probably due to the condensation of a considerable portion of the higher boiling compounds in the smokehouse before they had reached the water-filled casings. The paper chromatogram (Plate X) of the phenolic fractions of the samples collected for the temperature study supports the data in Table IX. The patterns show a decreasing content of high boiling phenols as the collection temperature was decreased. The spots corresponding in location to 2,6-dimethoxyphenol in the known mixture were bluish green in samples B and C indicating a preponderance of 2,6-dimethoxyphenol, while in samples D and E they were purple which indicated that guaiacol was present in the greater concentration. The results of the temperature study indicate that collection temper- ature had a definite effect on the qualitative nature of the phenolic fraction of smoke samples collected in an atmosphere of 30% relative humidity. Qualitative Comparison of the Phenolic Fractions of Three Commercial 'quuid Smokes" This study was undertaken partly because of curiosity and partly because the "liquid smokes" were readily available. They were designated "A", "B", and "C" because very little was known about their method of preparation. The total phenol estimates for the diluted (1 to 20) solu- tions of "A", "B", and "C" were 4.7, 4.8, and 4.2 mg/100 ml, respectively. Plate X. ~89- Unknown samples applied in approximately equal quantities based on estimates of total phenols in collected solutions A - Mixture of known phenols (10 ug of each) characterized as follows: No l Identipy 3p phloroglucinol .08 pyrogallol .20 resorcinol .28 hydroquinone .32 catechol .40 3-methoxycatechol .44 2,6-dimethoxyphenol .70 papa derivatives of (7) .74—.93 Color of azo dya gold pale yellow brown light tan bluish grey grey bluish green greenish yellow Sample collected in water-filled NOjax casing at 93°C Sample collected in water-filled NOjax casing at 66°C Sample collected in water-filled Nojax casing at 38°C Sample collected in water-filled NOjax casing at 27°C _90- A B " c n” E Plate X. Paper chromatogram of the phenolic fractions of smoke samples collected in water-filled Nojax casings for the temperature study. _91- The relative quantities (percentages) of guaiacol, 2,6-dimethoxyphenol, and 2,6-dimethoxya4-methy1phenol, respectively, in each Of them were as follows: "A", 19-55-25; "B", 27-65-7; and "C", 20—58-21. Plate XI is a paper chromatogram.cf the phenolic extracts of the "liquid smoke" solu- tions. Their patterns in this chromatogram.were very similar as all three showed the presence of hydroquinone, catechol, and 3-methoxycatechol in similar proportions. In addition, the paper chromatogram.pattern for each confirmed the fact that 2,6-dimethoxyphenol was present in larger concentration than guaiacol. The results Of this study indicate that the three "liquid smokes" were qualitatively similar in their phenolic con- tents which suggests they were prepared by similar procedures. _92- Plate XI. The "liquid smoke" samples were applied in approximately equal quantities based on the estimates of total phenols in their diluted solutions. A - Sample of liquid smoke "A" to I Sample of liquid smoke "B" O l Sample of liquid smoke "C" D — Mixture of known phenols (10 ug of each) characterized as follows: Np; Identity .pg Color of azo dyp l phloroglucinol .08 gold 2 pyrogallol .20 yellow 3 resorcinol .28 brown 4 hydroquinone .32 light tan 5 catechol .40 bluish grey 6 3—methoxycatechol .44 grey 7 2,6-dimethoxyphenol .70 bluish green 8 para derivatives of (7) .74-.93 greenish yellOW' Plate XI. -93_ Paper chromatogram of the phenolic fractions of three commer— cial "liquid smoke" solutions. SUMMARY Gas-liquid, thin-layer, and paper chromatographic techniques were utilized for the qualitative analyses of smoke phenols in this study. The latter two techniques were complemented by the formation of charac- teristically colored.p-nitrophenylazo dyes on the developed chromatograms. A combination of the above techniques permitted the identification Of fifteen (15) phenols in the phenolic fractions of various smoke samples. Two of the phenols (3—methoxycatechol and 2,6-dimethoxyb4—allylphenol) had not been identified in wood smoke previously. Based on the chromato- grams of the various samples, it appeared that the major phenolic consti- tuents of the smoke produced for these studies included catechol, guaiacol, 2,6-dimethoxyphenol, and the papa derivatives of the latter two. There was evidence of the presence of unidentified phenols in low concentration in some of the samples. Qualitative comparisons were made of the phenolic fractions of various smoke samples based on their chromatograms. The relative quanti- ties of guaiacol, 2,6-dimethoxyphenol, and 2,6-dimethoxya4-methylphenol were determined from.GLC chromatograms and compared between selected samples. The paper chromatograms of the phenolic fractions of the various samples were compared visually to assess their content of di- and trihya droxy phenols. A visual comparison of the TLC chromatograms served to substantiate the findings of the other two techniques. Whole smoke and vapor phase smoke samples were collected from.a labor- atory model generator by two different systems and in two different media. —94— -95- The following Observations were made from.the qualitative comparisons of the phenolic fractions of these samples: 1. The phenolic fractions of whole smoke and vapor phase smoke did not differ markedly when collected in a Similar manner regardless of the procedure used. 2. The phenolic fractions of samples collected differently showed con- siderable qualitative differences. These results suggested that the smoke phenols were confined almost entirely to the vapor phase of the aerosol, and that the collection pro- cedure had a great influence on the kind of phenols collected for analysis. Smoke samples were collected by model absorbent systems (water-filled cellulose casings) in a commercial smokehouse operated under various conditions of relative humidity and temperature. Additional samples were obtained by washing the outside of the frankfurteratype casings to check for casing selectivity. In one study the temperature was maintained at 83°C and samples were collected using both the frankfurter and bologna- type casings with relative humidities Of 7, 20, 40, and 60%. The following observations were made from.the qualitative comparisons of the phenolic fractions of these samples: 1. The phenolic fractions were very similar for the samples throughout the relative humidity range, regardless of the type of casing used. This suggested that relative humidity had little effect on the kind of phenols collected at 83°C. 2. The phenolic fractions from the samples collected in both types of casing were very similar. —96- 3. The phenolic fraction from samples collected outside the casings were very similar to those collected inside. In the other study the relative humidity was maintained at 30% and samples were collected in the frankfurter-type casing at temperatures Of 27, 38, 66, and 93°C. The qualitative comparison of the phenolic fractions of these samples indicated that the collection of high boiling phenols de- creased with decreasing smokehouse temperature at a relative humidity of 30%. This suggested that some of the higher boiling phenols might have condensed in the smokehouse at the lower temperatures. A qualitative comparison of the phenolic fractions of three commer- cial "liquid smokes" revealed only minor differences. This suggested that the three solutions were prepared in a similar manner. The results of this study are inconclusive in many respects but do exhibit general trends as to the influence of various factors on the phenolic fraction of hardwood smoke. The qualitative comparisons based on visual assessments of TLC and paper chromatograms could detect only ' gross differences, but they were necessary since the di- and trihydroxy phenols could not be chromatographed by GLC techniques available. The trends indicated in this study seem to justify further research in this area utilizing GLC with temperature programming and appropriate column(s) to permit the separation of the entire Spectrum.of smoke phenols. This would permit comparisons based on peak areas for all the phenols, and such comparisons could be quite sensitive and accurate. BIBLIOGRAPHY Amerine, M; A., R. M. Pangborn and E. B. Roessler. 1965. Principles of Sensory Evaluation of Food. p. 151. Academic Press Inc., New York. Anonymous. 1954. Smoke curing -- chemistry. In Food Investigations Board Report 1953. Dept. of Sci. and Ind. Research (Great Britain). p. 30. H. M. Stationery Office, London. Anonymous. 1956. Smoke curing -- chemistry. In Food Investigations Board Report 1955. Dept. of Sci. and Ind. Research (Great Britain). p. 32. H. M. Stationery Office, London. Anonymous. 1958. Smoke curing -- chemistry. In Food Investigations Board Report 1957. Dept. of Sci. and Ind. Research (Great Britain). p. 16. H. M. Stationery Office, London. Bailey, A. 1947. High boiling hydrolytic derivatives of lignin. J. Am. Chem. Soc. 62, 575. Banks, A. 1950. Factors affecting the oxidation of the fat of dehydrated herrings. J. Sci. Food Agr. l, 28. Barber, E. D., E. Sawicki and S. P. MbPherson. 1964. Separation and identification of phenols in automobile eXhausts by paper and gas- 1iquid chromatography. Anal. Chem. £6, 2442. Bayer, E. 1961. Gas Chromatography. Elsevier Publishing Company, New York. Block, R. J., E. L. Durrum and G. Zweig. 1958. A Manual of Paper Chroma- tography and Paper Electrophoresis. 2nd Ed. p. 301. Academic Press Inc., New York. Brauns, F. E. 1948. In Nature of the Chemical Components of WOod. (C. J. west, ed.) p. 115. Tech. Assoc. Pulp Paper Ind., New York. Brauns, F. E. 1952. The chemistry of lignin. In Wood Chemistry. (L. E. Wise and E. C. Jahn, eds.) 2nd Ed., Vol. I, p. 409. Reinhold Pub- lishing Corp., New York. Brauns, F. E. and D. A. Brauns. 1960. The Chemistry of Lignin. Suppl. Vol. p. 616. Academic Press Inc., New York. Browning, B. L. 1952. The polysaccharide fraction of wood. 2. Analysis of non-cellulose polysaccharides. In Wood Chemistry. (L. E. Wise and E. C. Jahn, eds.) Vol. II, 1159. Reinhold Publishing Corp., New York. -97- -98— Browning, B. L. and I. U. Isenberg. 1952. Analytical data and their significance. In Wood Chemistgy. (L. E. Wise and E. C. Jahn, eds.) 2nd Ed., Vol II, p. 1259. Reinhold Publishing Corp., New York. Clark, I. J. 1964. Chromatographic separation of phenols on ion-exchange paper. J. Chromatog. 12, 65. Commins, B. J. and A. J. Lindsey. 1956. The determination of phenols by chromatography and spectrOphotometry of their methyl ethers III. The determination of phenols in wood smoke. Anal. Chim. Acta 15, 5540 Creighton, R. H. J., R. D. Gibbs and H. Hibbert. 1944. Studies on lignin and related compounds. LXXV. Alkaline nitrobenzene oxidation of plant materials and application to taxonomic classification. J. Am. Chem. Soc. 66, 32. Crouse, H. C., J. W. Garner and H. J. O'Neill. 1963. Determination of phenolic constituents of cigarette smoke by gas chromatography. J. Gas Chromatog. l, 18. Crump, G. B. 1964. Thin layer chromatographic analysis of simple alkyl phenols. Anal. Chem. 36, 2447. Draudt, H. N. 1963. The meat smoking process: a review. Food Technol. 12, 85. Fiddler, W., R. C. Doerr, A. E. Wasserman and J. M. Solay. 1966. Composi- tion of hickory sawdust smoke: furans and phenols. Agric. and Food Chem. _1_4, 659 . Foster, W. W. 1959. Technology of smoked foods. Food Manuf..§fl, 56. Foster, W. W. and T. H. Simpson. 1961a. Studies of the smoking process for foods. I. - The importance of vapors. J. Sci. Food Agric. 12, 363. Foster, W. W. and T. H. Simpson. 1961b. Studies of the smoking process for foods. II. - The role of smoke particles. J. Sci. Food Agric. 12, 635. Freedman, R. W. 1964. Quantitative gas liquid chromatography of phenols by complete trimethylsilylation of hindered phenols in presence of acidic oxides. Ana1. Chem..§§, 1389. Freeman, R. D. and F. C. Peterson. 1941. Proxinhate analysis of the heartwood and sapwood of some American hardwoods. Ind. Eng. Chem., Anal. Ed. 12, 803. -99.. Geissman, T. A. 1959. Principles of Organic Chemistry. 2nd Ed. p. 522. W. H. Freeman and Company, San Francisco. Gibbs, H. D. 1927. Phenol tests III. The indOphenol test. J. Biol. Chem. 12, 649 . Gibbons, N. E., D. Rose and J. W. Hopkins. 1954. Bactericidal and drying effects of smoking on bacon. Food Technol..§, 155. G005, A. W. 1952. The thermal decomposition of wood. In Wood Chemistgy. (L. E. Wise and E. C. Jahn, eds.) Vol. II, p. 826. Reinhold Pub- lishing Corp., New York. ‘ Handbook of Chemistry and Physics. 1966. 47th Ed. The Chemical Rubber Company, Cleveland, Ohio. Hawley, L. F. 1952. Combustion of wood. In WOod Chemistry. (L. E. Wise and E. C. Jahn, eds.) 2nd Ed. Vol. II, p. 817. Reinhold Pub- lishing Corp., New York. Hedrick, T. I., L. J. Bratzler and G. M. Trout. 1960. Lethal effect of smoke on surface microorganisms of cheddar cheese. Mich. Agr. Expt. Sta. Quart. Bul. 43, 327. Hess, E. 1928. The bacterial action of smoke (as used in the smoke- curing of fish). J. Bact. 15, 33. Hollenbeck, C. M. and L. J. Marinelli. 1963. Some chemical constituents of smoke flavoring. Proceedings of the Fifteenth Research Conference. p. 67. AIM.I.F. University of Chicago, Illinois. Husaini, S. A. and G. E. C00per. 1957. Fractionation of wood smoke and comparison of chemical composition of sawdust and friction smokes. Food Technol. 11, 499. Jahnsen, V. J. 1961. The chemical composition of hardwood smoke. Abstr. of Ph.D. Thesis. Purdue university, Lafayette, Indiana. Jensen, L. B. 1943. Action of hardwood smoke on bacteria in cured meats. Food Research 85 377. Jensen, L. B. 1954. Microbiology of Meats. 3rd Ed. p. 199. Garrard Press, Champaign, Illinois. Kochanowski, J. 1962. Bacteriostatic properties of liquid smoke. Sym- posium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 29. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. -100- Krueger, W. W. 1953. Principles of Microbiology. p. 280. W. B. Saunders Company, Philadelphia. Kuriyama, A. 1962. The thermal decomposition of woody substance (in relation to smoke generation). Symposium on advances in the engin- eering of the smoke curing process. (D. Tilgner, ed.) p. 5. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Kurko, V. 1959. Antioxidative preperties of the smoke components. Mayasn. Ind. SSSR 39, 19. (From.translation by E. Wierbicki) Kurko, V. D. and L. F. Kelman. 1962. Smoke Phenols. Publication of the VIIIth EurOpean Congress of Meat Research Institutes. No. 49. Kurko, V. 1966. The role and importance of individual smoke phenols in inhibiting the oxidative deterioration of smoked meats. Report of the 12th European meat research conference. p. 25. American Meat Science Association, Chicago, Illinois. Lea, C. H. 1933. Chemical changes in the fat of frozen and chilled meat. V. The effect of smoking and the influence of atmospheric humidity on the keeping prOperties of bacon. J. Soc. Chem. Ind. .52, 577. Miler, K. 1962. The oxidation phenomena of curing smoke components. Symposium on advances in the engineering of the smOke curing process. (D. Tilgner, ed.) p. 15. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Morrison, R. J. and R. N. Boyd. 1959. Organic Chemistry. p. 585. Allyn and Bacon Inc., Boston. Pettet, A. E. J. and F. G. Lane. 1940. A study of the chemical composi- tion of wood smoke. J. Soc. Chem. Ind. 52, 114. Porter, R. W. 1963. Fractionation and identification of some compounds in wood smoke. Ph.D. Thesis, Michigan State University, East Lansing, Michigan. Porter, R. W., L. J. Bratzler and A. M. Pearson. 1965. Fractionation and study of compounds in wood smOke. J. Food Sci. 39, 615. Randerath, K. 1963. Thin-Laygr Chromatoggaphy. p. 175. Academic Press Inc., New York. Roberts, J. D. and M. C. Caserio. 1964. Basic Principles of Organic Chemistgy. p. 904. W. A. Benjamin, Inc., New York. Shewan, J. M. 1949. The biological stability of smoked and salted fish. Chem. Ind., §_5_, 501. -101- Sikorski, Z. 1962. Changes in the curing smoke as influenced by high voltage electrostatic field. Symposium on advances in the engineer- ing of the smoke curing process. p. 24. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Simon, S., A. A. Rypinski and F. W. Tauber. 1966. water-filled cellulose casings as model absorbents for wood smoke. Food Technol. 29, 1494. Simpson, T. H. and D. Campbell. 1962. .Methods of analysis of smoke de- posits. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 90. Yugoslav Institute of Meat. Technology. Beograd, Yugoslavia. Stanley, T. W., E. Sawicki, H. Johnson and J. D. Pfaff. 1965. Comparison of methods for the spectrOphotometric determination of phenols in automobile exhaust fumes. Mikrochimica Acta 1965, 48. Stuckey, B. N. 1962. Antioxidants. In Symposium on Foods: Lipids and Their Oxidation. (H. W. Schultz, E. A. Day and R. O. Sinnhuber, eds.) p. 139. Avi Publishing Company, Inc., West Port, Connecticut. Sundt, E. 1961. Paper chromatography of phenols. J. Chromatog. 6, 475. Tilgner, D., A. Makowiecki, W. Barchowiec and DjuPHak-Dju. 1962a. A comparison of curing smoke quality as influenced by the methods of smoke production. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 40. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Tilgner, D., K. Miler, J. Prominski and G. Darnowska. 1962b. The sensoric quality of phenolic and acid fractions in curing smoke. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 37. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Tilgner, D., H. Ziminska and M. Wronska-Wojciechowska. 1962c. The chemi- cal and sensorial characteristics of the steam volatile components in smouldering and friction smoke. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 35. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Tilgner, D., K. Miler, J. Prominski and G. Darnowska. l962d. The influence of smoke production parameters upon its amount of phenolic and acid fractions. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 18. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. -102— Tilgner, D., K. Miler and J. Prominski. 1962e. The methods of estimation of phenolic compounds in curing smokes and smoked goods. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 92. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. Tilgner, D., Z. Ziemba and DquHak—Dju. 1962f. The quality and composi- tion of curing smoke as influenced by water vapour. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 21. Yugoslav Institute of Meat Technology. Beograd, Yugo- slavia. Tucker, I. W. 1942. Estimation of phenols in meat and fat. J. Assoc. Offic. Agr. Chemists 25, 779. Wesserman, A. E. 1966. Organoleptic evaluation of three phenols present in wood smoke. J. Food Sci. 31, 1005. watts, B. M. and M. Faulkner. 1954. Antioxidant effect of liquid smokes. Food Technol. 8, 158. White, A., P. Handler, E. L. Smith and D. Stetten. 1959. Principles of Biochemistgy. 2nd Ed. p. 57. MeGraw-Hill Book Company, Inc., New York. White, W. H., N. E. Gibbons, A. H. Woodcock and W. H. Cook. 1942. Smoked meats. I. Bacteriological, chemical and physical measurements on smoked and unsmoked bacon. Can. J. Res. 20D, 263. White, W. H. 1944. Smoked meats. II - Development of rancidity in smoked and unsmoked Wiltshire bacon during storage. Can. J. Res. 22F, 97. Wild, F. 1953. Estimation of Organic Compounds. p. 90. University Press, Cambridge. ‘Wilson, G. D. 1961a. Smoke composition and factors affecting smoked meat flavor. Proceedings of the Thirteenth Research Conference. A.M.I.F. Circular No. 64, p. 113. University of Chicago, Illnois. Wilson, G. D.’ 1961b. A summary of the technical aspects of smoking meats. A.M.I.F. Special Report No. 26. Wolkowskaja, I. L. and I. I. Lapszin. 1962. Bactericidal and fungicidal properties of smoke solution. Symposium on advances in the engineer- ing of the smoke curing process. (D. Tilgner, ed.) p. 26. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. ~103- Ziemba, Z. 1962. Some problems of colour development in smoked food products. Symposium on advances in the engineering of the smoke curing process. (D. Tilgner, ed.) p. 50. Yugoslav Institute of Meat Technology. Beograd, Yugoslavia. APPENDIX -104- Appendix I. Some properties of identified wood smoke phenolsa Name Structure Mol. wt. LM.P.(°Q B.P.(°C) 0H phenol [§::E] 94.11 41 182.0 H CH3 g-cresol 0 108.13 30 191.5 0H m-eresol 0 108.13 11 202. 8 H3 H p-cresol 108. 13 36 202 . 5 H3 H OH catechol 0 110.11 105 240. 0 fl resorcinol 0 110.11 110. 0 276. 5 H H hydroquinone (::> 110.11 170.5 286.2 H aAvailable data from Handbook of Chemistry and Physics (1966). -105- Appendix I. Some properties of identified wood smoke phenolsa (continued) Name Structure Mel. wt. M.P.(°C) B.P.(:C) 0H H3C H3 2, 6-dimethy1phenol 0 122.16 49 212 I 3,4-dimethy1phenol <::) 122 16 65 225 CH3 - H3 0H 3,5-dimethy1phenol ' O 122.16 68 219.5 'H HO OH pyrogallol (::> 126.11 133 309 0H pholoroglucinol (::> 126.11 109 -~- HO OH 'H CH(CH3)2 thymol O 150. 21 51 233. 5 H3C I ‘ OCH3 vanillan 152.14 81 285 .9. H0 aAvailable data from.Handbook of Chemistry and Physics (1966). -106- Appendix I. Some properties of identified wood smoke phenolsa (continued). Name Structure Mel. wt. iM.P.(fC) B.P.(°C) I 'CHB guaiacol 0 124.13 28 205 0 .CH3 4-methy1guaiacol 0 138.16 5. 5 221.8 3 CH3 4-ethy1guaiacol <::) (152.19)* -- -- 0 'CH3 4-propy1guaiacol <::) (166.22)* -- -- ( H2)2CH3 I 'CH3 4-a11y1guaiacol Q (164. 20)»? 10. 3 252 H2-CH=CH2 4-viny1guaiacol <::) (162.18)* -- -- =CH2 I H3C0 0CH3 2,6-dimethoxyphenol <::) 154.16 56 258 aAvailable data from.Handbook of Chemistry and Physics (1966). *Calculated values. -107- Appendix I. Some properties of identified wood smoke phenolsa (continued). Name Structure Mel. wt. 'M.Pa(:C) B.P.(1C) 0H H CO 'CH3 2,6-dimethoxy-4-methylphenoI (::> (168.19)* -- -- CH3 0H 2,6—dimethoxy-4-ethylphenol <::) (182.22)* -- -- 0H H3C0 OCH3 2,6-dimethoxy-4-propy1phenol (::) (196.25)* -- —~ H2)2'CHB H 1-naphthol O 0 144.17 93 288 H 2-naphthol 144.17 122 295 0 0:) aAvailable data from.Handbook of Chemistry and Physics (1966). *Calculated values. Appendix II. -108- Standard phenols and their source. Compound Source phenol Baker .gfcresol Mathieson E-cresol Mathieson .pfcresol Mathieson 2,4-dimethy1phenol Aldrich 2,5-dimethy1phenol Aldrich 2,6-dimethy1phenol City 3,4-dimethy1phenol Eastman 3,5-dimethy1phenol Eastman 2-ethy1phenol Aldrich guaiacol City 4-methy1guaiacol (creosol) Cliffs Dow 4-ethy1guaiacol City 4-a11y1guaiacol (eugenol) Aldrich 2,6-dimethoxyphenol City 2,6-dimethoxy-4-methylphenol Cliffs Dow 2,6-dimethoxy-4-ethylphenol Cliffs Dow 2,6-dimethoxy-4-a11y1phenol Aldrich catechol (pyrocatechol) City resorcinol Mallinckrodt hydroquinone Eastman pyrogallol Eastman pholoroglucinol Aldrich thymol Eastman vanillan City 3-methoxycatechol Aldrich veratrole (methyl ether of guaiacol) Eastman Aldrich = Aldrich Chemical Co., Inc., Milwaukee, Wisconsin. J. T. Baker Chemical Co., Phillipsburg, New Jersey. Baker = City = City Chemical Corporation, New York, N. Y. Cliffs Dow - Cliffs Dow Chemical Co., Marquette, MiChigan. = Eastman Organic Chemicals, Rochester, N. Y. Mallinckrodt = Mallinckrodt Chemical Wbrks, St. Louis, Missouri. Mathieson = Mathieson Coleman & Bell, Norwood (Cincinnati), Ohio. Eastman -108- Appendix II. Standard phenols and their source. Compound Source phenol Baker ‘g:cresol Mathieson Imycresol Mathieson .pfcresol Mathieson 2,4-dimethy1phenol Aldrich 2,5-dimethylphenol Aldrich 2,6-dimethylphenol City 3,4-dimethy1phenol Eastman 3,5-dimethy1phenol Eastman 2-ethy1phenol Aldrich guaiacol City 4—methy1guaiacol (creosol) Cliffs Dow 4-ethy1guaiacol City 4-a11y1guaiacol (eugenol) Aldrich 2,6-dimethoxyphenol City 2,6-dimethoxy-4dmethylphenol Cliffs Dow 2,6-dimethoxy-4-ethylphenol Cliffs Dow 2,6-dimethoxy-4-a11y1phenol Aldrich catechol (pyrocatechol) City resorcinol .Mallinckrodt hydroquinone Eastman pyrogallol Eastman pholoroglucinol Aldrich thymol Eastman vanillan City 3-methoxycatechol Aldrich veratrole (methyl ether of guaiacol) Eastman Aldrich = Aldrich Chemical Co., Inc., Milwaukee, Wisconsin. Baker = J. T. Baker Chemical Co., Phillipsburg, New Jersey. City = City Chemical Corporation, New York, N. Y. Cliffs Dow - Cliffs Dow Chemical Co., Marquette, Nfichigan. Eastman = Eastman Organic Chemicals, Rochester, N. Y. Mallinckrodt = Mallinckrodt Chemical Wbrks, St. Louis, Missouri. Mathieson = Mathieson Coleman & Bell, Norwood (Cincinnati), Ohio. ”71111111111111!fllflflffltflifllmfilflllm