384 NORTHALL-LAURIE A NEW METHOD FOR THE A NEW METHOD FOR THE DETERMINATION OF TOLUENE IN COMMERCIAL TOLUOLS. BY D. NORTHALL-LAURIE F.I.C. (Published by permission of the Controller of H.M. Stationery Ofice.) Theoretical Consideratiom.-When a mixture of two chemically similar substances infinitely miscible such as benzene and toluene and which do not form a mixture of constant boiling-point are boiled the relation between the molecular composition of the vapour and the molecular composition of the mixture is constant and can be expressed by the equation where XI end X are the relative masses of the constituents in vapour W1 and W2 the relative masses of the constituents in the liquid and C a constant which is usually a ratio of the vapour pressures. AS the constituent of lower boiling-point boils off the percentage of the higher boiling-point conetituent rises and so does the boiling-point of the mixture the composition of the vapour changing in the same manner I t thus follows that if a mixture of benzene and toluene is distilled at a uniform rate in a flask without still-head condensation into an efficient condenser the composition of the first drop condensed will be equal to the composition of the vapour and as distillation proceeds, the percentage compoaition of the distillate will be in inverse ratio to that of the mixture distilling.If half the mixture is distilled in the above manner and assuming equal quantities of each constituent present in the mixture the benzene will be in excess in the fraction distilled and toluene in excess in the residue.Again if the relationship between the percentage composition of the constituents to the boiling-point is known the amount of each constituent in the fractions can be easily determined. Mixtures of benzene and toluene toluene and xylene of known composition by volume were prepared and their boiling-poilits noted. These were plotted against the percentages and a curve drawn from which the composition by volume for all boiling-points could be read off. Mixtures of the three homologues will behave under distillation oonditions in a similar manner but in this case the relationship between the percentages of the constituents in the fractions is more complicated. In this case two fractions must be collected. The benzene will then be in excess in the first fraction the toluene in the second and the xylene in the residue; there being a proportion of the three homologues in each fraction this proportion varying according to the percentage composition of the original liquid, As the three isomers of xylene occur in commercial xylene and as the percentage of each isomer varies in different samples slightly varying boiling-points will be obtained in practice.I find however owing to the boiling-points of the three isomers being so close-namely 143" 1 3 9 O and 138" C.-the error due to taking on DETERMINATION OF TOLUENE IN COMMERCIAL TOLUOLS 385 fixed boiling-point does not affect the results in any marked degree. A more complete separation of the mixture into its three constituents would be obtained by using an evaporator still-head and collecting the fractions up to 110-6" and 139" C.instead of in equal amounts. This is of course the procedure adopted in fractional distillation. I t is thus seen that plain distillation tends to separate the three constituents, fractional distillation giving a more complete separation. However any method of analysis depending on fractional distillation is tedious and likely to be unreliable unless conducted with great care. Consequently if sufficient separation of the constituents is obtained by plain distillation into three fractions so as to give first and last fractions of markedly different boiling-point a, method of determining toluene would be possible. After a considerable number of experiments I have devised a practical method depending upon the principles above cited.Method.-A known volume of the sample is distilled at a uniform rate from a distilling4 ask into an efficient condenser. As soon as one quarter of the volume has collected the receiver is changed and a further half distilled over. The distillation is then stopped. The boiling-points of the first fraction and the last-namely the residue in the flask-are then taken in an apparatus of special design. On reference to a table or graph prepared as a result of experiment with known mixtures the quantity of toIuene is read off. This method depending upon the collection of fractions always of the same quantity and noting the boiling-point has advantages over the alternative method of collecting fractions to a definite temperature and measuring them inasmuch as in cases where there is a predominating amount of one constituent the quantity of one or more of the fractions would be small and difficult to measure accurately.Boiling-Point Apparatus,-The boiling-point apparatus required for use in taking the boiling-point of the fractions is shown in the sketch on p. 386. This can be easily constructed from a distilling-flask. In the neck of the flask is placed a thin glass cylinder with a hole in the side corresponding to the outlet of the neck of the flask and sealed to its sides. This glass screen protects the bulb of the thermometer and insures a steady temperature preventing variations from draughts or other causes. Process.-Two hundred C.C. of the sample measured at tap-water temperature are placed in a distilling-flask of just sufficient capacity with the neck cut off close above the side tube The contents are distilled at the uniform rate of 7 C.C.per minute through an efficient condenser into a 50 C.C. measuring cylinder. When 50 C.C. have collected the cylinder is quickly changed without stopping the dis-tillation for a 100 c ~ . cylinder. As soon 8s 98.5 c . ~ . have come over the flame is extinguished and the contents of the flask allowed to cool. By the time that all drops from the condenser have ceased 100 C.C. will have been collected. Should the first 50 C.C. be turbid from the presence of water a pellet of calcium chloride is added and the cylinder well shaken and allowed to stand. The contents of the distilling-flask are cooled under the tap transferred to the special boiling-point apparatus and the boiling-point taken as follows : The apparatus is connected to a reffux condenser and boiled at such a spee 386 NORTHALL-LAURIE A NEW METHOD FOR THE that the condensed fluid runs back at the uniform rate of 1 drop per second.The thermometer should be divided into tenths of a degree and should be tested to ascertain if it is accurate between temperatures of 80’ to 140° C. When the temperature has become constant - which takes about ten minutes - the ther-mometer is read the necessary corrections for atmospheric pressure and length of exposed stem made and the figure noted. The apparatus is then disconnected emptied dried out and the liquid contents of the 50 C.C. oylinder-which by this time will be quite dry-added.The boiling-point is noted in a similar manner and on reference to the graph the vertical co-ordinate of which represents the boiling-points of the first fraction and the horizontal co-ordinate of which represents the boiling-point of the last fraction, the percentages of toluene and benzene present in the sample can be read off. The difference between the sum of these percentages and a hundred represents the xylene. Advantages of the Process.-There is no need to measure the fractions very accurately as slight variations in the amounts collected will not affect the value of the boiling-points determined. This obviates the necessity of cooling the fractions to one temperature which is a source of delay in hot weather. If the sample is at tap-water temperature the fractions will necessarily be at the same temperature thus eliminating any error due to expansion DETERMINATION OF TOLUENE IN COMMERCIAL TOLUOLS 387 The only part of the process requiring special care is in the taking of the boiling-point which would present no difficulties to a qualified man.A determination can easily be completed in from thirty to forty minutes. After the first distillation is finished the boiling-point of the fractions obtained can be taken whilst the second sample iB being distilled; and consequently the process takes no longer than the time required to distil 150 C.C. at the rate of 7 C.C. per minute. Accuracy.-It will be noted by inspection of the graph that the difference in the boiling-points is quite appreciable for 1 per cent.difference in the toluene present. There is no difficulty in determining the percentages with an accuracy of 0.1 per cent. Limitations of the Process.-The new method will give direct results for all possible combinations of toluene with benzene and xylene in gamples containing from 50 to 100 per cent. toluene the accuracy of the results being greater the higher the percentage of toluene present in the sample. The value of great accuracy in determinations of toluene in commercially pure samples is apparent it being important to be able to report upon the quality of small impurities in commercial (' pure toluene " now required for the manufacture of tri-nitro-toluene. For samples containing less than 50 per cent. toluene such as 5Os/9Os benzols it would be neces-sary to add a known quantity of pure toluene so as to bring the results on the graph, This is easily done by taking 100 C.C.of the sample and 100 C.C. of the toluene the error produced by doubling the result being more than offset by the greater accuracy of the determination. There is no limitation as to the percentage of benzene or xylene present so there is no need to add anything but toluene to the sample. Parafin.-Any paraffin present in the sample would appear in the analysis as toluene and the necessary correction must be made for this material. This can be done by the method suggested by Dr. Colman-namely to take in the case of very pure toluene the gravity of the sample; or in the case of toluenes of lower per-centage the specific gravity of the fractions distilled over between the temperatures of 107' and 115' C.using a Young 12 bulb head. Every unit in the third place in the specific gravity less than the specific gravity of pure toluene,-namely 0.870-will represent 0.66 per cent. of paraffin in the toluene found. Example.-The specific gravity of a sample was 0.868. This is two units less than 0-870 2 x 0.66 equals 1.3 per cent. paraftin. Therefore if the sample contains 60 per cent. toluene the paraffin present would be 0.78 per cent. Carbon Bi-sulphide.-This material does not occur in toluenes containing over 60 per cent. but is likely to occur in 50S/9Os benzols. I t is necessary to remove this material before the process can be employed. This is easily done by means of alcoholic potash the carbon bi-sulphide being removed as xanthate.The alcohol used must be either pure ethyl alcohol or industrial spirit. Ordinary methylated spirit contains paraffins. Further Note on taking the Boiling-Point.-As the whole process depends upon the acourate determination of the boiling-point care must be taken to always take readings under similar conditions. The apparatus used for this purpose should be This difference increases with the increasing percentage of toluene 388 NORTHALL-LAURIE A NEW METHOD FOR THE of the exact size specified. The flask should be supported 34" above the top of the Bunsen burner which burner must be screened from draughts by means of a cylinder of wire gauze. The flask should rest on a square of asbestos with a hole 14" in diameter.The stem of the thermometer must be screened from convection currents by means of a square of cardboard slipped over the cork. An auxiliary thermometer should be suspended within 3" of the thermometer, SO as to give the air temperature immediately surrounding its stem. The whole apparatus should be well protected from draughts. The rate of boiling must be so regulated that the drops falling back from the reflux condenser do so at the uniform rate of 1 drop per second. The most convenient way of regulating this rate is to employ a metronome. With these precautions it is quite easy to obtain rapid observations within one-tenth of a degree. Boiling- Po& Corrections.-It is of course necessary on account of the changes in atmospheric pressure to make corrections in the boiling-point.The relationship between the boiling-point of liquids to pressure is expressed by the equation AT= (760-p)(273 x t)C where AT is the difference between the observed boiling-point and the boiling-point at 760 mm. pressure t is the boiling-point of the liquid ; p is the atmospheric pressure reduced to 0" C. and C is a constant. The values of C are-For benzene . . . 0.000121 , toluene . . . 0~000?20 , xylene . . . 0*000115 Within the usual limits encountered the mean value of this constant can be employed and the following corrections which have been determined from the above equation will apply : Pressure. 770 769 768 767 766 765 764 763 762 761 760 Correction t o be made. - 0.45 - 0.4 - 0.35 - 0.3 - 0.25 - 0.2 - 0.2 - 0.15 - 0.1 - 0.05 0.0 Pressure.759 758 757 756 755 754 753 752 751 750 749 Correction t o be made. + 0-05 + 0.1 + 0-15 + 0.2 + 0.2 + 0.25 + 0.3 + 0.35 + 0.4 -t 0.45 + 0.5 Pressure. 748 747 746 745 744 743 742 741 740 739 738 Correction to be made. + 0.55 + 0-6 + 0.65 + 0.7 + 0.75 + 0.8 + 0-86 4- 0.9 + 0.9 + 0.95 + 1.0 Stem Correctio.ns.-As a portion of the stem of the thermometer has not been entirely surrounded by the vapour of the boiling liquid another correction is required DETERMINATION OF TOLUENE IN COMMERCIAL TOLUOLS 3 8 9 0.35 0.43 0.52 before the true boiling-point is obtained The corrections to be added have been calculated from the formula 0.000143 (T - t ) N when T is the observed boiling-point, t the temperature of the stem outside flask and N the length of the mercury column in degrees not heated by the vapour.The following table embodies the values obtained and should be used for the purpose of making the necessary corrections correction to the nearest 0.05" being taken : 0.40 0.50 0.60 T - t. N. 10 20 30 40 50 60 70 80 90 100 50'. 0.07 0.15 0.22 0.29 0.36 0.43 0.50 0.57 0-65 0.72 60". I 70" 80". ' 90'. 1 100'. 0.12 0.23 0.34 0.46 0.57 0.68 0*80 0.91 1.02 1.14 0.13 0.26 0.39 0.52 0.65 0.78 0.91 1.03 1.17 1.30 0.1 4 0.29 0-43 0-57 0.71 0.86 1.00 1.14 1.29 1 -43 l l o o . -___. 0.16 0-32 0.48 0.63 0.79 0.95 1 *10 1.27 1-42 1.57 Note on the GYapL-It will be noted that besides the black lines running in a sloping direction and gradually becoming more curved as the percentage of toluene becomes less there are horizontal red lines running across the graph. These repre-sent percentages of benzene and the graph therefore enables not only the per-centage of toluene to be determined but also the percentage of benzene. The sum of these percentages taken from 100 will give the percentage of xylene. Conclusion.-The method above described is quick is applicable to all commer-cial samples gives results with an accuracy of 0.1 per cent. and is especially valuable for determining the purity of "pure toluene." It is of course only applicable to commercial toluenes that have been washed free from phenols and unsaturated bodies