ll \llllll'} J‘ I 1 I JIHIIWI ll ‘ ! I 0300 0'1\| THE MICRO DETERMINATION OF HALC’GENS BY THE CARIL‘S METHOD Tide-sis for the Degree of M. S. MICHIGAN STATE COLLEGE Gale E. We“ 1940 (I) ‘ 1 {1'13 LJILEIL JJE'JQtJLllil1.13.31L1“. k: lixli/GE. 1 ;-v err .41 J..J‘J CAIHUS 4.3THCD .4 L‘lirio'l'o‘ Submitted to the Graduate School of niehigan State College of Agriculture and Applied Science in partial fulfilment of the requirements for the fiegree of gnSLELifigF SLIfiLCE DEPARTgiLT Lfi‘tnflLnldTRY 1940 (‘PFT‘~II‘~"’.'I< Y DEW. ”T 5 45 W 4 23/51 ' J‘- ' “ ‘ ' l). ‘E'JIOZIIJEJG4.ELT '3“ ins writer wishes to express his sincere appre- ciation to hr. Almer leininger, Associate Professor of Chemistry, whose able guidance and helpful sug- gestions made possible the completion of this work. pub L? «It “In ‘1 Cr; ’l‘zL-JL’S or ccmmns Introduction ------------------- - 1 Reagents ———————— - ---------------- 7 Apparatus —-—’ ------------------- 8 Experimental -— ----------------- 10 Summary --- -------------- ----—--- 21 i0 {0 fiicliography ----‘ """""" ‘“"' ILTRODUCTICI There are numerous methods available for the micro determination of halogens in organic compounds. among these is the Pregl (1) combustion method. this involves the burning of the compound in a stream of oxygen. The products of combustion pass over red hot platinum contacts and the halogen is absorbed in sodium carbonate containing sodiun bisulphite which reduces any halate present. The halide ion is then precipitated in a nitric acid solution by silver nitrate. the silver halide is then filtered, dried, and weighed. The dis- advantage of this method is that it is not suitable for volatile substances and those which decompose with great difficult". The sodium peroxide bomb method, as modified by Elek and Hill (2), gives good results. The sample is fused in a micro bomb with a mixture of potassium nitrate, sodium peroxide and sucrose. Zhe fusion is dissolved and any halate formed is reduced by hydrazine sulfate. The halide is then precipitated and weighed as the silver halide. The disadvantage of this method is the dif”iculty in the handling of liquids and volatile substances. in the Lacherl and Krainick (3) method, the organic .C‘ substance is oxidized by a mixture 3; concentrated sulphu- ric acid, silver dichromate and potassium dichromate. the halogen evolved is collected in a mixture of 0.01N sodium hydroxide with acid~free hydrogen peroxide and the excess alkali hack titrated with 0.01L hydrochloric acid. This is a rapid determination but has the disadvantage that it cannot be used on volatile substances. Iodine cannot be determined by this method. The Willard and Thompson (4) method involves the oxid- ation of the sample by fuming sulphuric acid to which is added ammonium persulfate. The halogen is distilled into alkaline arsenite solution and then precipitated as the silver halide. Ehis method is not suitable for the deter— mination of halogens in volatile compounds. The Stepanow method for halogens in organic compounds has been modified by Rauscher (5) so that it can be used for micro analysis. In this method, monoethanolamine or dioxane is used to dissolve the sample. hetallic sodium ‘ .1 is ended and the mixture refluxed. The solution is then acidified with nitric acid, filtered and the halide preci- pitated by silver nitrate. The silver halide is then de- termined gravimetrically. ihe first micro Carius determinations were carried out by Emich and Donau (6). Pregl (7) further developed the method. In this method, the organic compound is decomposed 1 V by meeting with concentrated nitric acid and silver nitrate I11 t a hirh temperature and under pressure. the silver hal- L La ide formed is filtered through a filter tube and then dried (fl and Jeighed. The advantage of this method is that it is adaptable to solids, liquids, and volatile organic halogen compounds. Samples which decompose with difficulty can also be determined by this method. The main disadvantage of the method as now used is the length of time required for the combustion LhiCh is considerably longer than that required for the other meth- ods. Also when volatile compounds are analyzed, the solu- tion contains glass Splinters from the capillary contain- ing the sample which makes it necessary for the precipi- tate to be weighed in a filter tube which is then reweighed after dissolving the silver halide with potassium cyanide solution. In the Carius method, as in most of the others pre- viously mentioned, the silver halide is determihed gravi- metrically. However, a weighed amount of silver nitrate may be used and the excess silver ion determined volume- trically. The Volhard method is probably the most used of these volumetric modifications, at least on a macro scale. The original Volhard (8) method consists of acidify- ing a solution of the halide with nitric acid. To this is added a measured excess of standard silver nitrate to pre- cipitate the halide and a quantity of a ferric ion solu- tion which acts as the indicator. Then without filtering .Lx off the silver halide, the excess silver nitrate is deter- mined by titrating with a standard thiocyarate solution. X , , Ag f- CI 8 ‘—-) £15.30}; 8 ‘ .‘u‘e"" + 661iS’-=->r‘e( CLS )" This method was found to be inaccurate when determin- ing chlorides. The slower the titration was carried out, the lower were the results obtained. Burgleman (9) reported that the indistinct endpoint was due to free nitric acid which destroyed the color by oxidizing the thiocyanate ion. Breschel (10) found that the ferric thiocyanate com- plex was decomposed by the silver chloride according to the equation: Fe(C1\IS):/+ angel _._, 6AgClx-S+ Fe“? He recommended filtering the silver chloride before titrat- ing the excess silver ion. Rosanoff and Hill (11) proved that the silver chlor- ide interfered in the determination causing low results and they also recommended filtering off the silver chlor- ide. Kolthoff (12), however, reports that an error is in- troduced during the filtration due to the carrying down of silver nitrate by the silver chloride. This error is not dependent upon the concentration of silver nitrate present so one can always deduct 0.7% chloride from the amount found. Mothmund and jurgstaller (13) suggested the covering of the silver chloride with tolvene or benzene. This method worked fairly well when the chloride concentration was small but Kolthoff claims there is an error of 0.6} Vander durg and Koppejan (14) investigated the in- fluence of a few colloids on the Volhard titration. They found that in the presence of soluble starch, gelatin, agar-agar, gum arabic, and gum tragacanth, the red colora- tion of the ferric thiocyanate complex appears indistinct and too late. The latest modification of the Volhard method was re- ported by Caldwell and hoyer (15) who used nitro benzene which forms an insoluble layer over the precipitate and thus prevents the silver chloride from interfering. The authors obtained good results by this method when working on the macro scale. There are several other methods which have been adapt- ed to the determination of small amounts of silver. Among these is the titration of silver nitrate with a standard potassium iodide solution using palladious ni- trate for indicator as devised by Schneider (16). The sil- ver nitrate is precipitated by the potassium iodide and the slightest excess of potassium iodide is converted by palladious nitrate to the brown palladious iodide. A method devised by thean and Van Slyke (1?) involves the titration of the excess silver ion with potassium io- dide using starch and sodium nitrite as the indicator. The first excess iodide ion is oxidized to free iodine by the sodium nitrite giving a blue color with the starch. The object of this work was to study the Volhard meth— od on a micro scale and its application to the determina- tion of halogens in organic compounds in an effort to speed up the Carius method. REAGEKTS a 0.1N solution of silver nitrate was prepared from solid, chemically pure silver nitrate which had been heat— ed in a platinum dish between the temperatures of ZSOJ-SOOO for approximately fifteen minutes. This solution was then diluted to the strength desired for the following analysis. an approximately 0.1I ammonium thiocyanate solution was prepared from C. P. ammonium thiocyanate and was then standardized against the standard silver nitrate solution. more dilute solutions are prepared as desired by diluting the standard ammonium thiocyanate solution. A 0.01N sodium chloride solution was prepared from C. P. sodium chloride which had been finely ground in an agate mortar and then carefully heated in a platinum dish until all moisture was removed. A saturated solution of ferric alum indicator was pre- pared to which was added concentrated nitric acid to clari- fy the solution. APPARATUS hicro burettes The micro burettes are of 5 m1 capacity. They were calibrated by determining the amount of water delivered.. They are graduated in .02 ml and can be estimated to the nearest .002 ml. All burette readings were taken with a magnifying glass. micro balance The Kuhlmann micro balance was used in weighing out all micro samples. Carius combustion furnace. The furnace used for the decomposition of all the organic compounds was an electrically-controlled furnace manufactured by the American Instrument Company. With this furnace it is possible to decompose four samples si- multaneously. Weighing tubes A long stem weighing tube is used for the solid samples. It consists of a soft glass tube of 30 mm. length and 2 mm. diameter to which is fused a glass rod about ll cm. long. A ground glass fitting top is used in all cases. For weighing out the liquids a glass bulb is made by drawing out soft glass tubing into a capillary 2-3 mm. in diameter, and then making this into a thin walled bulb. One end of the bulb is sealed and the other end drawn out to a fine capillary. To the sealed end is fused a glass rod 30 mm. long which facilitates the breaking of the bulb in the combustion tube. Combustion Tubes The combustion tubes are made of Pyrex glass about 20 cm. long and 10 mm. in diameter. 10 EXPER I LEI: TAL The purpose of the first part of this work was to study the Volhard method on a micro scale. various amounts of standard 0.05L sodium chloride were transferred by a calibrated pipette to a 10 ml. beaker. To each sample was added a few drops of dilute nitric acid and a measured excess of standard 0.05h sil- ver nitrate. The sample was coagulated by careful heating and then filtered by means of suction through a Pregl fil- ter stick into a 50 ml. Erlenmeyer flask. The precipitat- ed silver chloride was then washed with small portions of a 1% solution of nitric acid. Approximately one ml. of the ferric alum indicator was added to the filtrate. The excess silver nitrate was then titrated with the 0.05h ammonium thiocyanate to the first color change. The fol- lowing results were obtained: TAdLE I - mg.LaCl : ml. AgL03:ml. LH4CLS:mg. KaCl : Error : 0.05000 : .OfigfiBN : FOUND : mg,_w_ 5.845 2.610 .460 5.864 +0.019 5.845 2.565 .450 5.842 -0.005 5.845 2.515 .410 5.851 +0.006 5.845 2.690 .550 5.846 +0.001 5.845 2.690 552 5.840 -0.005 5.845 2.505 .405 5.857 -C.008 5.845 2.551 .440 5.843 -0.002 5.845 2.510 .410 5.855 -o.Oll 11 The volume at the time of the titration was between 15-20 mls. The results obtained in Table I were better than ex— pected because the burettes used can only be estimated to the nearest .002 milliliters and an error of .002 ml. causes an error of .006 mgms. in the calculated results. Because of the difficulty of getting the endpoint within .002 m1, solutions 0.01L were tried. When a small amount of wash solution was used, the results were con- sistently low which was probably due to incomplete wash- ing of the precipitate. When more wash solution was used it became necessary to evaporate the filtrate. Inconsis- tent results were obtained by this method. Because of the filtering step required in the above procedure, attempts were made to determine the excess sil- ver nitrate in the presence of the silver chloride. In the following work, determinations were run in an attempt to adapt the Caldwell and hoyer method to the mi- cro scale. A sample of the standard sodium chloride solution was transferred to a 50 ml. glass stoppered Erlenmeyer flask. The sample was acidified and a measured excess of 0.05N silver nitrate added. Approximately 20 drops of nitro benzene and one ml. of ferric alum indicator was added. This mixture was then shaken for about 60 seconds or until the silver chloride was coagulated and covered by the nitro- benzene. The excess silver nitrate was titrated with the 12 .05h ammonium thiocyanateo __n_.£3..1:~1‘- .11 l -1 "l __ -. mg.LaCl. : ml.AgL05 : ml.LH4CL$ : mg.LaCl : Error : .05000 : .0485QEW_} FOULD_ : mg: 2.925 5.500 2.572 2.924 +0.001 2.925 5.500 2.574 2.917 -0.006 2.558 5.400 2.675 2.558 0.000 2.558 5.290 2.565 2.550 -0.008 2.558 5.400 2.678 2.529 -C.009 1.754 2.940 2.574 1.746 -0.008 1.754 5.200 2.680 1.755 -0.001 1.754 2.900 2.570 1.756 +0.002 The results obtained in Table 11 indicate that the silver chloride has no effect on the endpoint. Upon thorough shaking the silver chloride is removed from sus- pension and at the time of the titration, the liquid above the precipitate is perfectly clear. The nitrobenzene also seems to have the property of inhibiting the darken- ing of the precipitate. The volume at the time of the ti- tration was between 15-20 ml. The main difficulty when using 0.055 solutions is over—running the endpoint. Because of the difficulty of getting the endpoint within .002 m1., more dilute solutions were tried under the same conditions as before, using ni- trobenzene to prevent the silver chloride from reacting. 15 TAJLE III _ mg.LaCl : ml.ng103 : m1.LE4QLS : mg.LaCl : Error .02500 : . 25153 : ECULD : ms. u...) 2.558 4.000 2.586 2.558 0,000 2.558 4.000 2.588 2.555 -0.005 1.754 5.100 1.886 1.757 +0.005 2.925 4.000 1.991 2.920 -0.005 2.925 5.716 1.706 2.922 -0.001 1.169 2.810 2.001 1.165 -0.004 1.169 2.020 1.211 1.175 +0.002 A blank of .004 ml. was deducted from the amount of ammonium thiocyanate used. The volume at the time of titration was between 15-20 ml. Determinations were run on the standard sodium chlor- ide solution using approximately 0.01H solutions of sil- ver nitrate and ammonium thiocyanate. Blank determina- tions were run using the same volume of solutions and as near as possible the same conditions as when running the chloride determinations. Varying amounts 0f nitric acid were added to determine whether or not it had any influence on the endpoint. 14 TL; BLE I V Cone. BLOB m1.: ml. 1H CLS . Volume at : .012fififi : Solution ml. ‘“ 0.1 .012 20 0.2 .011 20 0.5 .012 20 1.0 .011 20 2.0 .012 20 5.0 .014 20 4.0 0.1 20 5.0 0.25 20 Satisfactory blanks cannot be determined when more than 5 mls. of concentrated nitric acid is added. The color of the ferric thiocyanate comples disappears in 5-10 seconds which is probably due to oxidation of the the amount thiocyanate ion by the nitric acid. However, of nitric acid used in all cases is considerably less than 5 m1., so it does not cause inaccurate results. TAdIE V LigdaaCl : m1...g1.05 1:11.14H401-s : 1.1552101; Error _ : 0.01000h: .01216E : EOULD : mg. 1.754 4.000 0.797 1.749 -0.005 1.754 5.980 0.777 1.755 -0.001 1.754 4.295 1.018 1.757 +0.005 1.754 4.540 1.217 1.757 +0.005 1.754 4.500 1.197 1.748 -0.006 1.169 4.000 1.582 1.171 +0.002 1.169 4.000 1.587 1.168 -0.001 15 For all of the above determinations a blank cor- rection of 0.012 ml. was deducted from the amount of ammonium thiocyanate used. The volume at the time of titration was between 15-20 m1. 0 A number of determinations were run on the standard sodium chloride solution, using varying amounts of ni- trobenzene to see if it had any effect on the titrations. TAfiLE VI mg, .1:aCl :ml . ngLO 3: ml.hH CLS :hitroben- :m5.LaCl:Error :.01000N .01216N zene drops: F0010 : ;::__ 1.754 4.500 1.199 10 1.750 -0.005 1.754 4.505 1.202 10 1.749 -0.004 1.754 4.540 1.225 20 1.754 0.000 1.754 4.540 1.227 20 1.750 -0.004 1.169 4.000 1.580 50 1.175 +0.004 1.169 4.000 1.580 50 1.168 -0.001 It is evident, as shown by these results, that vary- ing amounts of nitrobenzene has no influence on the re- sults. However, it was found that when the volume of the solution was greater than 55 m1., it was impossible to completely coagulate the silver chloride. when the solution was titrated the silver chloride would inter- fere,making it impossible to get the proper endpoint. the Volhard method was found to be adaptable to the micro scale, when using solutions 0.025E and 0.011. The next step was to determine whether or not it could be used ~in connection with the micro Carius method. 16 THE CARIUS DETERLILATIQE A 4-8 mgm. sample of the solid organic compound was _weighed out in the weighing tube and transferred to the combustion tube containing 19-20 mgms. of accurately weighed silver nitrate. Approximately 0.25 ml. of con- centrated nitric acid was added and the tube immediately sealed by means of an oxygen flame. The liquid compounds were drawn into the previously weighed micro bulb by care- fully warming the bulb and dipping the capillary into the liquid. 0n cooling, the liquid will be drawn into the bulb. The capillary was sealed and the tube reweighed. It was then placed in the combustion tube containing sil- ver nitrate and concentrated nitric acid. The tube was sealed as before. the weighing bulb was broken by hit- ting the combustion tube against the hand. The combus- tion tubes and contents were placed in the combustion furnace and heated at such a rate that by the end of three hours the temperature had risen to 250°. This temperature was maintained for another three-hour period. The sam- ples were then allowed to cool. The combustion tube was pulled part way out of the furnace and the tip heated with a small flame to drive out the condensed liquid. It was then heated strongly and the pressure inside the tube caused it to break at the weakened point. She contents of the tube were flushed out with small portions of hot water into a 50 m1. glass-stoppered Erlen- meyer flask. Approximately 10 drops of nitrobenzene and 17 one ml. of ferric alum indicator were added to the flask and the mixture thoroughly shaken for 60 seconds. The excess silver nitrate was then titrated with 0.0251 ammonium thiocyanate. For halogen compounds containing bromine and iodine, no nitrobenzene was added because the solubility of sil- ver bromide and silver iodide is less than that of si1- ver thiocyanate. analyses were run on a number of organic halogen compounds with the following results: TABLE VII 351111511111 .bample- “3151033 1111. 111401.25 111801". Calculated 1 “EL- 1 mg. 1 .02515 -1 :11 1 )0 hall__e____ Chloro-benzoig 5.549 11.675 1.372 22.64 22 53 “Cl 5-067 17.135 2.714 22.69 4.816 10.715 1.281 D'.65 7.060 17.445 2,230 "”0.60 Chloro 4.725 12.416 1.796 20.91 20.86 acetanilide 4.002 9.812 1.355 00.91 8.411 15.294 1.144 20.85 6.028 10.557 0.995 21.91 5.457 9 941 1.056 20.84 5.755 14.269 1-991 20.82 8.875 15.456 1.555 84 p. p. dichloro- 7.000 25.177 3-305 25.55 26.89 az “o y benzene 4.518 15.246 1.796 26.84 5.515 17.058 2.559 26.65 18 iunLliLl/lll QumILE :Sam mple: n 3103 :ml.1H 053:Theor:0alculated ; ; Inr _ .025 5 --_.]§._--.E.-.i€{ halide fichloro Anthra- 5.140 12.557 2.051 14.61 14.71 quinone 6 956 16.067 2.612 14.67 5.111 8.25 5 1.064 14.66 p-Bromo 4.674 12.286 1.994 57.55 57.60 Acetanilide 4.704 12.096 1.955 17.41 6.145 16.094 2.610 57.65 p-bromo 4.555 8.676 1.250 54.28 85 ‘ Diphenyl 7.951 10.977 1.211 64.54 2 Cl-5-Hydroxy- 4.195 11.001 1.400 24.86 24.95 ; toluene ,MM 8.852 21.645 2.565 25.15 5.602 11.566 1.115 .5.15 m-chloro phenol 4.577 11.166 1.180 27.58 27.86 4.110 10.794 1.250 27.62 Iodobenzoic 6.520 11.171 1.595 51.17 51.59 Acid 8.655 9.848 0.915 51.54 7.568 11.225 1.457 51.27 The following determinations were run using the same procedure as before with the exception that 0.01L ammonium thiocyanate was used to titrate the excess sil- ver nitrate. SadLE IX SAhPLE Samplezfig505 :m1.LH4CIS: lheor: Calculated m2. :225 : 012615 : 34 : 2anlgxifig Chloro 6.684 10.490 1.770 20.91 20.91 ““etanlllde 6.648 9.615 1.580 20.89 4.266 11.195 .22 21.01 19 The main disadvantage of the Carius method, as used in the previous analyses, is the time required for the decomposition of the sample. ihe six hours required is considerably greater than the time necessary for the other methods of determining organic halogens, thus decreasing the value of the Carius method for routine analysis. because of this, attempts were made to speed up the \I'. determination by increasing the temperature and decreas- ing the time of combustion. 0n the following trials the temperature was raised to 5000 in 60 minutes and then h held at that temperature for another 60 minutes. The conditions for the titration are the same as before “WM using .011 ammonium thiocyanate. T141313 X ‘ _BahrLE :SqmplezAshozsmlohH 0N5 The°r‘caICUlated : ImgA timi. :.0126 N : ;g :p% “halide Chloro 4.500 10.675 2.955 20.91 21.00 acetanilide . 6.417 13.594 5.245 20.94 Chloro benzoic 5.166 9.168 1.655 22.64 22.71 Acid pp dichloroé 5.591 9.975 1.450 26.55 26.75 azoxy benzene 5.880 11.602 1.915 26.60 p bromodiphenyl 5.555 9.599 5.150 54.28 34.35 4.856 15.545 4.570 54.26 bromoacetanilide 4.080 10.594 5.450 57.55 57.88 5.622 15.458 4.920 57.84 2 chloro 5-hydro-5.009 14.092 5.775 24.86 24.95 xy-toluene 6.279 11.749 1.975 24.98 20 These results show that the time required for the .1 _,. ' ‘ ‘ a < 1 decomp031t1on can ce decreased. In all the aoove analy- ses the samples were completely decomposed by a two-hour combustion period thus considerably decreasing the time required for the Carius determination. ”I. .r v‘:-) 1 -. F0 JUhhARY A modified micro Carius method for determining halogens in organic compounds using the Volharfi method to determine the excess silver nitrate was found to give good results. and 0.011 solu- ()1 The Volhard method, using 0.02 tions, was modified by using a small amount of nitro- benzene which makes it possible to carry out the ti- tration in the presence of the silver chloride. It is not necessary to use nitrobenzene when analysing compounds containing bromine and iodine. it was found that the organic halogen compounds were completely decomposed by a two-hour combustion period as compared to the six hours previously re- quired, thus considerably decreasing the time required for the Carius determination. ‘1 L' 1'. \H- -'W ‘Ln'.’o_‘n.l- 1‘ J‘ ()1 ,p. A A A A A A A A \7 OS ‘1 (14) (15) (16) (17) BIBLIOGRAPHY E. firegl; "Quantitative Crganic Licroanalysis," r. blakeston's Son and Co. Inc., lhiladel- phia, 1957, Third English Edition, p. 94 Elek and Hill; J. Am. Chem. Soc. 55,2550 (1955) F. Pregl; loc. cit.,u. 108 Hillard and Thompson; 5. Am. Chem. Soc. 52,1895 (1950) Rauscher; Ind. and Eng. Chem. Anal. Ed. 9,296 (1957) Emich and Donsu; honatsh Chem. 50,745 (1909) F. Pregl; loc. cit., p. 104 J. Volhard; J. Frakt Chem., 117,217 (1874) V a“l‘~; '. " '., G. Burgleman; 5 Anal. Chem. 16,1(1877) Dreschel; J. Erakt. Chem. (new series) 15,191 (1660) Rosanoff and hill; J. Am. Chem. Soc. 29,269 (1907) I. L. Kolthoff; Z. anal. Chem. 56,568 (1917) Rothmund and Jurgstaller; Z. anorg. Chem. 65,550(19l9) Van der jurg and Koppejan; Chem. Jeekbl 21, 167(1924) Caldwell and hoyer; 1nd. and Eng. Chem. Anal. Ed. 7,56 (1955) J. Schneider; J. am. Chem. Soc. 40,585 (1918) hacLean and Van Slyke; J. Am. Chem. Soc.57,ll28(l915) 1.. .. ~‘c‘av- ., Q I Z- ‘W- ‘ \ K)- K»: P ; Aug-aw? " .. , u! I t . 1 ‘l v . u - n 1 ‘ . I :I‘A’~ A '7‘- V 'b” I y ‘ . u . v , a. u . :. . .‘ I I ~uk’ ' n . . " I v '_ . ' I -\ I“ ‘0 . . , n w . . h; . c '. I v ' \ 1.51573“, g . . n v I, 'c ' . l ‘- v. x I. u. ' V. x . |-' ~. I. Q IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 31293 02446 6728