EXCHANGE A Comparative Study of the Slow Combustion of Methane and Methods for the Determination of its Combustion Products DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School of the Ohio State University BY FRANK CARL VILBRANDT, B.A., M. A. The Ohio State University 192O A Comparative Study of the Slow Combustion of Methane and Methods for the Determination of its Combustion Products DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School of the Ohio State University BY FRANK CARL VILBRANDT, B.A., M.A. \\ The Ohio State University 1920 The slov; combustion methods consist in passing the combustible gas slowly into an oxidation chamber filled with oxygen or air, the combustion initiated by an elec- trically heated coil. The combustion pipet of Clemens Winkler {Ztsch. anal* Ghem. 28, 288 a 1889) was modified by the following investigators": Dennis and Hopkins (Dennis 1 Gas Analysis 1 1913, 148), Bretschger (Treadwell- Hall. 'Quantitative Analysis', 3rd Edit. J. Wiley & Sons, p. 790), Burrell and Siebert (Bur. Mines Bull. 42, 17, 1913), Voldere (Chem. Ztg. 31, 1296, 1907), Scott (Ind. Eng. Chem. ]., 118, 1909), Burrell (U.S. Pat. 1, 176, 199, L1917J), Anderson (Ind. Eng. Chem. 11, 292, 1919) and Demorest (Private communication). The main disadvantages of this method are the incompleteness of combustion, for- mation of oxides of nitrogen, breakage of apparatus in the hands of non-experts and errors due to oxidation of the confining material. The capillary method of combustion, introduced by Coquillon (Compt. rend* 85, 1106, 1877) consisted in pass- ing a previously mixed quantity of the gas and oxygen or air through a heated platinum tube, oxidation produced only in the presence of the platinum. 'This apparatus was later modified by Drehschmidt (Ber. 21, 3242, 1888), Wohl (Ber. 57, 433, 1904), and Roe hart (Ztsch. anorg. Chem. 38, 78, 1908); The same method modified by using quartzware instead of platinum for the capillary, but using various forms of platinum in the tube for catalysis were intro- duced by Sabatier ( *Die Katalyse 1 1st edit. Leipsig, p. 23), Wilson and Mason ^Proc. Chem. Soc. [London 3 21, 1296, 1905) and Mathers and Lee (Chem. Eng. 17.* 159, 1913). The disadvantages in this method such as leakage of gas through hot platinum and the high cost of platinum have been fairly well taken care of in the quartz tubes. Each modification introduced was designed to over- come some inherent error of manipulation or apparatus. ' The explosion method is essentially an expert^s method,, not capable of giving good results in other hands; the capillary method requires something better than the easily deterioratable quartz, and the slov/ combustion method re- quires the elimination of breakage due to local heating of the capillary and oxidation chamber joint, proper manipu- lation or apparatus design to insure completeness of com- bustion and the reduction of error by the use of larger quantities of gas for the combustion. The investigation carried out primarily studied the slow combustion method with the view of eliminating the above undesirable features and to study the influence of the oxidation effects of nitrogen on the analyses when air was introduced, and to study the effects of the products of incomplete combustion of the methane. EXPERIMENTAL A modified Burrell common train Orsat laboratory set- up, with a Kempel temperature and pressure compensator was used. The absorption reagents used were potassium hydrox- ide, alkaline pyrogallate s and acid cuprous chloride, pre- pared according to Dennis ( T Gas Analysis 1 , 1913, pp. 160, 225, 232, 233). The methane used was prepared by the Glad stone -Tribe method (J. Chem. Soc., [.London! 25, 684, 1873; 45, 154, 1883) from methyl iodide and Zn-Cu couple, but pure methane could not be mad.e by this method. Other methods yielded no better results. Methane analysing 84. 55/ methane 12. 10/ nitrogen and 3.35?o oxygen by the Dennis-Winkle r method was used for the investigation. STANDARDIZATION OF THE TEMPERATURE OF THE HEATED COIL. The references in the literature to red heat, cherry red, bright red, white heat and dull red for the temperature of the electrically heated coil was too inde- finite. Following the suggestion of Burrell and Siebert (idid. p. 19) who measured the current flowing through a platinum coil, but gave no coil dimensions for reproduc- tion of their temperatures, in this investigation, the temperatures were recorded as the amperes flowing through a #24 B, & S. gage, platinum wire, with a five millimeter diameter coil of four turns two millimeters apart. MODIFIED ELECTRODES. Difficulties due to the use of electrodes proposed by Bretschger (loc. cit.), Dennis and Hopkins (loc. cit.) and Porter and Ovitz (Bur. Mines Bull. I f 24, 1910) by breakage, leaks and oxidation of the mer- cury in the seal, led to the design of a new electrode consisting of sealed in platinum coil making contact with the current supply by means of mercury, held in the elec- trode tubes by rubber tube joint, with a large copper rod to which the supply wires were connected. (Seet^) Figure 1.) SLOW COMBUSTION PIPETS AND METHODS OF ANALYSIS. In order to make a thorough survey of the difficul- ties inherent in each of the more modern pipe ts in com- parison with a new proposed pipet, the following pipets and methods were studied: Dennis (ibid. p. 149) and Burrell and Siebert (ibid. p. 79). These were modified to make the Burrell and Siebert either water cooled or air cooled by providing an air jet to play against the pipet and also a continuously flowing film of water over the same. In addition the two types of pipets were supplied with the new central burner. CENTRAL BURNER PIPET. The central burner pipet (Figure 1.) is provided with two capillary stop-cocks, B and C. Manipulation of the apparatus is carried out as follows: With C closed, air or oxygen is drawn into the pipet by lowering the mercury leveling bulb attached to the pipet at F. With ten cubic centimeters of the air or oxygen the mercury in the burner tube K is forced out by closing B and opening C. Connections are made with the supply of gas to be analysed; with the temperature of the coil attained by passing a measured current through the coil, the gas is slowly drawn into the combustion pipet by opening the stop-cock on the lowered level bulb; the gas on entering burns with a small flame at the mouth of the quartz burner tube K. When the gas has all been passed into the chamber, by manipulation of the cocks and level bulbs, about twenty cubic centimeters of the gas is withdrawn and passed back and forth into the apparatus several times to insure complete combustion of all the gas which would normally be retained in the capillary connec- tions. .This type was also supplied with water and air cooling. These pipets were constructed of transparent quartz to enable the study of cooling with water and air, not possible with glass apparatus, and to eliminate the frequent interruptions necessary to replace the easily broken glassware. The capillary quartz tubes were of special construc- tion, with enlarged ends to enable fitting on tubing; several different sizes were studied. The effect of design of apparatus, reduction of ap- paratus temperature with water and air, effect of cutting off the current after the initiation of the combustion, the effect of varying the temperature of the platinum Figure I. Central Upright Burner Pipette wire, the effect of varying the oxygen to gas ratio, the effect of rate of oxidation and presence of nitrogen were studied with the above equipment. CALCULATIONS. The results obtained in the above study, were calculated by the use of the formulae of Voldere and De Smet (Ztsch. anal. Chem. 49, 661, 1910), Anderson (ibid. p. 302) and by the author. Using the general formula for the complete combustion of the hydro- carbons of the methane series: 2 = nG0 2 * ( n+1 ) H 2 2 Since the total contraction is equal to the sum of the volume of the hydrocarbon and of the oxygen required minus the carbon dioxide formed, or T.C. equals V plus Qg minus COg then Og required equals 3n plus 1. 2 COg equals nV T.C. " 5 plus n . Since the gases deviate from the gas laws according to Wohl (Ber. 37_, 429, 1904), Raleigh, Leduc, Baume, Perrot (J. Chem. Phys. 7, 367, 1909) and Dennis (ibid. p. 139) it was necessary to make corrections for these deviations. The equations for the reaction of methane and hydrogen with oxygen then becomes; 0.999 CH 4 plus 2.000 Og equals 0.954 COg plus 2 HgO 2.004 Hg " 1.000 Og " 2 HgO or, COg equals 0.995 CH 4 T.C. ts 1.499 Hg plus 2.007 CH 4 In order to eliminate repeated development of correct equations on the basis of carbon dioxide remaining in the gas after combustion a correction factor was developed for the different hydrocarbon values following the plan of Anderson (ibid. p. 302) 6 SUMMARY OF COMBUSTION DATA. The data obtained in this investigation is compiled, in Table 1, the results being averages of from three to six determinations, includ- ing all data such as type of apparatus, modification, color of heating coil, current flowing through wire, rate of gas flow, ratio of oxygen to gas sample, analyses and estimated average completeness of combustion. DISCUSSION OF RESULTS. FORMATION OF OXIDES OF MERCURY. Many determinations not recorded in Table 1 were eliminated because of forma- tion of oxides of mercury due to the high local heating of parts of the apparatus in contact with the mercury con- taining fluid. Mercury oxidation dees not affect methane percentages but, by increasing the total contraction, in- creases the hydrogen percentages. Two forms of the ox- ides of mercury were found in the pipets during some of the worthless combustions;- red and brownish yellow. The red variety occurred at high coil temperatures and the yellow at low heats. INFLUENCE OF WIRE COIL TEMPERATURE. The results show incompleteness of combustion with the coil heated with below 6*8 amperes; additional experimentation showed that no ignition of the gases resulted when less than 3.5 ampere were used to heat the coil, consisting of four turns of #24 gage platinum wire, 2 millimeters between coils and 5 millimeter coil diameter. After initial ig- nition of the gas the temperature of the coil had no effect. INFLUENCE OF RATE OF COMBUSTION. After visible com- bustion has once been started the rate of gas flow up to 25 cubic centimeters per minute has no effect on the analyses, except that precautions must be taken where mixtures approximating explosive proportions are used that destructive explosions do not occur. INFLUENCE OF RATIO OF OXYGEN TO COlViBUSTIBLE GAS. With ratios approximating 2.0 oxygen to 1.0 gas, or no excess oxygen, results are erratic. With 25% excess oxygen, or with ratios of 2.5 to 1.0 uniform and agree- ing results were obtainable . Nothing is gained by in- creasing the oxygen ratio above this value. 1 1 \^f 7 . Cx> ro 1 1 1 1 I 1 1 1 1 1 i i i Series I'ku a' p o cr P DJ o er" P exi ro H- en rf^ en cn en en if* rn oo ex) exi exi Nos. of runs. _ _ _ _ O = = = o o Type of Ap- s 9 to c+- paratus 3 H- r^ H^ ~ ~ = = = *r w -s p t-j | r? f" 1 CO ro -J 1 = = = = s = = = o o o = = o jjj y o 5 H"- w n- CD H- Modifica- _ _ _ _ _ P O = = = c** C P 3 tion H- O "" H- J3 3 CM 03 CT3 W g W = = = = = Color of H- H- s ~ = = 5 5 M Wi re o tr a co -o ? Os Anipers flow- = = r = CO s s: ro cn ing through yq coil* ^ cn o i o ro ro ex) ro ro ex) ro ro i M tf, ^ i w Ratio of c tr< en O O en O O en O en a> O ro Oxygen/Methane ^3 KJ o CO -o co co co co co co en en cn en '~^ f^ H^ en h"> h^ i 1 i- 1 en o co exi cxj ro 4i**cn <- {> t-J O tf* O O co HJ o -~o * rv> ro o/ P (6 ro j- 1 co co T si cn r\> O M co t^ CO OD CX) -0 CO CO -0 n> * ex; cn CO CO 1 H- 1 en > OT Cb tP- ro O O Methane $ CO ro o co ex) K* O en co o O O O rfs- i-j O exi ro Hydrogen CX) -o h^> ro -a ^ M i Gi H- 1 K- 1 o/ CO o O cn ex) en cxj t CD CO CO s* cn ro f * H- ' i ' > en en co CO ^ CD O cn o> ~3 Residual Gas OD co en cc co O -o co ro it- *4 " "^ y i.0 CO CO H- 1 !- cn cn t^- co O ex) ro . M Estimated Av, CO CO CO CO '^3 CO CO CO O co o co cn o> O ai O) a> -o a o . ->3 co ro cn -J. * CO Combustion I_J en ro ro o -j ->j co en CO ^ CO % o Hj W -< 1 Remarks a CD on ^cra 7 i ? Series No. H-* CO p td H- ? ro ro 0* P i cn o S _^ CD = 3 -$ f*T P M H- 1 *t tO c i "I M cn o co P < cTP ro H"-< O CD H- 1 1 cr P ~" H- 1 ~ O O = o CD CO CO a CO rfs- - I 1 cn 1 P H- CO co i o (> cn .-a 6 Dennis- Continuous Dull 818 12 Winkler Heating Quartz Vari- -b 5 " " " able No s. of runs. Type of Ap- paratus Modifica- tion Color of oo Wire Amperes flow- ing through coil- gj ^ o > Rate of gas flow i CD- cc/min, ^ t- 1 H3 W ro CXI ro ro ro ro cxi ro CXI ro ro Rate of cn cn O o cn o cn O cn C Oxygen/Methane S M > O co CO CO -^ Co co CD Co CO CO co c rto 2 o t* *f* > CO if" ri^ rt^ 03 f> H^ n cn * 3- i_> ro ,)5, M cn cn cn cn cn cn CO c f W " co en H cn ro cn cn ro CO ro * /" co a- CO CO 03 Co cxi 23 co co rf*- CO CO CXI CD ro CO CXI * cn cn CO * O ro O o O O O ro a CO Hydrogen !_, ro CO ro ro cn oa -a cn CO ^ CD cn M CO CO cxi * * cn % I- 1 cn cn cn ro O i- 1 cn cn cn cn M Oi \-> cn S Residual Gas Co -3 cn CD ^ OJ if* tf* cn i> cn a-/ ro tf* cn CO cn cn cn to cn cn / co CD CD CD co co CO ro CO CD co CD O O co CD CO CO CD CO CD * Estimated Av. Completeness of -0 cn cn cn CD Ij b GO rf^ CXJ CD Combustion .. HJ CO H-J CXI cn CD ro cn >^ CX) CO & 3 K 3 tc / P OT P OT c+- O H- Remarks O I-* O **J 3 3 1 7 C-, 1 P CXI M ro i CX) ~r c c S> g c V. V H- 1 1 P CXI P d- N > d- ; o> > -5 1 1 L *f ro i P H- d- CD CO CO 1 H- 1 1 to H CD 1-i I *i t N >- K C! f- H O 5 a O a i P d I 5 I -i _< I-J 1 * C = C ll r 1- P! H- C *-< o O~ P o t C t- 1 CD SL i i P ~ N k ) r 1 H i, j fr t 1 L o ro p s & i P H c o 1 CXI O *> d- f P 3 ^ ? 2 w H- d- -o 9 Series Wo. IIos. of runs. Type of Ap- paratus Modifica- i Color of g Wire Amperes flow- ing through ^ coil o ,_, ^ ^ !-> ^_, _ ro ro ro Rate of gas fbw GO = co CO ro o cn cc/min. 3 f cx) cxi ro CX) ro ro ro ro 03 ro ro ro * ^ Ratio of Pd O O ro o cn cn cn cn O cn ro ^ Oxygen/Methane I M CO CO co CO CO co CO CO co CO CO CO ^ ^v cn lf^ cn > ir h^> = tf* lf rt^ H^ Cn"2n 2 tr 1 O co cn ro cn *> cn cn CD lp cn cn cn of ^ ro IP 1 O'J H CO ro t-< o % (r) CO ~ CO CO CO CX) co ex) co CX) CO Cxi co CXI CO CX) CD CO CO CO Methane W co JL Oi to CO CO co CO O Q ro "^ -a cn * * H- 1 o O o M a o O o O O o O Hydrogen to l_i cn cn cn -0 CD CD rf^ ^ -o -0 d co CX) -o * CO h^ *^ CO CXJ co -0 to / M M \-> \-> K- 1 w ,_, M M M t-" H Residual Gas o ro cn rf* cn cn cn cn cn cn cn cn Co O _j CD cn CXI Jk r^ co en vp> to 6 Oi Co to H- 1 8 e o ft CO to to CO to to to CO to CO to O CO to CO CO O CO CO to CO Estimated Av. Completeness of cn ro Oi cn O s cn CO CD CD ro s \-> \-> CXI ro ro Nos. of runs- _ _ TJ """ H- M Type of Ap- "- T^r ** P paratus - H- O o ^ HJ H' H- ^^* O M M 2 z 3 s H- 1 H- O H- 3 CD 01 <+ H- cn w H H- Modifica- : - CD' ^ CD M tion fT^ * * H- , v __ x 0- O ^ Cd 1 = = H- H- Color of Wire co pt- 8 cn * H- 1 - CD o ro ^ H- CD Rate of gas flows r* cc/min co w i-3 M ro ro CXI ^ exi ro ro CXJ ro ro Ratio of o M en ro en O O en ro o en ro Oxygen/Methane ^ ^ 00 oo CD CD CO CO CD 00 CO -a A TT O t~J O * rf* rf* i 1 en ro Cxi CO c*- ^ ^_, en en exi exj H- 1 O CX! CO ? to ro CXJ tO en to to O CO cxj /o CO ^ 00 CO O OP O3 Cxi tT- CD CO CO exi exj i^ CD CX) CD I- 1 ro Methane en o -o O CO O O CX) o en tf en o ro ui to ro H* oo ro f-< o O O M H H- Hydrogen ro H -si O3 ff>- ^ M cn CXJ o/ ro -o ri^ -