I. Coulometric and Amperometric Methods of Analysis. II. Spectrophotometric Investigation of the Copper (II) Monobromo Complex. III. Determination of Carbon by Wet Combustion Citation Farrington, Paul Stephen (1950) I. Coulometric and Amperometric Methods of Analysis. II. Spectrophotometric Investigation of the Copper (II) Monobromo Complex. III. Determination of Carbon by Wet Combustion. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/hka4-a051. https://resolver.caltech.edu/CaltechTHESIS:08052025-224132798 Abstract The conditions under which electrolytically generated chlorine can be used for secondary coulometric titrations With an amperometric end point have been investigated. A procedure is described by which triposittve arsenic has been titrated in quantities of from 30 to 800 micrograms with an average error without regard to sign of less than 0.5 micrograms. the method proposed by Wooster (30) for calculating corrections for the coulometric titration of iodide in hydrochloric acid has been studied. Data are presented to show that improved methods of calculation give more accurate results. For the coulometric titration of iodide by means of electrolytically generated bromine perchloric acid may be substituted for hydrochloric acid. The perchloric acid requires no purification and the titrations are just as accurate as those made With hydrochloric acid. A metallic bismuth micro-reductor has proven satisfactory for quantitative reduct1on of ferric iron. Samples containing 234 micrograms of iron have been reduced and titrated with electrolytically generated ohloll"1ne to an accuracy of better than 1%. Iodine monochloride has also been used for the oxidation of ferrous iron. Recommendations are made for further investigation. A combination of coulometric and amperometric techniques has been used to determine a dissociation constant for iodine monobromide and a constant for the oxidation-reduction equilibrium involving cupric copper and bromide ions. For the iodine-iodine monobromide half-cell, data from this investigation yield a potential value of -0.92 volt as compared with a value of -0.88 volt obtained from the literature. On the other hand, a potential value for the cupric-cuprous bromide half-cell calculated from this experimental data differed by 0.06 volt from data found in the literature. A study of the indicator currents caused by various oxidation states of vanadium established the fact that tripositive and quadrapositive vanadium give rise to an indicator current. Also, it was observed that a reducing agent capable of replacing hydrogen ion would probably be partially oxidized by the indicator circuit; a special method would be required for the coulometric titration of such a substance, The equilibrium constant for the formation of the monobromo complex of copper (II) was determined by means of a spectrophotometric method. In solutions of unit ionic strength the constant was found to be 2.1 ± 0.25. For the determination of carbon in organic solids and relatively nonvolatile liquids, a scheme has been devised to use the Van Slyke-Folch combustion solution. An apparatus is described for situations where a Van Slyke manometric apparatus cannot be provided. The results are accurate to ± 0.05 mg of carbon. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry and Chemical Engineering) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Minor Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Swift, Ernest H. Thesis Committee: Unknown, Unknown Defense Date: 1 January 1950 Record Number: CaltechTHESIS:08052025-224132798 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:08052025-224132798 DOI: 10.7907/hka4-a051 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17596 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 08 Aug 2025 18:25 Last Modified: 08 Aug 2025 18:27 Thesis Files PDF - Final Version See Usage Policy. 33MB Repository Staff Only: item control page

I ., COULOME'.rRIC AWD AMPEROMETRIC METHODS OF ANALYSIS II , SPECTROPHOTOMErRIC INVESTIGATION OF THE COPPER (II) MONOBROMO COM!'LE.l III e DETERMI NAT ION 01,, CARBOU BY WET COMBUSTION Thesis by Paul St ephen Far~ington In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy Califo~nia Institute of Technology Pasadena~ Califoxnia 1950 ACKll OWLEDGEME.MT The advioe and guidance of Professor Ernest H. Swift are most sincerely appreciated; his assistance has been invaluable in the pursuit of these investigations . Professors Carl Niel1".snn and. Norman Davia.son hava contributed man;; helpful suggestions. Dr. c~ w. Gould constructed the special glass apparatus required in Pa~t III of this thesis . The $ncourage- ment of my wife, Barbara , and her assistance in the prepare.ti.on of this manuscript ara gratefully acknowledged . ABSTRAO'f The conditions under Which eleetl'olytically generated chlorine ean be used for secondary QOUlometrie titrations With an ampet"ometric end point have been investigated. A p:rocedue ts desel'ibed by which triposittve arsenic has been titl:ated in quantities of from 30 to 800 microgt"ams with an average e:r,:,or without regard to sign of less than 0.5 micrograms. the method propos~d by Wooste:r (30) for ealoulating cor• l'&Qtions for tha ooulometrie titration of iodide in hydrocb.loric aeid has been studied. De.ta are p:resented to show that imp:roved methods of cateulation give m0~e aceui-ate results. For the eoulomet~1o titration of iodide by means of eleetl'olytically genel'ated ba-omine perehloi"io acid may be substituted for bydi-ochloric seid • The pe:rchloi-ie aeid l'eq_u it'es no purification and the titl'ations al'e just as aeourate as tl1ose made With hyd:rochlol'lO aoid. A metallic bismuth miol'e ..?'eduetoi' bas proven satisfactory :for quantitative :reduot1on of fei-ttie 1ron. Samples eontaining 234 mic:n.•ogl!'ams of iron have been reduoed and titrated with eleetrolytically generated ohloll"1ne to an aeau:raey of betteithan l~. Iodine monoehlol'ide has e.lt=JO been used :for the oxi- dation of ferrous iron. Recommendations a!'e made foii ftu.'ther investigation. A combillS tion o.f ooulomet~1o and am:pe:rometr:1.o teohniques has been used to det ermine a dissociation oonstant for iodine monobromi d.e and a constant for tha oxidation-reduction equili brium involving cupric oopper and bromide ions. .li'o:r the iodine- iodine monobromide half- cell, data fr om this investigation yiel d a potent ial value of -0.92 volt as compared with a value of - 0.88 volt obtained from the literature. On tho other hand, a potent ial value foi~ the oupr io-ouprous bromide half- call calculated f:rom this experimental data d.if:fered by 0.06 volt from data found in the literature . A stu d.y of the i ndicat o-r currents cause d by various oxidation states of vanadium established the fact that tripositive and qua.d:raposi ti ve vanadium give ::r.is e to an indicator current. Also, it was observed that a reducing agent capable of repl~cing hydrogen ion would pr ob ably be partially oxidized by the indi cator ci:rcui t; a special method would be required for ·the coulo• metric titration of such a substanoe, The e quili.brium constant for the formation of the monobromo complex of copper (II) was determined by means of a spectrophotometric method. In solutions of unit ionic strength the constant was found to be 2.1 ~ 0.26. For the determinat ion of carbon in organic solids and relatively nonvolatile li q_uids, a scheme has been devised to use the Van Slyk:e-Folch combustion s olution . An apparatus is de"' · scribed f or s ituations where a Van Slyke manometric apparatus cannot be provided . carbon. The results a.re ac curate to ± 0.05 mg of TABLE OF CONTENTS Page I• COULOMPJTRIC AND ,.,')J/IPEROMJID:RIC MJID~ HODS OF ANALYSIS Introd.uction o 1. 2. 0 4 0 0 • C O Q G C O e O 1 0 Tho Use of l!lleotrolytically Gene:ra.tsd Ohlorina in Coulometric Titrations. • " ,0 Coulometric Tit~ations The Coulometric Titration· of Iodide in Hydiioohloric Acid ., • ,, " • ,, ., The Ooulometl'io Titrat ion of Iodide in Pe~ohlorie Acid, • ••• o • • 3. 4 16 0 23 c. The Coulomet~io Titration of Iron • • • 26 • Equilibr ium Studies A. Equilibria in Iodine li&onobl."omido Solutions '° •• o e • o • • • 0 • 0 e 37 o ., • 0 51 Study of the Suocessive Oxidation States of Vanadium ~ 0 e Oxida:tion- Redu.otion Equilibria in Cupric Bromi de Solutions. • • 4. \> 0 ' C • G O ii 61 0 0 0 6 0 • • 0 67 SPECTROPHOTOMETRIC INVESTIGATION OF THE COPPER (II) MONOBROMO COMPLEX • o o b 0 O O 0 69 References • o • 0 0 ., • " • 0 80 • 0 " 105 1,1 References II. III,, o ~ D 0 0 .... o c G 0 e • • • o e e o " DNf~JRMI NATION OF CARBON BY WEr COMBUSTI ON References . " " 0 O O 6 o G 78 I .. COULOMhiir'RI C AU D AMPEROMETRIC METHODS OF ANALYSI S Introduction A coul omet:ric titration may be defined as an elect:rolytic process in whi ch a. reaction is effected quantita tively with known electrical efficiency end the number of equivalents is determined by meas uring the amount of elect:ricity consumed., As discussed by Szebelledy and Somogyi (28) there are two types of coulomet:ric J)Q:oceysses: the primary or direct process in which the desired reaction takes place at a suitable electrode , and the secondary or indirect process in which an intermediate half- cell react ion is caused to talce place at the electrode and the electrolytic produat then causes tha desired reaction •. Lingane (14) has desc~ibed examples of the primary process in Which certain metal ions were reduced at a mercury cathode . , Inasmuch as the cathode voltage was cont rolle d, selective reductions coul d be aecomplished . . However , the curi·ent decreased exponentially dut'ing the titration , and the titration time was therefore pr olonged. Examples of the secondary pr ocess have been described by Szabelledy and Somogyi (29). In these procedures the inter- mediate half-cell :reaction was the electrolytic oxidation of bromide to bromine; the bromine then oxidized such substances as thiocyanate , hydrazine a.nd hydroxylamine. Since these re- ducing agents al'e not susceptible to stoichiometric anodic oxidation , it is apparent that the secondary process can be used to extend the range of application of coulometric methods. Furthermore 0 the pl'esance of a high ooncentration of the intermediate substance (bromide, for example) provides two additional advantages: 1, A constant electrolysis oul1rent can be maintained with assurance that the elect~ode efficiency is known . This permits the use of the relatively simple constant current - time method for determining the amount of electricity used in a titration. In comparison with the cathode-voltage controlled process the titration is more rapid. 2. Aoourate micro titretions are possible because the current and current eff'icienoy rematn const ant re gardless of the ooncentrat:J.on of the substance to be determined. Recently , electrolytically generated bromine has been used for the titrat ion of thiodi glyool ( 23)., trip osi ti ve arsen:tc ( 20) • tri positive antimony (1) ; and io dide (34). All of this work differed from that of Szebelledy and Somogyi in three majox respects: the constant current-time method was subst ituted for the chemical crnulQmeter , an .amperometl'io end point was used in place of visual indicator•s; and the quantity of material to be determined was re duced to the micro soale . With similar appara- tus cuprous copper (19) and iodine (21) have been used. as inter - mediate substances. Tha amperometric end point consists of detecting an excess of the intermediate by means of a diffusion current set up between two polarized pl atinum electrodes in a stirred solution. In the case of r eversible half- call reactions the diffusion currant is linearly proportional to the concentration of the subst ance which is present in low ooncont:ration. In addit ion , this quantitative feature of the arnpercmetrie method hes proven useful in the study o:f aqueous dissocia.tion equilibris, 0 and the :results of suoh investi gatit'°ms al'e described below. I ~l, , THE USE OF EL,ECTROLYT ICALLY GENERATED CIILORilt E I N COULOMETRIC TI TRtl.1!IONS Until the present wo:rk only three intermediate half- cell ,, ~eactions had proven s atisfac t o~y for coulometric determinationso, For the titration of ~educing agents, bromine has been employed rather extensively (1 , 20, 23 , 29 ; 34) and recently Ramsey; Faxrington; and Swift {21) have shown that iodine can be used for the titration of ars en:te. Cuprous copper has been used for the coulometrio tit~ation of certain oxidizing agents (19). In order to increase the number of substances which could be titrated by such processes~ a stronger oxidizing intermediate than bromine was sought. Since the halogen half- cells are so rea dily reversible at electrodes , chlorine seemed to be the logical choice for investigation. Wooster, Fa:r1•ington , and Swi f t (34) use d chlorine as an intermediate in the coulometric titxation of iodide in 2 formal hydrochloric acid and observed positive errors 0£ from o.5 to 1%e Sinoe these errors may have arisen from the iodine monochloride formed in those titrations a study has been made of the oxidation of tripositive arsenic by electrolytically generated chlorine and the results are presented below . Reagent~. All chemicals us ed were nReagent Grade . " One volume formal sodium chloride solutions, prepared from the sia.lt , were found to cont ain no ext!'aneous oxidi!&ing or re ducing agents . Sul furic acid {2. 5 formal ) was purified by bubbling chlorine gas through the solution for 2• 3 minut es. Aft e~ boil - ing the a oi d for one hour , no excess chlorine remained in s olution and only a very small amount of reducing ma terial was detected . The reducing mate~ial present i n hydrochloric a ci d was det e:emined by coulomet::eio titrati on wi th chlorine . rhe stoichio- 1 metri c amount of potassium c lo:rate was then added ·to 8 .'E' hydrochlol'io acid. Aft err one day ess ent i al ly all of the chlorate had 1~eacted and the a ci d was :r eady- f or use . Standard sol uti ons of ·t1•i posi t i ve arsenic we:re prapazied in the following way : Bur eau of Standar ds arsenious oxide was dried for two hou~s at 110- 115° a. A 0 . 5 gm s ample of tha oxide wa s a ccurately wei ghed out and diss olved in 10 ml of water containing 1 gm of so dium hyd~oxide . The resulting solution was neutralized with 8 ml of 2.5 F sulfuric acid and about 180 ml of boiled distille d wate~ were added , then t he t otal wei ght of the stock solution was de te:t"mi:nad . Wei ghed samples of the st ock solutions were diluted t o appropriate volumes to provide standard solutions for analysis . Stock solutions we:re disoa.rdad after 5 days: dilute standard s olutions were used only on the day of p~apar ation, Labo1•atory distilled water was boiled 15-20 minutes to remove a small amount of oxidizing a gent, presumably chlorine . Ap-pa:ratu.s . r.rhe npparatus described by 11eier , M.ym.•s ~ =~nd Swift (19) was used With the follo wing changes: The generator cat hode instead of the generato1· anode was enclosed i n a shield. In orde1· to improve current regulation , the laboratory D., C. supply was replaced by a simple voltage- regulated rectifier . The t n• same basic apparatus was used for all eoulometric investigations; only t ho modifications will be noted in subsequant sections . Preliminal'y Adjustm~m.~ • The current of the genel'ation oircu.i t was determined by measur.ing the voltage dro p across a standardized resistance through Which the generation. current was passing. All voltage measurements were referred to a cell which had been checked against a bank of cells obtained from the Bureau of Standards . Generation rates of about 10- 8 and 10- '1 equivalent per second ware used . Yihen not in use the indicator elect:rodes were shorte d to- gether and stored in a solution o.5 Fin sulfuric acid and 0.2 F in sodium ohlo~ide . Before each series of tit~ations both indi- cator electrodes were connected to the generator anode and chlorine was gena?ated at the high ~ate for about 30 seconds . To maintain the sensitivity of the electrodes they were given the same treatment f or 10 seconds after each blank and every titration. nitration Procedure . All solutions had a t otal volume of 45 ml and were 0.5 formal in sulfuric acid and 0 . 2 formal in s odium chloride; the initial indicator potential was set at 0.30 volt. The considerations influanoing the selection of the various conditions of the procedure ara discussed later. A correction for reducing material in the reagents was made by generating in a blank solution of the composition given above for short intervals of time and recor ding the values o:f the indicator currant. A plot of indicator curren·t vs . generation time was constructed and the linear curve was extrapolated to zero time; the intercept on t'he time axis was design.at ad as the "'blank time• 11 When the apparatus had not been in use for several hours the first one or two "blank time" values were usually e:r:ratic and were not included in the average "blank time" value ~ At least three consistent values were used in obtaining the average i,blank time•'' To make a tit:ration the proper amount of sulfuric acid and sodium chloride solution ware mixed in a titration oell with sufficient water to make the volume either 35 or 20 ml . A 10 or 25- ml portion of standard arsenite solution was pipetted into the cell and the total volume was then 45 ml . As soon as the titration cell was attached to the stirring apparatus, generation was started . By means of the bui lt- in potentiometer the genera- ting current could be checked during the course of the titration and the currant was held constant by making small adjustments. Near the end- point the indicator current underwent a momentary reversal, and at this point generation was stopped immediately. After a few seoonds the indicator current rose slowly and reache d -8- - a steady value of less than 10 microamperas .. Generation was continued :fo-r sholl't inte:rvals and values of the indioato.r cur rent and time were recorded . The linear curve of indicator current vs time was extrapolated to the time axis , and the point of intersection was taken as the "titration time" inas"" much as tho chlorine excess is assumed to be zero at the extl'apola.ted ze:ro indioa.tor ou:rrent ,. By subtracting the average ''blank time" fr om the 0 ti tration time" a oorrected ti t:ration time was obtained . The weight of arsenic was calculated :from the values of the corrected titration time and the rate of genera.ti en . Discussion !:Qt,ential Difference A1mlied Ac:ross ,the Indioatq,Lfil._eotrodes . Chlorine was generated in a blank solution until the indicator cu~rent was 25 microampa~es with an applied potential difference between the indicator alect~odes of 500 mv . The indicat or po- ·tential was ·then val!'ied between values from 50 to 1000 values of the indicator current were recorded . mv and Before each change of indicator pot ential the current was adjusted to 25 mioroamperes at 500 mv by the generation of chlorine . thus obtained are presented in Table I . The data When these data were plo tted, the curve of indicat or current vs indicat or potential was similar in form to the one obtained by Ramsey for iodine (21) except that the flattest portion of the curve was fr om 400 to 600 mv . Although it would have been dasi~able to use an indicator TABLE I Values of Indicato~e Curr@nt and Indica.1;o:r Potential for a Solution Containi ng Ohloride a and Chl ol'ine b · Potential (Millivolts) Current (Microampe:res) Potential (Millivolts) Current (M:i c:roam:peras) 60 4.0 553 26 .1 106 7 .5 617 27 e5 175 12.2 696 29 .0 217 14.2 755 30.5 286 18.0 798 32.l 319 19.2 84 8 33.2 362 20 .7 935 37 C 417 23 .0 1042 43° 500 25 e0 a Sodium ChloJ.•ide 0 .2 F . b Chlorine was generated u_ntil indicato~ current was 25.0 mioroamperes at an indicator potential of 500 millivolts . The chlorine ooncent~at ion was appr oximatel y 2 x 10-6 F. c The indicator cur:rent was changing rapidly so that these values al"e unce:rtain. potential \Vhi oh lay (>n this flat :portion of the curve • this was no t p?"act ioe,bl e because high i ni t i a.l i ndica to x ou1"i-ents {that is, befo1.·e beginning generation ) we:re obtai ned in blanl{: solutions. I n order t o reduce this i niti~l valua of the indicator cuT-rent to l microampel'e or lesf1.; it was neoessar.y to 1ise a potential of 300 mv. Thi s potent i al pl"oved t o be satisfaotOl"Y fo:r the t i t ra- tions ; linear cu~vea of indioatol' oui-ient va time wel'o obtaine d for bl ank ti t :r.at:i. one ,, and t he oleot:rode sensitivity was goo,d . The Effect of pH on Indicat or Current. Inasmuch as chlorine tends to hydrolyze even in sli ghtly acid solutions, it was desirable to determine the maximum pH at which a stabl e indi• oator current oould be obtained vvhe:n tha chloride was 0.2 F . In a solution buffered to a pH of 3 with phosphoric aci d and dihydro gen- phospha.te no indicator current could be observed when the low rate of generat ion was used. For a solution in which the pH was 2 an indicator current was obtained but it decreased a t the rate of l mioroampare in 4 seconds as soon as generation was st opp ed; this is not ad.equate stability for quantitative measurements. A stable indicator cur~ent was obtained in 0.05 F H so ; however; to include a f actor of safety, solutions o.5 F 2 4 i n H so were used in the titrations. 2 4 Indi cat or Current Phenomena . The indicator currant response was slower :fo:r: chlol'ine than :for b:r:omine o:r iodine . When eithe:c a blank chlori de solution o:r such a s olut ion containing arsenic was pl a ced in the cell and a potential was applied between the indioatol' eleotr,odes O t b.€1 indicator Cl.l'r:cent surged to a high val ue then decreased slowly. In eit he:r case several mi nutes elapsed before tha m.1rrent reached a steady va lue of about 0.2 mioroampexe. In practica , generation was stal'ted i n a blank solut io n as soon as the indicator ourrent dropped to 1 micro-am.pel'e ,instead o:f wait i ng until 0 .2 mieroampere was reaehed ; no difference in bl ank time wa.s observed e When arsenic was titra- te d. , generation was tiegun :i.mmediately after closing the i ndicui- tor ci r cuit because the indicator current always dlt'opped :rapidly and underwent a reversal near the end point and the value of the indicato:r current during the titration had no significance . Afte~ the first pe~iod of generation in a blank solution and after the current reversal at the end of s. titration , the i ndicator current l'Ose so slowl y that a period of l to 3 minutes was necessal'y for steady state to be at tained . Fol: all other readi ngs of indioat or current a steady state was reached i n 10 t o 15 seconds . The current ravei-sal near the titration end poi nt i s simi• lar to the one obs e~ved by Myers and Swi ft (20); however , with the p~esent apparatus no hydrogen enters t he sol ution horn t he gene1•ator cathode c When the potential is fit'st applied to the indi cator elec t ro des , the cur?ent surges to a hi gh value (mo~e than 50 microampares) then de creases . During the passage of this oul'rent a smal l quantity of hydro gen o:r possibl y arsenic i s probably bei ng produced on the indicator oathode . Neax the titration end point the concentr ation of chlorine increases -12- :rap:i.dly and there is a transition of indicator electrode potenti.als and rea ctions. If there is a significant lag in l"emoving adsorbed hydr ogen {or possibly ersenic) from t he indicat or cathode , a cell v.rou.l d be established in apposit ion to the mpplied potential and a current would flow i n the !'eve:irne direction. The Effect of Chloride and Acid on the Coulometric Tit l"ation. In view of the posi t ive errors obtained by Wooster in the titration of io dide with chlorine in 2 F hydrochloric acid (34) ~ it seamed advisable to use the lowest practicable chloride and hydrogen ion concentrations~ As discussed above 0.5 F H2so4 was taken as a sa.f'e minimum acid conoentration. With a chloride ion concentration of 0.05 F there was evidence that the genera- tor anode efficiency was slightly lass than 100%p while 0.1 F chloride ion seemed perfectly satisfactory. Confirmatory titra.... tions were made using the chloride and aoid concentrations given in the :procedure above, and the results are listed in Table II. To investigate the effects of varying acid and chloride concentrations, titrations were made according to the conditions listed in Table III. It should be noted that each stock solution was titrated by the regula~ prooedu~e, so that the data i n Table III may be directly compared with the data for the oorTesponding stock solutio11 in Table II. The arsenic found v1ha11 t itrat ing in 0.05 F s odium chlo:rida solution was about 0 •.2% greater than the value obtained in 0.2 F sodiu.~ chloride solution; this i ndicates that the current efficiency may not have bean 100% in the Oa05 formal chloride soluti on , Consid ering ·that chloxin0 would be expected to oxidize the platinum anode i n the presence of suoh high chlori de concent~ations , it is su~prising that significant positive er?o~s we~e obtained only i n the t itTations in 6 F hydrochlo:tic acid$ Conclusions Chlo?ine oan be used as the intermedi ate :fo:r a seconda'.1'.'y coulometric p:t'ocess in Which use is m,'3.de of pl atinum electrodes and an amp eromet:ric end point e Th.e acid conaent:ration should not be lowe:r than 0.1 volume :formal, and the chloride eonaentration should be between 0.1 and 2 fol.'mal, In the coulometric titration of arsenic by this method an accuracy of .t 0.2% was obtained fott 300 to 800 micn:ogram qu.ant i ties of a rsa:ni c; 30 to 100 mic:rogTam quantities we:re ·t i tra:bed vvi t h an average e?'1~or of less than 1%.i The material presented in this section has baen a ccepted :for publication by Analytical Chemistry (3) . -14- TABLE II Coulometrio Titrations of Tripositive Arsenic By Means of Electrolytically Generated Chlorine Numbe:r Arsenic (micl'ograms) Taken Found Ert'Ol' %Error I 29.8 30.2 29.9 29.9 29.9 0.4 0.1 0.1 0.1 1.34 0.33 0.33 0.33 I 82.4 83.l 82.7 82.6 0.7 0.3 0.85 0.36 0.24 82.6 0.2 0.2 0.24 I 321.9 321 . 9 321.9 322.3 321.9 o.o o.o o.4 o.o o.o o.o 0.13 o.o I 802.5 802.2 801.8 803.5 802 . 6 -0. 3 -0.7 o.5 0.1 -o.og o.oe 818.6 817.8 817.9 817.8 817.8 -0.5 -1.3 -1.2 -1. 3 -1. 3 II III s 8 819.1 795.6 796 .l 794.8 794.5 796 .l 795.3 o.5 -o.a -1.1 o.5 -0.3 -0,04 0.01 -0.06 -0.16 -Ool5 -0.16 -0.16 0.06 -0.10 -0.14 0.06 -0.04 Roman numerals indicate stock solution used . TABLE III Effect of Aeid and Chloride Ion Concentrations on Titra.tions of T:ripos itive A1..-senic Nu.mbexII 8 Arsenic (miol'ograms) Found Taken Solution o.5 F H2,so4 ; 819.l o.o5 F r.faOl %Error 819.4 819.4 ., 819.0 819.8 0.04 0.04 - 0 .01 o.09 ,. 817.4 , 817.4 -0.21 -0.21 -0.21 817 09 817.9 ... 0.15 -0.15 -0.16 '• II 0. 6 F :a:2so4 ,; 819.l 3 .0 1!, NaCl II o.5 F H2so ,, 819.1 4 Saturs,ted NaCl III III a 817.8 2 F HCl 6 F HOl 817.4 795 . 6 ' 795 . 6 " 794.9 796.4 795.7 -0.09 -0.09 0.10 0.01 798 .4 800.5 799.2 0.35 0.62 o.45 794 . 9 Roman numerals indicate ,stock solutions used . - 16- I • 2o A. COULOMETRIO TI TRATIONS The Coulomatria Titration of Iodide in Hydrochloric Acid The coulometric titration of iodide in hydrochloric acid by means o:f eleotrolytieally generated bromine and an amperometrio end point was first investigated by Wooster (30) . In Figure l it oan be seen that the indicator current at the beginning of an iodide titration may be significant. This initial indioatOl' current introduces a oorraotion in addition to those described in Part I-1. Al though Wooster obtained goo d resul·ts, his method of calculating corrections involved several approximations. It appeared tha ·t a more rigorous method o:f calculating col'rections mi ght improve the accuracy of the titrations. Furthermore, Wooster stated (32) that positive errors were obta ined if the indicator current was allowed to flow during the titration .' In order to clarify these points the experiments described below w·ere undertaken . merimenta,!_ Reage~ts.. Six formal hydrochloric acid solutions were prepared from reagent grade concentrated acid. The acid available commercially was found to contain as much as 3.5 x 10- 8 equivalent of reducing agent per milliliter, The amount of reducing agent pxesent was determined by electrolytic oxidation and was removed RT a: w 2 0.. § u ~80 3 /o--: =:~:::o, a:60 / A/o o: / ~ 3 8~ /4 ~20 / : / 40 ~ ,,.-,,- ~ '/ ct ct _, '-~ ~ ct / ' g B if° "Q ~ ✓ 8~ 8 I I I 20 40 60 80 / I I I I I :::::.gr," I 100 120 140 160 180 200 220 GENERATION TIME (SECONDS) Figa.i-e 1. Bebav1o1!' of Indieetor _OU.iittent . dui-tng Tlt~atlone ln Hyd~oohlorio Aoid Solutions. A. No b:romide pteeent by boiling the 6 F acid with the calculated amount of 3% hydrogen p eroxide . Excess pezioxide vvas dest?oyed by boiling the acid fo~ ab.out 15 minutes. One formal sodium bromi de solutions were prepared from reagent grade sa,l t. These solutions, when tested, we~e found to contain no extraneous oxidizing or reducing agents . Stock 0.1 F solutions of potassium iodide were made up by wei ght from reagent grade salt whioh had been dried for 1 hour at 110° c. iodate. The potassium iodide used was found to contain no The solutions were made 0.01 F in sodium carbonate to minimize air oxidation. Dilutions of the stock solutions, 0.005 F in sodium oarbonate, were used directly fo:r ti t1~a.tions. Apparatus. The apparatus was the same as i n Fart I-1 except ·that the indicator ano de was 1.6 by L,5 cm and the indicator cathode was 2 by 2.5 om. P:reliminaru Adjustment. When not in use the electrodes were stored in a solu'tion 2 F in hydrochloric acid and 0 .1 F in sodium brom.i de. Before use the indioato:r electxodes which were shorted together during storage , were made the generat or anode , and then bromine wa s generated on their surfaces for a period of about 10 seconds on the high rate (10 milliamperes). The electrodes were given the same treatment after each blank and every titration. Ti trati on Procedure~ The i ndicat or po t enti al was set at 138 millivolts as suggested by Wooster (31). The solutions for titration were 2 Fin hydrochlor.ic acid and 0.1 Fin sodium bromide with a total volume of 50 ml. Before each sel'ies of ti trations seval'al blank solutions v1ere titrated in order to determine the blank time. For a titration, tha acid ; bromide and watel' were mixed in a titration call. Upon the sddit ion of a sample o-t dilute iodi de solution, the initi al indicator current was recorded , the generation currant was adjusted and generation was started. Readings of the indicato:r cu.:t"rent were made at sho?t intervals of generation to establish the initial slope of the indicator current curve. By means of the built- in potent iometer th,e generating CU?'l'Emt was checked throughout the couTse of the titrations, and adjustment was made whenever :necessary- to hold the current at a constant value . In general tha indicator cir- cuit was left open until the equivalence-point was nearly reached; variations f rom this procedure ara noted in Table IV. As the current began to rise after the minimum point , readings of indicato~ current were recorded after several short intervals of generation •. Disoussi on Corrections for Initial Cut>rent, Solution BlankL and End Point . Two methods were used for obtaining these corrections. Method A (Wooster), approximate in nnture, aonsisted of generating in a solution, prepared as was that to be used for the titration but containing no iodide, until a suit- able indicator current readi ng was obtained , usually from 25 ·to 40 mi croamperes ., This cu.l·:t•en"G va.lu.e divided by the gener ation time in seconds gave a factor, by whic h subsequent i ndicator cu:rrent readings could be converted i nto equivalent genera:tion times. These measurement s were repeated until the reproducibility indi.oated that the desil'e d cons-t anoy of elec- trode sensitivity was being obtained . I n the sub- sequent iodide titra tions . t he initial current was '.i'eco:rded , then gener ation was oonti nued unti l the i ndicat or cu~rent had gone t hrough the minimum; and had increas ed to values of the same magnitude as those obtained for the blank solutions . The initial current and the final cu-rrent reading ware then converte d to equivalent times by application of the above :factor. The time equivalent to the i nit ial current was added, and the time equivalent to the fi nal current was su'bt:raeted f rom the ·total gene:ra-tion time. In Method B {Faxrington), the correotion for the initial cui-rent was made by :recording values of ·the indicator current du:ring the initial rise, then ex-trapolat ing to zero cur?ent the linear curve obtaine d by plotting these 1,alues against time . Ti me equivalent to the distance f:rom zero time to the intercept of this curve was added to the generation time. When generation was star·ted with blank solu• tions the~e was a shoxt time inte~val before tla current began to rise , indicating the presence of a small quantity of some substance capable of reducing bromine . A oorxection \i.ra.s made f'or this material by continuing generation until the plot of indicator currant ve~sus time was linear , then extrapolating this curve to zel'o current . The time equ;h a.lent to the distance from this intercept to zero t ime was subt:racted from the generation time , The end point of the titration was obtained by continuing genera• tion , after the indicator current minimu.ra , until the plot of indicator current versus generation time was linear • then e:Jt:tl!apola.ting this curve to zero current . Titration values obtained by these two methods are tabu- lated in Table IV . It oan be seen that the theoretical titra- tion value is more closel y approached ·by us:i.ng Method B. When tit:rating iodide in hydrochl ori c a cid in t he absence of b?omide , Wooster ( 33) reported large posi·t iva e:rrors (up to 8%); suoh err ors appear to be largely due to the method of oa.l oulation (A) and actually are more naa,tly on the o:rdei< of o.5%. The ocour:renca of positive e!'rors when the indicator cur l"ent was allowed to flow during tha titration was not confi rmed . The experiments described in this section have been included in a recent publication (34). . TAB. E IV Conroarison oi Hethoa.s _,. and For Calcu..lating The End Points of Iodide Ti. t~at:tons Iodide, JUorog£ams Sexies I (indieat or current off) Fowid Taken Method A :Methoo. B 13/4;.4 J.34.7 134.7 1 36 .0 134.1 135 .1 134.2 Av~ II (indicator current off ) 134.4 Av. III (indicator CUl'l'Emt on for hal:f of titl'ation) ( ind:i.oato:r current on 134.3 continuously) IV (indicato:r current on through minimum point only) Av. 134.3 V Cindicat ol' current off) 1 34 .3 Av. 134.4 1 35 . 6 134 . 9 135.5 135.2 134. 3 133.9 135.4 1 34 .4 133. 6 134.7 134.1 1 35 . 6 134 . 3 134.6 134 .l 133.3 l33e4 Av . 134 el 1 33.8 133.8 1 34 .6 134 .2 134 .7 133.8 134 .2 134 . l 134~2 1 33 . 9 134.l 133.9 l34e0 I - 2B~ The Ooulometric Ti t~ation of Iodide in Pex chlo~ic Acid The ooulometric titrat ion of 1.odide in pe:rohlo1•ic acid is essentially tho same as the titration des cr i bed in the preceding sec t i on (I • 2A ) , and it was studied only as a supplement ·to the i nvesti gation of the iodine monobromide equilibr11tm • .Experimental Reagents . The 6o% pel?ohlorio acid required no special :pu?"ifi ... catiOlle Potassiurr1 iodide containing no iodate , was dried fo~ 1- 2 Stock solutions we:re made up by weight to contain approximately 10- 4 foTmula weight of iodide per gram of hours at 100° c . solut ion f and wa:re made 0 . 01 F i n sodium cal:'bonate to minimize e.ir oxidation. Ths iodide concentrat ions varie d lass than 0 . 1~ f:rom values o1)tained by ti tr.at ion to the · iodine monoohlo:ride end- point With a s t andard iodate solution. Portions of the stock solutions were delive~ed from weight bu~ets and diluted with 0 . 00 5 F sodium carbonate. Only :freshl y p:rapared dilute sol utions were used for t i trati ons . ~oaratus . The apparatus used i n Part I-2A. was also used for this investigation. Procedure . A total volume of 45 ml was us ed and the solutions were 0 . 2 Fin sodium bromide and 2 F in :pe:rchlorio acid . In - 24all other respects the titrations were performed exactly as in Part I - 2A . Calculation of the amount of iodide found was made exactly as in Method B of the preceding section. The results of con- firmatory titrations are pre s ented in Table v . Dis cussion The most important advantage of using perchlorio acid instead of hydrochloric acid lias in the purity of the former . Reagent grade perchlorio acid contains so little reducing material that it may be used without special treatment , while untreated hydrochloric acid causas excessive blank time (See Part I - 2A) . In addition , the platinum electrodes are less subjeot to oxidation in perchlorio acid solution because there is no complexing effect. If titrations :require more than 200- 250 a econds ., low results will be obtained; p~obably because of lo ss of iodine . The~e is soma evidence that low Tasults will be obtained if the solution is above 25° C during titration. For best results it is ~ecommended that the temperature of the solution be 18- 23° c. TABLE V The Titration of Iodide i n Pe~chloric Acid by Means of Electl'olytically Genal'ated Bromine Iodide (micrograms) Taken Found 45.3 45 . 9 45 . 4 113,3 113.2 113.2 113. 4 134 .O 134.5 134,5 482.7 482,8 484 . 8 482.l 484 .1 484 . 1 484.1 1203.6 1201.2 1204.6 1203,9 1203. 9 1201.9 1203.9 - 26-- The Coulometrio Titration of Iron 1~20 0 Tula.ny investigators have recommended procedu:i:•es for the determination of iron; however , very accurate methods for daterming miorogtiam quantities of iron are somewhat limited . For example , Cooke , Hazel and McNabb (2) find errors up to 2% when titrating less than O. 5 mg of :ferrio iron with ohromous chloride ; this is typioal of many vol umetrio procedures . On the other hand Mehlig and Shepherd (16) describe a spectrophotometric procedure whieh oan be used to determine small quant ities of iron with an accul'·acy of 0.2%. In studying the composition and properti es of blood ; investigators at the California Institute have sought a. convenient method for deter mining il'on in small amount s . To them the available volt1.metxic and colorimetric techniques did not appear to be satisfacto~y. In an effor t to dev elop a rapid and accur- ate method for the titration of micro quant ities of iron the application of a secondary coulometric process to such a. titrati on was investigated. Inasmuch as no intermediate has yet been developed Whi ch will quanti•tatively reduce foz- ric i:ron., methods of oxidizing ferrous iron were considered . The titrat i on of ferrous iron by means of eleotTolytically generated chlorine has bean studied , and further development of this procedure is re commended . A new appr oach to the problem involved the use of iodine monochlorida as the oxidizing agent . A small amount of iodide was oxidized to iodine monochloride by the generation of chlorine, than ferrous iron was addE.)d to the titration call . The iodine monochlorida end point was restored by the genera ... tion of ohlol'ine and the time :giequ.i:red fo r that oper ati on was ta.ken as the ti t1 at ion t i me . 4 Fur thel' investigation of' this system should be p1:ofita.bla . Appsratue. The appara tus was the same as that used in Part I - 2A ., Preliminary Experifilents . 1. To a 40 x 80 mm titration cell containing 15 ml of 6 F HOl in 40 ml of solution was added 5 ml of 0.005 F FeSO. 4 Chlorine was then generated in the cell at the rate of 10- 7 equivalent per second 1 and, if the ferrous iron reacted stoiohiometrically, no chlorine excess should be indicated before 240 seconds of generation. After about 10 seconds of generation the indicator current was greater than 50 microamperes . When genera- tion wa s stopped, the indicator current decreased fairly rapidly indicating that an excess of chlorine had been generated before the equivalence point and that the ferrous iron was reacting at a measurable rate with the chlorine . An accurate titration would be impossible under suoh conditions. 2. To a titration call containing 15 ml of 6 F HCl in 40 ml of solution was added 1 . 5 x 10- 5 moles of KI . Chlorine was generated until the iodide had been converted to iodine mono chlori de an d the indicator current was at the minimum point of 2 mic:coa-mperes (Se e F:i. g o 1) . U-pon the addition of 5 ml of - 0 . 005 F Feso the indicator eurrent increased t o a high value . 4 When chlorine was generated the indioatotr current dec:i::.•eased but passed thl?ough a mini.mum cul:' l"ent of 20 mic!'oamperea before 200 seconds of generation. Again there are indications of a slow rate of rea ction wi th ferrous iron and a successful titration is impossible unde~ these conditions . 3. To a titrat ion cell containing 15 ml of 6 F HCl and. 10 ml of 85% H Po was added 7 0 5 x 10• 6 moles of KI , and 3 4 chlorine was genel'ated until a minimum i ndicator curr<:mt of 3 mi c7eoarnperes was a t ·tainad . Upon the addition of 5 ml of 0.005 F the indicator cur~ent i ncreased to a high value (off 4 sea.le) o When ohlori11e was generated , the i ndica-t o1• curl'ent de- Feso c~eased steadily and ~eaohed a stable value of 3 microamperes after 241 seconds of titration. Three successive titrations by this procedure used 241 . 2 ~ 241 . 9 , and 241.3 seconds of genel'a- tion re spectively. 4. These conditions have defini t e possi'bilitias . Experiment 3 was repeated with only 5 ml of 85% H3ro 4 • As the equivalence point was approached , instability of the indica.tor curren·t with time gave evidence that ·the oxi dation of ·the :remaining :f'err-ous i:ron was not instantaneou.s e Evidently the conce:n.tration of phosphoric a ci d {o~ H3Po ') is oritioalci 4 5e To a titration oell cont aining 15 ml of 6 F HOl and 10 ml of 85% H3 Fo 4 was adde d about 2 ml of a dilute solution of Feso4 • Chl orine was gana rated , and the oxidation of ferrous iron appeared t o proceed smoothly. The indicator current m- ..-29.,,. creased as for a :normal end point i and the goo.d stability of the indica tor cu!'rent ga,ra enr:tdence that, under these concli• tions O there was no measu:ral.lle rate o-:f reaction between chlo:r:i.ne and ferrous i~on . 6. In order to provide for the addition of 10 to 15 ml of soluti on from a reductor a 125 ml titration vessel was used. To the ti·tration cell oonttaining 25 ml qf 2 F HCl and 20 ml of 85% H Po was added 2.5 x 10- 6 equivalent of Ia , and the total 3 4 volume was 60 ml o Afte~ oxidizing the iodide to iodine monochlorida, several milliliters of dilute Feso4 solution were added. No difficult y was . experienced in titrat ing to the iodine monochloride end point; evidently the :reagen·t propol'tions are correct for these volumes. , 7. Experiment 6 was repeated with only 15 ml of 85% H3Po4 • Inasmuch as a measui-allle reaction :rate was observed near the end point , the 20 ml of · u Po oap. be oonsidal'.'ad the minimum re3 4 quirement :for approximately 75 ml o:f solution. To a titration cell containing 20 ml of 85% H Po and 3 4 9 ml of 1 F NaBr was added a small crysta l of Faso 4 • llromine a. was generated and almost immediately the indicator current b egan to i n crease. Whenever generation was stopped , the in- dicator current decreased rapidly: evidently the ferrous iron and bromine were rea cting at a measurable rate even though the H Po was 3 4 present . Investigation of Micro Reduct ors. In Ol'der to avoid large v0lumes of wash solution, i'edu{)to:re 0£ small size wete p1epa1ed. UnltHJ$ crtn~~Wiae specif18 d the eoluxm 0f r(ldueing tnate.3;'1al was .about .v to 8 om higb and e mm .·t.1'.l diameter. ~'h~ tit~!ltio.n plto- ce.d~e· deaC"~ibed tn pralimill$l'Y . eJq,e:riment 6 wa.s •used be:oa:u,ee a~ 1n$wt _stmo.sphort1 Wa$ unntce~ealry and th$ aeouraey of that pi-0•<:HtdUJe appeated to be suf:fio.1 ently go.oil to test itedv.e1H:>l's• Instead of ueing the minimum tndi,oato.r . cufl'ent as e. 1tef.Cl)180nce JH>int • a . value o:t . the in<U,C$tot eu:rt'tnt on the iodin~ st:de of the mini.mum PQ.int Via$ tak~n b$es,u se o:f th~ giteatet m&tai :res,.. :po,nse for E)atb seo.ond of g4.l»t:rat1.o n in t .hat l'egton, A d:llute solut.1on: of itien~ l F 1n iOl, was pl'epai'ed by dleeolV"ing a weighed sample of Ju:t~. il'On wi:rte., . 1~ SilYtJ .Bf.ldu.gtQJr.* Finely di'vided metsl.li~ sll. var was p3'e- pat"ed by ~edue_i:ng silvei- ion w1t .h ooppe:r, in a sotu,·U..en alightly aei.d ifi.$~ W1 th eo ~ A 5 ml po:rtion of the . (ii lute iron sQlut:lon 5 was :passt« th•()lltg)l the f$du.oto'r s.nd into the iQdine monoehlo.ride s01ution tl,1 about 3 mi11ut1s. Ten m1l.lil11Ha1rs of 0.5 F ROl W$t"e used to wash the teduc,t()~. Suocessi'V'e titt"ations l'equi:red pro• g:rees!v-ely shorte:r titvation times, 1ndieating that the small silver :reduetol.' was probabl.J being :rapidly exhau$ted by the fo"tmation of' a su:rfiu,e coati.ng of sil.vel' ehloild&. Effo~ts we1e made tf> maintain a ftesh ail V8i- $UYf.a et by inse:rting an anode in the solution above the s11 ve~ oolu.mn and applJing a potential b,e tw8en the silver ec>t'UJiin end the poeittve eleot~OdEh When a pltltinm.n anede •as 11;sed; oonsidete.bltt hfd,togen 1t1as evolved hotn the silvel column; Wb.en a silvtr wt?:e anode we,$ usei, tbtl anode surface appea:red to be po1at,1J1;ed afte~ one or two aamplea ha.d paased th:ough the :,eduetot" • In au.ob small stze the ail vei- JH)dueto:r was unsatisfaetoty,. 2. Z1n.o .Redu.e.toi-. A s'Ul:':faoe OQat.ing of amalgam was applied to 30 tnesh ~i:no bJ pou~ing the id.nQ into a ve~y dilute solution of ,·, ' ' RgCl • 2 ( ztne coltllnn about 7 om bJ.gh) and .anot'he:t of the size tn,d ieated above. With both reduoto~s the titi-stto:ns va:rted by stv<J11al pei,toent. Inasmuch as nydl'Ggen bubbles a;ppeal.'ed in the column; :1 t 1$ probabl.e that portions of tht qo1umn w&:te bloekEHl. 3 .• . ,OagJpip ,llE!duc1a~,l:.. A ·20% oa<lmiU.m amalgam was p-repatt.ed as d.il"eoted. by ielge:r ( '1) e th$ s.~l ga.m is solid wnan oool. and oe.n be: granulated. When 'f;he cadmium i'eduotor was used. the titl'&• tions we-re moYe cona.iitent than with the silver or zino reductoi-a; howevtti,; the t1ttation V$l'WH3 still. val:'ied about 31'. Hydrogen was fo.rmed in sml1 amount and may aoe.ount foi- the va:riation. Pel"hapa s tower hfdtogen ion oonoenti-ation in the il'on solution would give betteft ,:esul..ta. With the cadl'niwn reduetor. 4• Bismuth :tte<ln.r~to~• Altbongh Someya { 24) used a bismuth amal- gam to ~educe feir~ie i:ron ., no 11~fel'ences he.V$ been found in whieh metallle b1$muth has been used as a reduoing agent. Finely divided metallic 'bienmtb (b1aek) was ,Pl.'epated by lmtne~sing a ettlp of zlno 1n an aoid solutt.on of »101 3 • About S em of this bismuth in a eection o:f a 50 ml bu.1."et oonst:l.tuttd the fit st t~ial reduet()t. In a $E>-i "ita qf :t'<i>U:t' ti tratione using th1a :reductor the maid.mum va~iation in titration timt WS.$ only a l i ttla over 1%. Similar re~ml t-;s we:re obtained with a :r.eduoto:r 6 mm in diameter . The bismuth reducto::r. appea:rs to be the best of the reduotors studied o :Recluction of I:ro:n QX Bi smuth and Ti t:tati011 with Kl.ect:eolyt$.eallz Generated Ohlo:r:i.ne . In ora.ei· to more t,horoughly test the b:i.amuth redu.cM.on of iron and to tzry a motie st:raight:forv7a:rd typo of t i - trat i on , the chlorine tit ~at i on of iron was investigated., Apparatus . The apparatus was the same as that used in Part I -2A. Reag~~t!* The matet·ial i n 8 F HOl which was 1.•educing to •ohlorine oxid.ized by the addition of KC10 3 as described 1.n P.al.'t I • 2A. Phosphoric acid. (85%) was freed of material. :reduci ng to chl orine by a ddi ng an exce8s of chlorine , allowing the noi d to stand f or sev-ere,1 dayr1 , and then boiling the ao1. d vigo1.'ously t o remove excess chlorine . A stock s olut ion 0.1050 Fin fe rri c iron was p~epa~ed by dissolving FaCl3 • 6H20 in l F HCl . fer:x·icyamld.e wo.s negative . A t est for :fert"OU8 i:¥.'On with Standardi.zat ion of the farrio solu- tion was cal'l'ied out by the iodometrio })l~o cedut·e suggested by Swift ( 25} • Sodium thiosulfate f.01.~ the titration WbS stanc1a1:1d-- ized against 10:0 using the s taroh end point . Dilut i ons of the 3 st ook fer:ric solution ~ 1 1'1 in HCl , wel'e used for cou.lometric titrati on.s . Procedure . The usual pr oces s of shorting the indicat or elec- tro des and genera.ting intermediate on their su:rfacas :for se ... varal seconds wa s carried out after each blank and tit~ation . For these ti t:rat;ions an :tnte:-i.·mediate ,:ate of genor.ation, 4~131 x 10- 8 e quivalent pe:r second , was use d. Tho i ndicatott potential wao 300 mi llivolts . To a 40 x 80 mm t i t ?' at:ton cell w• .s add0d 20 ml of 85% H PO , 3 ., 0 ml of 8 F Af·tol" pla cins tho • oell in posi t:i. O11 on the stirl":i.ng appar a tus , tank ni t1•ogen was 3 4 01 ,· and 7 ml of wate r. bubbled through the solution for 4 to 5 minutes . The glass ni t:rogen i nlet tube wa s t hen wi thdxav1y1 to the upp el' :part of the cell , and a constant flow of nitrogen into the cell was main• tained f ol'.' the duration of the titration . A 5 -ml sampl e of dilute fex~ic ahlo:t•id a solution was pipetted i nto the reductoawhich had been filled v1ith freshly praoipi tated bismuth (bismuth column 8 cm high , 6 mm in diametal') . By meairn of a stopoock the flow through the reduc',.;ox was :trngu.ls. ted so the;!; 2 to 3 minutes wa1•e :requit'EHl fo"J: the sample to flow into the JGi tz>a1;ion cell . Ten millili t e:rs of' 0 . 5 F lICl wore use,1 to wash the radm;rlJor; the to·taJ. volume in the tit ration cell was then 45 ml . The indicatoii ai:r-oui•t; wa s olosed 0 and chlorine Wt.=t s gene:t:<a'"' ted until the indicator curr ent :reversed to a negative value . At that :po:tn:t generation was stopped and t he indicator current increased to a positive value in about 30 secondse Readings of indicator current values iive:re then recorded after 5- seaond :tn... tervals of generati on . Reagent blanks were prepared and swept out wi th nit:rogen . A 5 ml portion of the HCl used in the sample was passed thrqugh the reductor, followed by 10 ml of Oe5 F HCl ~ Ohlo~ine was - 34- then generated for 5-sacond intervals and the corresponding values of the indicator currant were recorded . The linear curves of indic ator current vs . time were extrapolated to the time a.xis . Tor the titration curve the intersection was taken as the ~titration time . " The intersection in the case of a blank wus designated the "blank time . " By subtracting the average 11 blank timen from the "ti t?"ation timett a corrected ti tration time was obtained . Discussion The results obtained by the above procedure were accurate to about 1% and indicate that further investi gation should ba profitable. With samples containing 233 . 7 micro grams of iron titration values of 233.0, 230 . 7 and 234 . 8 micrograms were found . The bismuth micro reduotor appears to be quite satisfactory for this application . If allowed to stand in acid solution ex- pose d to the air 0 the bismuth will dissolve fairly rapidly . After the reductor has been allowe d to stand for about t wo days, the bismuth column seems to be more tightly packed : with freshly pr eci pitated bismuth is :recommended. of the metallic bismuth has not been checked . replacement The purity In particular , iron should not be present ; therefore , a sample of the precipitated and acid washed bismuth shoul d be dissolved and a test made for iron and :perhaps copper ., The amount of R3Po4 used in the above pro ce dure signifi- cantly increases the solution viscosity and decreases the indicator electrode sensitivity • .After the addition of ferl;'ous solution from the :reducto:r , there was ,an indicator our~ant of about 30 microamper es ., This current deo:reused during the titration and ourl':ent :reve:rsal occurred near the end point as mentioned in the procedure . Perhaps some hydrogen was being p~oduced in the reductor; reduction of the acid concentration in both iron solution and wash solution should be tried . Suggestions for Further Investigation Chlorine Titration 1. R~duce the acidity of the titrat,ed solution with the possibility of reducing the n ~o requir~ment . 3 4 ~. Use 1laH2Po plus HC10 in place of H Po ; there may 4 4 3 4 be less reducing material present . 3. Use the maximum permissi ble pH (about 1,5) , and t~y fluoride as a substitute fo? H Po • 3 4 Iodine Monocnloride Titration 1. As a refer enc a point take an indicator current value on the chlorine side of the minimum point ; the meter sensitivity is greatest in that region G 2. In order to shorten the perio d during which iodine and iodine monoohloride are i n contact with the platinum electrodes, two possibilities appear f or reducing the iron and than running it rapidly into the titration solution: a) . Reduce the sample in a smail bulb , flus hed with N o~ CO , by a dding finely divide d bismuth and 2 2 swirl ing foT a few minutes 5 By means of a stopcock , drain the sample rapidly thxough a plug of glass wool and into the titration cell . b} . Construct a reduction column with a 20 to 25-ml chamber placed below the column stopcock . Flush the chamber With inert gas , and collect the sample and washings i n the chambeT . Through a second stopcock quickly run the sample and washings into the titration cell . EQUILI BRIUTui STUDI ES ' A. Egui.lib:ria i.n Iodine Mon ob romi de Solutions During an io d:ida titration the i ndicator current passes through a minimum value i n the region of the io di ne monobromide ,· equivalence point (See Fi g . 1) . Sinoe the indicator current is limited by the io dine concentration bafo~e the minimum point end by the bromine concentration afterward , it appears that the value of the minimum current is related to tha concentrations of the halo gens in equilibrium with iodine monobxomide . A mathe- matical correlation of the minimum indicator current with the cupric bromide equili"brium has been made by Meier (17); Kolthoff and Lingane (8) pre sent oaloulations for the silver iodate equi librium i n which only cathode :reactions are involved . In order to simplify study of the iodine monobromide equilibrium interhalogen chloride oompiL exes 'I.JI/ere eliminat ed by using perohlo:rio acid i n pl a ce of hydroohloric acid~ 1¥Pel'imental ~eagents o The reagents desc~ibe d in Part I - 2B were used f or these experiments . (),~ /{ ApE~ratus . The instrument described in Part I - 1 was used with two modific at ions: a)o For the iodine monobromide e quilibrium studies the microammetar was replace d by a Leeds and North:t'up - 38- box- type, reflecting galvanometer . A shunt re- s i stance of approximately one ohm made the gal vanometer sensitivity one micl'oarnpere per' sea.la division • .As a result the potential drop across t ha galvanomata:r was always much. l ess t 'han one millivolt, and the potential applied to the in• dicator electrodes was held constant without a djustment. b ) . In order to study the mechanism of the indicator diffusion current s , two titration cells were connected by a glass bridge containing 2 formal perchloric acid . A sintered glass disk in the bridge retarded flow bet ween the cells. Both cells were equipped with pl atinum electrodes and constant speed stirrers . It wa s then possible to separate t wo indicat or electrodes and to determine which pairs of :i.onic or molecular specie s woul d ca.rry i ndicator currents . Procedures and Result~. The data f or Fi gur e 2 were obtained by recording indicator current as a function of appl ie d potential for a solution which was 2 Fin parchlorio aci d and 0.1 F in sodium bromide and which oontainad 1. 62 x 10- 5 mole of iodide titr ated to the point of minimum cu:rrent. · Since the flattest portion of the indicator current- volta ge curve lies in the range 40- 70 millivolts , an indicat or potential of 65 millivolts was chosen for the study of the iodine monob:romide equilibrium. I 0 40 ,,, ~ .,~ a. E 35 0 0 ...u i 30 0 25 0 ., I C: :: 20 ::, (.) 8 I I lo 15 -0 C: 10 5 D2 .04 D6 D8 .10 .12 Potential Difference Between Indicator Electrodes (Volts) a. Flgui-e Yetiatlan ot IndJc,atot: ouii-•nt with Ptlen\tal Difte,r•noe Between lndl~sto1 El.eetl'-o.e~ toi,, a Solution Oonta.tntng 1.62 x 10• 5· Molt ot lotU..IUt Mtnob•omide. Data for equilib?'lu.m ealculsit:tons were ta.ken in solut1ens 2 Fin perebio~lo aetd and either 0.1 F o:r o.3 Fin sodium 'bl'o• mide. About 98% t:f' the known aniount et iedide tn each solutton was oxidlzed to iodine tnonobromide by the so.di tion of eatui-ated b:rcmd.ne wate,r • By g$!le~nt ing at the constant 11at e _()f l.039 x 10"'"8 equi ve.lent per secQnd t<-1r snott in1;alivals of ti.me ox.1• dat1on of the iodine waa .(:ompl$ted~ and the 'V':a1ue of t:}1e indi• oator ourri-ent was obs$J1~ed ~i-ing this p~QcQss. A typiea.l set of data talcen in the :wegion of tb.e iodine monob:romtde $.quivs~ lEJnee point ts ahown 1n Il1igue. $. When 6 x. 10•8 ntole of iodide was tit:rat~d, a small. cl.n'rent (0.3 to o.6 mio:rQanrpe:re ) \f$-S noted st the m:tniinum p<d.nt • With this quantity of iodine; mono• bitomide present the mini.mum :indicator cu~l'ent abould hate been essentially zei-o; o-o-nsequent.l u .• this obse~ved indicator e1u~rent was eonsldered to be a a-esidual elurrent Whieb should not be attll'ibutGd to halogens in. equilib11tum with iodine manobYomid$. T.he two-oel.l e.l'rangement d~sevtb ea in ttie fou:rth pa:ragre.p h of appal!'atns was use-d with eeveral pairs of solu:tiona. Vlith a solution 1n each cell,, a platinum $l$eta-ode in one eall was conneete<l as an indicator e1nod$ e.nd an.- ~leet~od€i in the otaer oel.l was made an indieatott cathod•• SS.nef!t the cii-,o ~i t was oompl~ted by the acid•filled btidge., eu.~t~n·t pase;J.ng througb the 'G lee• tiode.s was gove:rnsd by the nat~e and eo:ncent:ra:tion (j:f the halo• gen eornplex-es p~eijent in tba solutiQns. This in<lieato~ e~lH;Jnt; is thetefQ~e simil.aa:- to · the diffusion eu:i:reats obtained when the tle~ti-odee &re eontsine<l in ont cell. In 4?able. VI a~e p:,esented ,...... VJ QI :U 25 Cl. E 0 ... 0 -~ 20 ~ 'E ...~ I 0 ... ::, (.) 0 o .!:! 5 -0 C: 20 40 60 Generation Time 80 100 (Seconds) 120 Jlgur• 8. lt'ban.o, et l .a41oat;oa, -O llh•n" f.a tl:ut • ltsJ.cn• of tht 14JaimU. 1ola,. loi\tae M.onobrend.t• s.eo x 10•4 ftffl81• tha Tasults of a series o:f experiments in which all solutions we:re 2 F in pel'.'chloric acicl. and 0.1 F in sodium bromide . The potential dii':ference between 1jhe electrodes was 65 mv . Experiments l ana. 2 show that iodine, in the presence of bromide , Will react at both anode and cathode to give an indicator current. Such a mechanism had been proposed (33) as the explanation of the indicator currant maximum near the iodine equivalence point, Since many substances do not react readily on elect~odes , it is of interest to detel'mine whether IB1• 2 • undergoes reduction at the indicator cathode. Under the conditions of experiment 4 the ooncantration of iodine and the concentration of bromine are each about l x 10- 7 F in the cathode cell q Considerati.on of the indicator currents and the halogan concentrations of experiments 29 3, and 4 leads to the conclusion that IBr 2 does reset at the catho de .; Calculation of Io_r:!i.ne I~Ol]:,obrgm,i de Eguilib_r~ Q,onstt:!._ll!• It has been observed that befoTe the minimum point is reached the in• dicator current is linearly proportional to the iodine concen• tration . After passing through the minimum , the current becomes linearly proportional to the bromine ooncant~ation. Therefore, in the vicinity of the minimum point , it seems reasonable to represent the indicator current by the expression , i : k ( I * ) + k ( B:r * ) 1 2 2 2 ' (1) in which a sta:rred qu.anti ty indioe,tas the total concentration of I I I 2 3 4 s. Ih l.08 X 10-S 2 2 a B~ I 1.08 X 10-5 .. I ~ 3 25 6 .6:6 X 1,0 - 4 9 0 (M.1 c:r-o-ampe~ea) Indicator Our~ent 5.30 X l(J!f'O''f tl •.B x 10-4 5.8 X l(),•4 F•rmsl OEJ-nc,ent.~stion Cathode Solut&.on Aet-tv:e Clon.s:t i tuent a no X ;_fi/7 l •V~: 1 n-S ,.. Formal. Ooneentration tential d1:ffe-ren-c.e bGtween the ele-ctrodes was 65 millivolts. In tb.e:se solutions the iodine,. homine, and iodill'.$ m:ano.biiomid'& -exist pr1ina~1ly as. the ions IgB~- ~ Br -:-; and l k- - :res~$et1v~ly • . All so1.ut1ons 3 2 were 2 F in pe:rchloric 3t'dd and 0.1 F in sodium bromide, and the po .. 2 2 2 ·- -- l No. Active 8 Consti tue1rt Anode S1~1ut1.o n Diffusion Currents B&twe,en Two 1 -S Cllate·d El.e:eti-odee TABLE VI \ ·• :1 M ~ - 44- all species wliioh contain that halogen ·in the oxidation state indicate d by the formula . From the aquilibl:' ium ex-pression , (I *) (Br *) 2 2 substitution in equation. (l) gives Since the concentration of iodine monobromide is lar ge with respect to the iodine and bromine concentrations , it is essentially constant for these studies and m~y be represented by (2) Di:ffe:re:ntlating with :respect to time of generation , - di "I! dt kl d(I 2*) dt K*C 2 d(I 2*} - k2 (I *)2 dt _Q._ 2 At the point of mini mum current Therefore, at the minimum point ... k - k2K:·co2] d (I2*) ~ l (I *) 2 2 dt • k K*o o 2 • o 1 - Ka ( I ~ ... and (3) (I *) • 2 m Substitution in equation (2) and collection of te:rms gives as the relation for the minimum point current , im' i m : 20 o fi l k 2K* • The value of the equilibrium constant is then given by tho exp:r.•ession, (i ) 2 K* : __m""".- - 4k k C 2 {4) • 1 2 o Di.ffe:rentiation of equation ( l) with respect to time gives • Before the roinimu.m point i n the region where the b:romine concentration is very small compar ed to the iodine concent:ration . d {Bl' *) 2 is neglig:tble Hnd. ---dt d (I kl -d.t - ..dt2 di ' *) Tr,,,> • Denoting the generation ~ate as Kg (l . 039 x 10- 8 equivalent per second) and the volume of solution as V (liters) d(I *) - 2- = d.t -v· 2K - g and. kl : _ ( di_/ dt) V 2K g The value of¾¼ may be determined graphically by measuring the slopa of the linear portion to the left of the minimum point of the indicator GUrTent - t.imo curve (See Figure 3) . manner k 2 In a similar oan be caleula t@d f:rom the ge:ne:rat,or constant and the slope of the linear portion of the indioato~ cu~rent curve after > ·the minimum p oint.• Dat a for the evaluation of k , k , and im wa re obtained as 2 1 desoribed above . The values of 00 we~e calculated from the initial amounts of iodide added , assuming eomplate conversion to iodine monobromide . Values of K* ealculatad from equation (4) are pr esented in Table VI I. Since i.odine., bttomine; and iodine monob~omide all form complexes with bromide ion ; the equilibrium , shoul d be considered.,, Unde:r condj.ti ons app:eoaching those of the pTesent wo:i:.·k the dissoci ati on constants o:f lial~l' - and 1h•3- may be taken as o . 058 (22) nd o .055 (6} re spectively , A more recent value of the I Br- constant has been obtained (9) . but 2 the equilibrium was studj_eo., in neutral rather than acid solu• tions. ~ Faull (4) states a value of 2,7 x 1o•v fo~ the dissoo- iation constant of IBr 2•. With the aid of these dissociat ion constants values of K were calculs:ted :from values of K* and the results a:re given in Table VII I! Forbes and Faull {5) determined the value of K* to be 2 . 06 x 10· 8 in 0.9·7 4 F HB:r and 3 . 00 x 10· 8 in tJ: . 725 F HBr , and the oovr.asponding values o:f I{ are 1 . 82 x 10- 8 and 2$92 x 10- 8 • Using these values of K one calculates potential v·alues of • 0 . 87 and •0.88 volt for the half-oell, as compared withe valuo of -0 . 92 v which was cal-culated from the ave~age value of Kin Table YII . For these calculations the potential of the bromi de- t:rib!'omide half- call was ·taken as - 1.05 v (10)~ and 1 . 34 x 10· 3 M was accept ed as the solubility of iodine ( 26) . Since the agreement between the present work and that o:f Forbes and E'aul l is not particularly good , the experimental technique was altered somewhat in an effort to check the order of magnitude of the equllibrium constant . In the titration cell was placed a second indicator anode which had a sensitivity of one- half and two- thilJds that of' the regular anode when the concentrations of the ion reacting at the cathode we-re about Ool F and 10- 5 F respacti vely. 1for an · iodine mono·bromida soluticn at - 48 the point of minj_mu.m indicator current , the indicator current was observe d when each of the anodes was i n the circuit . S1,noe iodine is the only current limiting substance being oxidized. at tha indicatol' nod o , the dao1:eas a i n indicator currant , A im, observe d on switching from large to smal l anode is about onethird , or at most one- half , of the indicator current carried by iodi ne at the la~ge anode . From equat ions (2) and (3) it can be shown that the indi cator current c a1•rie d. by iodine at tho minimum point is given by the relation , (i 1 2 *) m : r: /k l k 2K*C o 2 • The highest possible ratio of f 1ro il *) 2 m wi l l give a minimum value for K* ; therefore , sin ce A i i s at mos t one- half of K*min • A series of six de terminations o:f 6 i with th:ree di f fa:rant m iodine monob~omide concentrations and two different bromide .. 7 concentrations gave an average value of 4 . 4 x 10 for K~min • )l• Because o:I: the rel a tively small values · of ~ i , these caleum l at i ons a1·e less precise than the ones involving i • Howevel' , m - 49- these experiments do serve ass check on the assumption that the minimum indicator curxent is determined primaril y by the iodine and bromine which are present at that point . 6. 7 4.3 1 2. 9 24.7 9. 4 17 . 7 8 . 39 6 . 30 7 . 98 6 . 76 8. 75 a . 02 7. 75 2. 38 1 . 16 z .20 2. 62 2.oa 2. 01 2o 46 (IBr*) 2 K*: (I 2 *)( Br 2 *) 0.100 0 . 100 0.100 0 . 100 0 . 300 0 . 300 0 . 300 E. 60 X 10- 4 5 . 60 X 10 - 4 1 . 12 X 10- 3 2. 24 X 10- !3 5 . 60 X 10- 4 l el2 X 10- 3 2o 24 X 10- !3 2 (F) Br - (F) I Br - (I2Br-)( Br - , K3 (!Br - ) 2 * cor~ected f or ~esi dual current 4 06 im* (microamperes) k2 (11a/ 10- 6F ) kl (ua/ 10- 6F) 0 . 816 1.10 o . 945 1.72 1. 31 2. 03 1 . 80 K* x 1 0 6 Dissociation Constant of I odi ne Monobr omide TABLE VII 5 . 64 7 . 59 6 . 51 6 . 78 5 . 16 a.oo 7 .10 K x 10'1 I 0 I (.j) Jh oiJ.dation • Redue'b°1on Ef!uilib:t'':ta In Oup:rie Bl!'oniide Solutlo-n~- While studying eoulo11H!t1td.c titl_tations in aalutions eon,.. taining cupr:le eoppeY! (0.005 F) and bromide (0.4 F) M·e ter (18) obse:r.-ved that the indieatoi- eur:rent never deea-eased to zel'o but always had an appreeiable minimum value. He de?ti?ad a l!elationship between the minimum indieato?t cu.tt~tant e.nd the constant fol' the equilibieium, Assuming that the potential value for the b;romlde•tl'ib1tomide hSlf•cell was well established, Me1eit then ealeulated a molal potential fo:r the half•oell • Although Meier's potential value for abOV$ eopper half cell was within 26 millivolts of the 'Value which ho oaleu.lated :f:rom data gi tren by Lstimezi {12,13) ~ the bi-omide oomplex:es of cup:ric- co:ppe:r.- had not been taken into account. When the data pl'Osented in :Part II of this thesis were e.vsi1able. oorl"ections fo~ th$ presence of au:s:r+ and GuB:r well'e applied to Meter's 2 data. These eol'11ections inel'eased tn• a1verg'Jn<te betw(Jen the pQte·n tial calculated fr~m amperomet;rie data and that ealcnlle;ted ti-om Latimel''s data. A coniplete 1:eohec,lt ef ae.leulati0ns tben .appeared to be 1n ol'der . On page 171 .o:f Oxidation Potentials (12) a value E : •0,06 0 i s gi von for the half- coll, 2 Br-+ Ou z Ou.Br 2• + e ... ; this is on o:f the values which Meier used in cal culating a :reference cuprous bromide - cupric copper half-cell potential which he compared with his experir11ontal ly dete:rmtnad value. However, if one combines the following data which have been checked to the original sources given by Latimer (11,12) E 0 : •0o522 K : 1.2 X 10- 6 then a value E0 : - 0.173 is obtained for the, half- cell For the half-cell one t hen obtains the value E0 : - 0 . 52 which is considerablW dif:fel"ent from the -0. 64 with Which Maier eompat"ad his experi mental value of -0. 666 . conditions v.ras made . An examinati on of Meier's experimental Inasmuch as the value o:f the indicator potential was found to be of considerable signifiaanoa in the study of the iodine monobromida equilibrium , an investigation was undertaken to detel'mine the effect of indicator potential upon mi nimum indicator current . Experimental !?,eagents. A stock solution of 1 . 01 F Cuso 4 w.~s prepared from reagent grade CuSO •5H o. 4 2 A ten fold d.il u-tion o:f the stoc1t . solution wa s standa:rd:lz8i iodome·tl'ically as di i-act ad b7 ·swift (27). The sodium thiosulfate solution was standardized a• gainst dl'ied potassium iodate. Stock solutions of 1 . 00 F and 6.0 F NaB~ were prepared by wei ght from the reagent grade salt which had been dried at least 2 hours at 120° c. Reagent g~ada conoent~ated H SO, was diluted to give 3 F 2 4 acid. A)2l;?B,l;'a:tru;s. Except fol" the mi croammeter the apparatus was the same as that used in Part I - 2A. The microammeter was replaced by a shunted galvanometer as described in Part I - 3A. The sensitivity of the galvanometer was about 3 divisions pe~ microampere. Determination of ,the Proper Potential for Stuc1~ing the Cupric B?"omidQ Eq,uilibri'!m,• A solution containing 5 ml of 1 . 01 F OUSO4 f 5 ml of V"' F H2SO4 • and 7 .· 5 ml of 6 F NaBr was made up to a volume of 45 ml with water. With an applied indicator potential of 100 millivolts the indicator current was carefully adjusted to its minimum value. .l:he indicator potential was 1 then varied and oo:rl'esponding values of the indicator current wa re reco:rde d. A plot of indicator current vs indicator po- tential is presented in Figure 4 . Inspection of Figure 4 shows that an indioator potential of about 60 millivolts is most suitable for this equilibrium study. I I I .. 20 I I BO 100 I I 140 Jndioator Potential. (Mlll.ivolts) 40 60 120 Figure 4. Vaa,iation of Indieatoi- Oui.-rent with Potential Differenc:,e Betwettn Indicato:r Electrodes fo~ a Solution Adjusted to the Point of Minimum Indicator Ou~rent at 100 Jlilltvolte. Solution 0.112 Fin Ou.S04 and 1.00 'B' in NaBr. -55- Procedure fqr Obtaining Eguilibrium Data . In view of the above exper.i ment, the indicator :potential was set a.t 60 millivolts . Data for equilibrium calculations were taken in solutions 0 . 33 F in II SO 2 and with a total volume of 45 ml. 4 By means of pipets Cuso and NaB:r stock solutions were added to a titration cell 4 (See Tabl e VIII for experimental concentrations). B-romine was generated in the call , and values o:f indicator current ware :recorded after each second of generation. Polarity of the generating electrodes was then reversed , and generation was continued until the indicator current had passed through the minimum point . When excess cuprous coppe:r was being generated , values of the indicator eu~rent were recorded after each 2- 4 seconds of generation. Tha value of the minimum indicator current 11·1 as determined by passing very slowly back and forth through the minimum point until a reproduci ble reading was obtai ned. Calculation of Cupric Bromide Equilibrium Constant . Meier (17) has de~ived an expTession f or the equilibri um constant ~ ~ : (CuBr 2- ) 2 (B:r · ) , in terms o:f the minimum indicat or current 3 and the indicator sensitivities on each side of the minimum point. Inasmuch as that deriva.tion is quite similar to the one in Part I - 3A of this thesis , a detailed treatment v,111 not be presented here . The oalculatio11s are be,sed on the assumption that the indicatoI' current is a linear function of the cuJ;>rous copper and bromine concentrations: - 56- because at constant bromide concentration the ratio of bromine to trib:romide is a constant. The ouprous copper exiets pri ... marily as Ou.Br - at the bromide concent~ations used . 2 If Kl : (CuBr 2- ) 2 (B~ - ), 3 Maia~ has shown (17) that - 4(im)3 Kl - 2'1k, 2k - in which i m (A) 4 : minimum indic ator current, , V ; 0 . 45 liter Kg : rate of generation (1 . 039 x 10~8 equivs,len.t per second) and k4 ·c 1 + 17 (Br "'" >] - _(diLdt}V 17 (Br •) 2K g • K2 -.. If ( Cut t ) 2 (Bl' ... ) 7 t then • Values of K we:re oal,.oulated by substituting into equa1 tion {A). The minimum cu~rent was a direet experimental measurement . Values of k and k 1 4 ware obtained by gl~aphi oal estimat ion of the indicato-r cu.l'rent slope and substitution of that value into the pl'ope:r e-quation . In or dal" to calculate values of x: 2 it was neoessa'l'y to col'.'rect for the bromide com• plexes of cupric copp er. of K0 1 : From part Il of this thesis the value 2.1 waa taken :for the equilibrium, j-i' Cu + B~ - ; Cu.Br +- , As an approximation for the formati on of the dibromide complex, Cu.Br+- t" Br"" : CuBr 2 it was assumed that the constants for the bromide complexes would have the sama r ati o as the constants f or the chloride complexes. From the work of McConnell and Davidson (15) one finds a value of 5.65 f or the ratio of tha monochloride constant to the dichloride constant. By applying this ratio the appr oxi- mate value of Ken : 0.37 was obtained f or the dibromide equili"' briwn given above . Higher bromide complexes of copper ware as- sumed to ba negligible i n their effect on the cupric or bromide ion concentxations, In Table VIII both the formal and molal concentrations are listed for euprio and bromide ions. These molal concentrations were calculated by ac counting for the cupric bromide oomple xes described above. Values of K2 calcu- l ated fl'om the molal cupric and bromi de ion conoent l'ations a.re presented in Table VIII. For the equilibrium , values of K for various oup1:ic and bromide ion concentj'.'at ions will be found in Table VIII . F~om the a~erage value X: 1 . 32 x 10- 16 ·the val ue of .E0 : -. 468 is obtained for the cell :reaction given ab ova$ If the val ua E0 : • l.05 is taken (10} for the li..al f • eell , than one calculated E : • Oe58 0 f or tha half- call ; As menti oned ea:rliel' in this section t he data 1.n Lat i mer (11,12 ) yield a value of E 0 : - 0 . 52 fo:r this hal f - cell. The source of g~eatast erro~ appaa~s to be an iudicatQr cur:-rent whi ch my be given by the following electrode reacti ons: Indicator anode 3B!'• = Br 3• -r 2e- Indicato:r cathode 2cun+ 4Br· + 2e· = 2 OuBl' 2- • Inasmuch as indic at or currant due to this decomposition augments the i nd.icat or cmn:ent due to cu_p:rous copper and bromine , an in- determinate er~or is introduced because it appears t o be impossibl e to separat e the t wo currents . Of course t;he decomposition curTent vari es wi th the concentration of cupric copper and bromi de ; therefore , it is difficult to make an estimate of the effect~ Activity coefficients have not been taken in.to account in all these calculations. The agreement between this work am the literature would probably not be improved by considering activities . TABLE VII! Constant for Cuprie Bromide Oxidation - Reduction Equilibl'ium no. kl (Div/10- 6 F) k4 (Div/10• 7 F) (Di V' • )'·~ ( F) 1 5.10 7. 5 1,00 5 .. 45 9 .8 0.112 0 . 0561 0.404 1.00 3 .6 8.41 7 6.16 4 . 41 8.6 0.0561 0.0224 0.0561 0,0561 Oe70 6 5. 20 5. 45 6.75 6. 75 10.6 5 9.10 9. 95 5. 40 6.16 8.65 2 3 4 (Cu.++) (Br•) (M) (Mu 2 0.0318 0.0162 0 . 90 0,95 3 0 .143 4 6 0.0206 0.0059 0,0206 7 0.0152 0,69 0 . 66 o . 98 0.66 o.95 No . 1 5 Kl X 10 1 . 39 2. 42 10 . 7 0.306 0.366 0.226 5,29 i ·ID (Cu++) , 5.l 4.3 20 4 1.00 1.00 0.10 1 . 00 X 10 K X 1016 4.84 0,287 1.61 15. 3 0 . 232 le50 1.32 o.zo4 1,,20 0.232 o.974 3.28 1.32 K 2 1.61 A,r K : (Br · ) (F ) 2 ( Cv..Br 2 - ) ( B:r • ~ 3 2 (Cu ++) (Br _ )'I * l scale division : o.ss mi o:roa.mperes 0,699 4. Study of the Sucoessive Oxidation St ates of Vanadium During a coulometric titra tion of tripositive arsenic (20) no change of i.ndicator current is observed until the equ.i va.- lenoe point is reached . Soma substances such as iodine cause ' an indica toT cu:rl'ent during- the oouree. of the •ti t:ra tion .( sea Figu.l'e l) ~ Professor Swift ' suggested thst the oxidation.- of vanadium through seve!'a1. 01r.ida.tion states might give rise to a series of indicator currant peaks . A possible applicatipn of suoh phenomena would be the development of differential titrations 11 ~erimantal. Reagents . A solution of NaV0 ize d by Dale Ileie:r . (0 . 152 F.) and Na 2 3 had been pr.a pe.red and standard- The solution was said to contain Navo 3 oo3 (0 . 2 F ) .• This stock .vanadate solution was diluted exactly ten- fold . and sufficient HCl was included to neutralize the Na co and gi ve a f'inal acid concentration 2 3 of 0 . 1 F . The hydrochloric aoid ( 8 li1) described in Part I - 1 was used he:re . For the re ductor 30- mesh zinc was given a light coat of mel."cu:ry by :reduction of Hg.Cl • 2 A smull Jon.es 1·eductor wa s prepared with a zinc colmnn 8 cm high and ab out "'I mm in dia- mater . A;eparatus . The instr-ument descl'ibed i n .Part I - 1 was used for this i nv estigation,. A gene:ration r ate of 1 . 054 x 10· 7 equi - valent per second was used for all experiments . Use of Small Jones Reduoto~ and Obse~vetion of Indi cato~ _ _...._._ _ _ _ _._......,....,..;;_,..,.;...,,;,..;;....___...,_,...._,. • --a,t,. ~ - . . . -- • ~ - ·- • • rte , - - . ...- ~ . . . . . , Cur rents ~ l. A 1- ml portion o:f ·the dilute lfaVO was equivalent 3 t o about 150 seconds of generation fo:r each eleci:;ron change of tho vanadium. Ab out 30 ml o:f a s olut ion 0 .1 Fin HCl and 0 ,2 Fin NaCl we~e a dded to a titration cell and the cell was flushed out wi th nit:rogen , For all succeeding operations an at mosphere of nitrogen was maintained in the cal l , One millilitel' of the dilute Navo solution was pipetted i nto the raductor and was 3 al lowed to pass into the aell at the rate of one drop about ever y 10 ae conds o Several millilit3rs of solut ion (O .l F HCl and 0. 2 F NaCl) wer e passed through the reducto~ to wash out the sample . With the indicator potent ial set a t 200 millivolts the indicator cu1•rent was ve-.ey high and could onl y be followed by usi ng a mieroammeter f i tt ed with a var iable shunt . When generation of chlorine wa s started ~ the indicator current decreased steadily f or about 40 seconds; then. the indicator current remained at a low va lue (about 3 micro&mperes) for about 140 seconds while generation oontinuad e A:ftel' a total of 180 seconds of gene~at io n , t he i ndi cator current increased l:'a.p idly Hnd di d not again :retu.x,n to a low va1·11e • even after 330 seconds of gane~ation . Similar resul ts were obtained upon repet i tion of this procedure . Inasmuch as Meier , :Mye:t>s , and Swift ( 20) observed no indtcatol' current for hydrochloric acid solutions con-taining V +4 t5 and V , the above exparimants might be interp~eted as follows : Rou.ghly 3o~i of the vana d:lum was reduced to the plus ·three Btate and 40 se conds were requi1•ed to oxidize 4 that v +3 ·to v + • Durj_ng the next 140 seconds the v +4 was oxidized to v +5 • then a chlorine end point was obtained . Howeve:r , before it can be definitely . +4 stated that Vt 3 and V suppot1t an i ndicator current; it mus t ba shorrn that in the above expei-iments the indi cator ourrant is not caused by hydrogen fl-om the reducto:r o 2. To a titrati on cel l was added 35• 40 ml of 0. 3 F HCl , then the oell was flushed out with COP. . Afte!' passing "" solution t hrough the reductor , CO 2 was bubbled. through the solution and the indicatol' ou:r!'ent 1 ml o f the dilu't6 Navo 3 {potantia.l 300 millivolts) was observed but 110 chlorine was generated . The indicator our,:ent decreased slowly and l'E:mched 10 micn.~oamperes in about 3 minutes . The same ope~ations wexa performed with no vanadium in the solution going through the :i?aducto:r . The ini ijia,l current in the soluti on was a g:reat des.l l ess than with ,ranadium present . Upon :passing CO ,. , thl'ou gh the solution , the indicator ~ - 64- current decreased rapidly to 2 microamperes in about 20 seconds. 3. Experiment 2 was repeated with a vanadium sample and puri f ied nitrogen was use d i n place of co 2 • The nitrogen was bubble d through a gas wa shing bottle filled wit h 3 F NaOH saturated with Na s o : 2 2 4 oxygen. this process was supposed to remove After t h~ sample had been i n the cell f or f ive minutes, the indicator current bad decrea se d somewl1at but remained at a high ( off scale) value. In view of the results of experiments 2 and 3 it seems reasonable to conclude that an indicator oul"rent has been observed which is caused by v r 3 and v +4 • 4. In previous exp eri ments no evidence had been obtained regarding the presence of v~~. Because of the dilute solu- tions used it was not possible to i dentify the oxidation state by color . About 3 ml of the stock 0 . 152 F Navo HCl and placed in a small flask . were acidified with 3 The :flask was flushed with co 2 , then granular zinc was added and t he flask wes stoppered . After swirling the contents of the ·flask f or 20- 30 -seconds , the solution turned purple . To a cell containing 0 . 2 F HCl and f lushe d wi t h nitrogen 3 drops of the pur ple solution were transfe rre d . The i nitial i ndicat or current decreased as chlorine was generated but there was no region of low indicat ol" current (below 20 microamp eres) prior to the f i nal cu~rent rise . Also, there was no additional indicator currant maximum or minimum -65- which might be expected during the oxidation of v ++to v T3 • A partial explanation of the above experiment is that v•• was oxidized at the indicator anode while hydrogen was being produced at the indicator cathode . However, the presence of hydrogen does not seem to account for the absence of a low ' indicator current during the oxidation of Vt- 4 • Uso of Potentiometric IndicaM,llii...~Zstem. Fo:r these experi - ments 0.0152 F Navo solution was reduced by granular zinc in 3 a stoppered flask; the flask was filled with co • One -milliii~ 2 tar portions of the reduced solution were transferred to the titration cell when needed. A Beckman No. 270- 6 calomal electrode was used as a reference electrode. A shGrt length of platinum wire and t wo pieces of pla.tinum foil (1 cm 2 and 5 cm 2 ) were tried as electrodes in the ti tra·ted solution. The potential difference be- tween the calomel electrode and a platinum electrode in tha cell was measured by a Beckman millivoltmeter. In preliminary tests, stable potential readings ware obtaine d when iodide was oY~dized in a solution of sodium bromide. However, dur.i ng a titration o:f tripositive arsenic with chlorine, potential values.: fluctuated 20- 30 millivolts when readings were taken. Several experiments ware made by introducing reduced vanadium solution into a titration cell containing 0 . 2 F HOl. atmosphere of nitrogen was maintained in the cell. An Chlorine was generated fol' sho:rt intervals o:f' time and potential r .e adings were recorded . In no case was it pos sibl~ to get a sat of good potential values for a titration . dxi f ted o:r fluctuated . Tha potential values e"ithar No one of the three. platinum electrodes mentione d above improved the ~esults . A l a:r ge calomel half-cell was prepared and connected to the titration cell by means of an agar- agar salt bridge . The po- tential between this calomel half- cell and a platinum wire electrode in the cell was measured by means of a "Queenn potentiometer . Stable potential values could not be obtained with this arrangement . Conclusions -··---·-Inasmuch ,•as vanadium can be detel'mi.ned coulomet:rically by electrolytically generated cuprous copp er (20) ~ further investigation did not appear to be profitable at this time . Two conclusions can be drawn fl"om the exp eriments using the amperometric circuit . 1. In a hydro chloric acid solution tripositive and quadrapositive vanadium cause an indicator current. 2. If a goo d reducing agent such as dipositive vana- dium is placed i n a titrati on cell and the indicator circuit is closed, a significant amount of the reducing a gent may be oxidized by the indicator anoda while hydro gen is being produced at the indicator -0athode . For the coulometrio titration of such subst~ucas a special method Will be required. References 1~ Brown , R. A. and Swift; E. H., J . Am . Chem. Soc. , ll, 2711 (1949). 2. Cooke, W. O., Hazel , F., and MoNabb , W. M., Anal . Chem., 21, 1011 (1949) . 3. Farringt on, P. s. and Swift , E. H., Anal. (1950) . In P~ess . Chem., _g_g,, 4. Faull , J . H., J . Am. Chem . Soc., 66. 522 (1934) . 5. Forbes , G. s. and Faull, J e H., J . Am . Chem. Soc ., .£§., 1820 (1933). 6. Griffith, R. o., Mo.Keown , A., and Winn , A. G.; Trans . Fa~aday Soc., _fill, 101 (1932) . 8 . Kolthoff and Lingane , Polarograph¥ , Interscience ; 7. Helger , B., Chemical Abstra cts, 43, 5688 (1949) . New York , 1946 , P. 4~5- 458 . 9. Korenman , I . M., J . Gen. Chem. (USSR) 11, 1608 (1947) . 10. Latimer , The Oxidation States of the El.~mants , Prentice- Hall , New York , 1938. Page 53 . 11 . Ibid , Page 169 . 12. I bid , Paga 171. 1 3. I bid , -· Page 173 . 14. Lingane , J . J ., J . Ame Chem . Soc., .§1., 1916 (1945) . 15. McGonnell , He and Davi dson , N •' ibid , 72 (1950). In Press . 16. Mehlig . J . P o and Shepherd , M. J •• Anal . Cher.a., fil:., 17. Meie:r, D. J •• M. S. Thesis , C,I.T . (1948) • 18. I bid , Page 38. 19, Meier p D. J .~ Myers , R. J., and Swift, E. H., J . Am. Chem . Soc., 11, 2340 (1949) . - 64 2 (1949). Pa.rt IV . Referenaas {Continued ) - 20. Myers, H. J . and Swift, E. H., ibid., 70 ., 1047 .{1948) • 21. Ramsey , W, J . 1 Farrington, P. s., and Swift, E. H., Anal . ·Chem., 22 , 332 (1950) . 22 , Ray , P., and Sarkar, P. v., J . Chem. Soc., ill, 1449 (1922) . 23. Saase , J . W., Niemann , C., and Swift , E. H., Anal. Chem.; ll, 1~7 (1947 l " 24 . Someya , 25 . Swift , E. H., J • Am. Chem. Soo ., .§1., 2682 (1929) • 26. Swift , E. H., 4.._Syst8m of Chemical An~ysis , Prent iceHall , New York, 1939 . Page 69 . 27. Ibid , Page 245. 28. Szebelledy, L . and Somogyi .• z., x., z. Amorg. Algem. Chem ., ~ . 368 (1926) . 313 (1938) • 29. S zebelledy , L . and Somogyi , an d 400 ( 1 938) . z. Anal . Chem ., 112, z., ibid .; ill, 385 , 391 ; 30 . Wooster, w. s ., M. s. Thesis , C. I.T . (1947 ). 31 . Ibid , Paga 10. 32 . Ibid , Page 11. 33 . I bi d , Page 12 34 . Wooste:r, w. s ., Farrington, P. -- Anal . Chem., 21, 1457 (1949)0 • s ., and Swift , E. H., II o SJ?ECTROPHOTOMETlU C I MVESTI GATI ON OF THE COPPER (II) MONOBROMO C01\i1PLEX In Par~ I ~3B . of this thesis ~he e.alcula tion~ r~qutred in:foTmation. conoe:rnh1g the bromide complexes of cupric copper. For many years it has bean knovm that oupric copper in concentrated bromide solutions (above 3.5 formal HBr) produces a ?'addish- brown to p-q.rple color; hqweval",. no one seems to have published an equtl ibrium constant for an~ of the bromtda compl~xes ~ MoOonnell and Davi dson (1 ) have dete:rmi~ed equilibrium constants for both the monoohloro and dichloro complexes of cu1;ric copper : the speot:roph.otometrie method which they used is , . rel atively v..ncqmplioatad fo-r · detarm:'-nation of a monohalide complex and :lt was adopted for thi~ i:nvestigntion ., All solu- tions used for these measui-ements were adjusteo. to an ionic strength of tmity. ~,!:!lrimental Reager;ts . A solution of Cu( 010 ) had been prepared by Harden 4·2 McConnell by dissolving Cuco •cu(OH) in HCl0 : the excess 3 4 2 HC10 was reported to be very small . A solution of Na s o 4 2 2 3 was prepai-ed and standardized against Kio • The Cu(Cl0 ) 3 4 2 solution vi, as then standa1·0.ized. iodomatrioally aa directed by Swift (2) 0 and the copp er ooncentra.tion was found to be le342 F e A stooJr solution of NaBr (0.100 F) was prepared by dissolving a weighed port ion of t .he :reagent grade salt in c~1loride- -.70- free water and diluting to the mark in: a volumetric flask . NaBr had been d~ied for 1 . 5 hours at 120° c . The Solutions o . 0050 F and 0 . 0250 F in NnBr were prepared by aceu:ra:te dilution of the stook solution. A standard solution of HC10 diluting 60% acid . (2 . 673 F) was prepared by 4 Standardization of tha acid was accomplished by the iodate method outlined by Swift {3) . A;e;earatu~ . A Backman Model DU Spectrophotomete1• wS. th hydrogen lamp wa s used for measurements of optical density. The rectan- gular :right p:rism quartz cells were of 1 . 00 em path length . l o Portions of the above solutions wel'e pi petted into a 50 ml volumetric flas~ so that the final solution was 0.250 F in Ou(c10 4 ) , 0.00100 Fin NaBr , and 0 . 249 Fin HOlo • Before 4 2 the addition of Na.Br • the vol urnatr1e :fl.ask ws,Pt flus hed ou.t with co 2 • After the solution in the flask was thoroughly mixed , a portion of the soluti on was transfer~ed to a sp ectrophotometer cell and a cell cover was placed in position with a thin film of gl'ease to exclude ab.• . A blank solution 0 . 249 F in H010 was used to balance the 4 spectrophotometer at each wave length . A copper oomparison solution was made up t o be 0 . 250 F :i.n Cu.(C10 ) end 0 0249 F in 4 2 HClO • 4 At a given wave length U1a spectrophotometer was balanced - 71- at ze:ro opti.cal density When a blank solu1; ion was in the ,l ight The copper comparison solution ii-Vas then placed in ._the path . light path and the optical density was dot·e:rmine d. Immediately; the copper solution containing bromide was placed in the light path and the optical density was measured . Thirt$en wave lengths from 255 to 300 millimicrons were used. The wave length intervals vm:ie d from l to 10 millimiorons depending upon tho change in optical density. Below 260 millimicrons the optical dens:i t :l:es of both cop- per solutions were high (above 0 . 6) and the diffe~enoe in optical dens ities was not large enough to pe~mit accurate calculations ( see Discussion), The aoppe:r complex appea~ed. to absorb most strongly about 280 millimicrons; however , the optioal density of the coppo11 complex solution in that region was too low {about 0 . 1) for accurate wo~k. 2. A solution 0. 250 Fin Cu(Ol04 ) , o.005 F in NaBr , and 2 0 0245 Fin HC104 was pl'epar ed as in ex.periment 1 . Opt:lcal den- sities were measured (as in exp . 1) at evel'y 5- rnillimicron intervsl from 260 to 300 millimicrons . The :results of this experi- ment indicated that 0 ., 005 If NaB:fl was a suitable conoentl:'ation for subsequent experiments . 3. The optical 4ensity of a solution 0.485 F in HCl0 4 and 0 . 005 Fin NaBr was measured using 0 . 485 F. HC10 4 as the blank sol ution . Over the range 270 to 295 millimio1·ons the optical density of the bromide solution varied from 0 . 002 to 0 . 004 . Proced~. All solutions ware prepared by pipetting standard solutions into 50 ml volumet?'ic flaslts a.nd di.luting to the li1lasks to contai n b::eomide wG re flushed wi·th eo marlrs. 2 just before the a ddi t:i. o:n of br omide s olution. :!.1he Cu( 010 ) con4 2 centretions were 0.250 F , 0 . 210 F , 0 . 170 F , 0.135 F . and 0.100 F, The HaB!' concentration was always 0 4005 F, and suf- f'ieient HClO was added i n ee,ch case to b!' i ng t he ionic 4 sJe;re11gth to unity . !!010 4 &oh blank solution aonta:i.:ned the amount of used i n the corresponding copper complex solution: each copper comparison solution contained the same amount of Cu(c104 ) 2 and H010 as the corresponding coppe~ complex solution , 4 The operation of transfe~ring the copper complex solution to a spectxophotomete:r cell was pe1•:formed in a lar ge beaker into which & stream of CO 2 we,s pa,{;1sed .. Thai; cell · was then clos ed with a Tubbe:r stopper whi oh had been trimme (l to fit the squai·e top of t he cell. Tha spectrophotometer was balanced w;ith · the bl ank solution at each wavelength • ·then the optical donsi tios o:f a coppel' com... plex solution and the corresponding aoppel' compa:rison sol ution we!'e mes.sured ~ Readings W'Ol'e taken at 5- millim.im:on intervals f'.rom 260 to 300 millimiorons . To check possible d'.ri fting of the readines , optical densities of each set of solutions were remaasu:red at several wavelengths abou·t 15 mi11utes afta:r the fi:rst determin.ati on . For solutions of the same ionio strength it i s a reasonably good assumption that tha activity coefficients of the ions of :tnt:e:rest are constant even though the composi t1on of the solutions is va:r:tad . As shovm by preliminat<y experiment 1 ; ve-:r,y dilute solutions cannot be used because tha:re wou.l d be insuf .... :f.icient complex to give easily measurable e,bsor,ption e In consio.eration of t hese facts a const ant ioni.c st:t•ength of unity was chosen foi· this invosti ge.tion . In the aquilibrimn , let ( CuB:r+) : x , then {cu ++ ) = a~ x and {Br•): b - x . Fo~ the equilibrium constant , K, one obtains the relation If a >> b and a >> x ; as a first approximation x '6- x =Ka 01' (1) By definition , and ul .,.- E1 X T\ in which D and D ai·e the optical densities of solutions (a ) and 0 1 (x) formal in cu++ and CuBr + respectively and with a total ionic str ength of 1 . 00 . The quantit ies E0 and E1 are the mola~ extinction coefficients oi cu+-+ and Ouih.· +- -respectively . is the optical densi ty of a s ol ution (a ) form.al in oo:ppe:r per• chlorate , ( b ) :fo:rmal in sodium broinid.a and with suf':fioient pe::rohlo:ric acid to bring the :tonic st? ength ta unity , :fu.?1ther usa a >> x of tho fa ot that Do "" 1 - yields the app:roxiroation E1 X +E O • o(;ll,' then E1, x : D0 l - D0 Subs tituting in e quat ion (l) . --~- Dl _ Ifs E. l b 1 + Ka __.... and Dol . Dl _ , ,.. _.. ,. b • 1 + Xa Inversion of ·this equation yields tho relation, (2) If CuBr + is the only complex fo~mad in signifi cant quantity, the11 a plot o:f b -- against 1 should be a st:ra:i.ght line $ When D -D 01 0 a the expe:rimental data \,"/ Ore plotted on su oh a g..eaph , the points were somewhat scattered but no smooth aurve othel' than a st:raight line eould be pa 1:rned thl' ough ·the :point s .. In equation (2) it can be seen that the equilibrium constant , x. is obtained f?iom the ~atio of intercept to s ope . Instead of evaluating K g'.t'aph.:cally. the method of least squares was used to co1•1~ela.t;e t· e data~ Although optical dcm- si ties had bean measured at n1.ne wave lengths .• the data used for tba calculation of K was that which was t aken in the region of maximum abso~ption by CuBr t- • After obtaining values of K, a check was made on the change of Ou ++ concentration by complex f ormation : the nmnerical appro::dma tions rnade in del'i ving equa- tion {2) are within exoa:rimental erl'ors~ Values of K caleulatecl f rom data at five wave lengths al'a presented in Table IX, By means of standa.l'd formulas t he pro- bable el.'r or was calcuJ.ated for each of the five slopes and in- tercepts . From an avel'age value of the slopes O inJGeroepts , and probabl e errors in those values , a probable error was calculated for the avel:'age value o:f .K : this probable erl'.'o:r.> was found to Fo:r each of the f:lve solutions used in this investi.gation (x), the concent ration of Ou.Bl' + , was calculated using the average value of K, From t he approximation mentioned above , then Values of the mol ar ext;inction coefficiont of OuBr + , E1 , ware calculated :for oaoh of the five s olut ione . and the ave:re.ge E1 was found for each of the nine wave lengt hs . Figure 5 shows the curve obtained by plotting log E against 10 value of 1 0 wave langth. S1.umna:tt Y: An equilibrium const ant for the formation of the monobromo complex of copper (II) in solutions of unit ionic strength has bean determined by a speot:rophotometrio method . At 22 ± 2° C the average value for that equi].ib:ri~m constant was found to be 2 .1 .:t 0.25. The molar oxtinotion coefficient of Ou.Br+, ~\ , been de• termined over the r ange 260 to 300 mil l imicrons , and l'eported graphically as a f unct ion of wave length. E1 is TABLE IX Equilibrium Constant ·"or tha Copper {I I } I,tonobromo Complex in Solutions of Unit Ionic Strength t-'t Wave Leng'Gh (Mil limio:ron s) (lite;/mola) 275 2 . 06 'fl ' 280 285 290 295 Av . 2. 08 References 1, McConnell, H. and Davidson , N., J. Am. Chem ., Soo., 72 0 2. Swi :ft , E. (1950) In Press . n., A System of Che,mica:i 11.,n alys i~, P:r:·enti oe- Hall. New York (1938). Page 245 . . a,s a,o 275 aao sss . 2,0 2t5 Wevt Len8'h (W.lllrnicl'one) Figure . s • . Mola'it lktinction Ooeff1c1.itnt of OuJt + tn Solutions ot Unit Ionic st,:engbh. IIL, Detel'.'mination o.f Carbon by Wet Combus'tton * Equipment recommended fo-:e use by a Chemical Va:rfa:re Sa?'vioa Mobile Unit was subject to the following special requirements: a),. The appa: catus should re.quire a minimum of space when · packed and w·hen assembl ed fo:r. use. b). \r'i'hen packed , the apparatus must withstand heavy shocks in transport~ The Van Slyke manomatric apparatus was designed to u.se the Van Slyke~Folch (4) combusti on mixtu:re for the determination of carbon in organie compounds . However, the Van Slyke apparatus is bullcy and fragile and does not meet the above re quirements . It a1'31,0aTed pos sib.:!.e to desig.a a. less expensive appa:ratus which could be assembled easily in a smal 1 space " Wit;h these object- ives in view this investigation was undertalrnn . Bl'iefiy, the proceduxe is as follows: A sample of 10 to 12 mg is heated with an anhyd:rous sulfu:r:ic; phosphoric, ohxomic , and iodic acid mixture in a stream of carb on dioxide - free air ~ and the carbon dimcide evolved is absorbed in sod:bun hydroxide . The carbonate thus fo:rm ed is pre cipitat ed as barium carbonate and determined acidimo"iii."tcally . * Tho Van Slyl~o- Fol oh combustion Based up on work done f or the Office of Scientific Research a nd Development under Contract 0EMsr- 325 with the Cs,li for nia I nst itute of Te chnology , ... 31 .... solution ha s been used in other ptooeduras (2,3) , but in these the carbon dioxide has been determined gra.vimetr:Lcally, Reagents .~ l?reparation of the combustion solution. Pou4." 60 ml of 307& fuming H2so in·to a flauk contain:i. ng 40 ml of 85% phos4 pho:d.c acid and add. 10 gm of C!.'0 3 and l gm of Kio • 3 Heat to 140-150° C and swirl or stir for 1- 2 m:l nutes •.. Cool and store in small glass- stoppel'ed bottles. As an altern tive to using fu.ming sulfuric acid, use 20 gm of P. o , 85 m1 of 95% H2so , 4 2 5 and 15 ml of 85% H PO with tha stated quantities of C:r:0 3 and 3 4 KIO 3 • Sodium Hydroxide {0.5 F) . wat er O Dissolve 5 gm of NaOH ;in 20 ml of add l ml of 1 F BaC\~ solution and c ent rifuge . Dilute the cent-rifugate to 250 ml and store in a siphon bottle pro tected by a soda lime tube . Barium Chlo:ride Reagent , l F in BaCl Pi•el imina1\y ;E;!cperi..m~l1.:tJ?.• 2 and 0 . 001 F in H<n . The following seventeen exper;iments are presented to on.tline the developments leading to the final pl.'ocadu:re and the appa:ratius shown in :b"i.gu:res 6 a:nd 7 , ( page 91 ) . 1. Tho 'first rather ol'u.de apparatus consiste d of a dryi ng toweJ~ , f illed with pellets of NaOH, and two 15-ml centrifuge tubes com1ected in series as are tubas A• in: . and G in ]figure 6 . ( Page 91) . The tuba oo:r:responding to combustion ·tube M was closed with a rubber stopper , the CO 2 was admitted into ab- sorption tube G by means of glass ·fmbing d:rawn to a fine tip , and a solution of' Ba(OH) ' 2 was used i n tuba G. A sampl e of 3 , 5 dini t:robenzoie a ci d was v1e i ghed int o the combus-hion tube . By means of a wa.t e1· aspiratol:' air was dr awn ·l:ih:t:•ough the e:pparatu.s for seve:r.al minutes , ·CO by the NaOH in the d~ying ..,cowe:i.• o be1.ng abso:rbed 2 To the combustion tube ware added 2 ml of comb1.tst:i.on s olution , and 2 ml o:f ldutul:'ated Ba( OH) and 3 ml of water wa!'e added to the abso:rpt ton tube . 2 While a stream of ail' was drawn through the e,pparatus, the combust i on solu.t:i.on w·as heated to boiltng for 6·•8 minutes . The absorpti on tube was then disconnected~ and the naco pl.'.'ecipitat e was 3 washed with t wo 3- ml 1;,o:r tions of water then t itrated with 0 .0462 F rIOl to a ma·thyl orange end point in hot s ol ution. The result was abou t 25% high and indicPtod that ·the p:rec:lpi tate should be vita.shed mot"e ef:ficien·tly ~ A tendency fo r solid Baco 3 to alo g the ~Gi.p o:f ·the i nlet tube in tbe absorption chamber was t:roublesome . ~• :J!;xpe:r iment 1 was i•epaated. t o the washing of the pi- e- cipj_tat0 . After the precipitate was drain0d of excess Ba (OH) 2 solution , 1 ml of water and 1 d-rop of 0 .1~ phenolphthalein ind:i.c ato:r were ao.ded to the precipitate. Very dilute HCl (about O. O4 F) wat1 added dro:pwise with stil'r:tng u.ntil the solution was just decol o:rizod . i faint pink colm: on the preei1li tate in• dicated that the acid was not present in excess , The mixture was centri:fttged , and the preci pi ta-ts 1tras washed with 2 ml of wat er . The ca:{bonate was ti t:r.ated to a. methyl orange 011.d point i n hot solution with 0.1932 F HCl . The result was ab-out 30% low 5:.ndic·ating that absor:pt:J.on was incomplete; perhaps the combustion :progressed too :rap:i.dlyc of the solid BaCO 3 Some was ca ?ried up on bubbles and deposited on the upper wall of the tuba end on the stopper . Afte'.l'.' was hing the p:rec:t~pttate and oentr.ifuging., the centrifugates were turbid; a small loss here bu.·b not enough to account fo:r the large eJ::ror. 3. Repeated experiment 2 wit l1 the following chan ges: a. The Ba(OH) b. was -replaced with o.5 F MaOH. 2 The oombusticYtl solution was heat ed slowly . c. One mil 1ili to:r of l l!" Bac1 2 reagent was added to the absol'ption tube to precipitate the 013rbonate . d. Afte~ the base adsorbed on the preci pit ate was neutralized with dilute HCl as described in 01::par:i.ment 2, 2 dl"ops of 2% B:raunsol wetting agent we re ad.dad before centrifuging. The results of three runs were 97 P 101, and 94% of the amotmt of carbon thought to be present. The above modification a.ppearea_ i;o correct the major difficml ties ; however O iihere is a tendency towa rd low t'es-u.lts .- 4. The drying tOWGl' filled Wi.th NaOi·I pallets was replaced by t wo bubbla:rs made from ta ... 1. tubes; the f irst was fi ll ed with 50% NaOR ancl the second Vii th 0 . 5 F Na.OH. In. earlier exp eriments combustion tube M had been an ordinary 15-ml centrifuge tube closed by a :ttub'bar stopper . A spaoial head was made on a male standar d tape:t< gl:' ound- joint ( 14/20) 1 inlet and ou.tle·t tubes - 84• ware pl'ov:i.ded in the sides of the head for gas f l ow • and an ·inlet ·tube fol' intl?odu.oing combustion solution was see.le d in the top. The female 14/20 jo:i.nt was sealed on a 1 5- ml centi-i.:fuga tube and this combustion tube {M) was fitted to the head , using H PO as a l ub:rioant and ai:r seal , 3 4 ceclure we re as follov,s : ao Modifioa:tions of the p:ro- After- most of the excess base on the p:rec:i,pi tate had been neu-t1;aliz ed and the mixtuxe had been centrifuged a.s described in expe:riment 2 , ·the remainiri·g ads orbed base was neut:ta.l izad in s i mil a:r fashion instead of washing wi th wa.t e:r . b. The carbonr;1.te was determined by addi ng an excess of standar d HCl an d backt i trat i ng W'i th standal'd waoH. The ave:rage of several :runs was still low by ab out 2%. Blank: runs avel'aged 0 . 016 mg of carbon Which was ab out 0 . 5% for most of the samples used c 5. Rate of flow of the gas st~eem t hrough the absorpt ion tube G was val'ied :i.n a se:ries of three experiments. The results we re consistently low indica'ting that the :r ecovery of co 2 in the ahsol'pti on tube mi ght not have been 100% af'ficient . 6. __Anhydrous Na oo was p repared by heating NaHC0 st 2703 2 3 300° C to constant ~-vei ght. Three samples of the Na 2oo 3 were weighed out and t:rea.ted by the pl'ooedure of expel"iment 4; H so 2 4 was used to decompose the oal'bonate. The avera ge result was 97% of the calculated value giving additional evidence that th~ eo 7. was ltOt completely- ebS()l."bed. 2 Thtil p:reeip1ta:tion and titration pl!'ooedul'es wer$ cheeked with weighed samples of tbe freshly p:re,pa1red Na co 3 (expe~lment 2 6). Ea.oh eai-b&nate sample was dissolved in a small amount of watel', then o.5 F Ne.OR was added and the pi'eeipi tat1.on. washing, and titl"a.tion we:tit per'fo~nied as tn ex_pe:rim~nt 4. The ;results of thl'ee tests avei-ag€ld about 0.5~ high eh.owing that the~o was n& loss .i nvelv-ed in the above opei,ations. a. A second absoiipti.on tube was placed in sezeies with the usual abao:r:be3:. The only alt$l!&t1on of proeedu:re was to add 5 d:rops of l F BaCl 2 to 1Ib.e fittst washing of the Be.Oo 3 p~eeipi tate and 3 di-ops to the second washing. One sample of Na oo 3 was tl.'fa&,ted as in e~e:riment 6 asd the 2 result was 98.3% of the. ee.lculated valucu 11e Baoo was 0btainE!Jd 3 in the second absol'.,fption tube. ~Ol' a second sample of le. oo 2 the ~esult wss only 94.7% of the esloul'a ted value: 3 a sign1fi• eant quantity of Bac.o was obtained in the second tube and the 3 preeipitate was estimated to be about 0.1 mg of ea:rbon. Obviously e. mo:re effteient absO:fption system wae needed. It waa dest:rabl~ to absoi-b the 002 in a oent:rifu.ge tube so that precipitation and eent3rifugation oould he effecte.d witb a mini-. mum of expesutte tG> the air. No sinte,,:ed glass units wo\tld fit into a 15-ml eentl'ifu.gtil tu.be; howevei,. a. small po,:oue dispereing unit beeame e.va1lable eemmeite1ally and one of those was ea-dered. 9. While the disperai:ng unit was on ordet, a diffel'4'nt type of appa:ratus was tried. Two 15.-ml centrifuge tubes, fitte.d - 86- with ground glass joints, were connected together by means of a glass tube which car~i ed two male ground- joints . Sealed into the bridge directly above ea.eh of the tubes i.vas an inlet tuba • each of Which was closed by a stopcock, A sample was weighed into one of the tubes and the joint of that tube was lubricated with syrupy H3Po4 • Through one of the inlet tubes the assembly was evacuated with a water aspi?ator. About 5 ml of 0 . 5 F NaOH were admitted to the empty tube, and 5 ml of combustion solution was admitted to the tuba oontaining the sample. The combustion solution was then heated to decompose the sample, and the assembly was shaken to aid in absorbing the CO 2 into the base ,. Precipitation and determination of the carbonate was done as in experiment a. Many expe:eiments were performed and several minor va:ria- tions of the apparatus were used . sults were low. The axoess CrO 3 In almost all cases the redecomposes at el evated temp- aratures producing oxygen which destroys the vaouum in the apparatus and hinders the transfer o:f co to the base ., For 2 this reason these experiments ware discontinued and further detail is omitted from this report. 10. A dispersing unit , stated to be a 50-mioron disperser, was obtained f:r om Wilkens- Anderson Co . The unit was constructed fr om a 2. 5- cm length of 6- mm o.d., 4- mrn. i . d. Al:frax tubing (Csr borundum Co.) which was cemented to a section of 4-mm glass tubing; the opposite and of the porous tube was closed with a white cement .. Thi s di sparser served as the gas inlet in -87- absorbing tube G. The only changes o:f pTooedure were the following : a. Because of the large area involved, the inside of the disperser was rinsed as well as the outside. The washings were added to the solution in tuba G. b. After the addition of the Bac1 to precipitate the 2 carbonate~ the 15-ml centrifuge tube (G) was stoppered with a rubber stopper fitted with a short section of glass capillary tubing. The tube was then placed in a bath of boiling water for 2~3 minutes to allow coagulation of the precipitatao Eight samples of a dipic acid (recrystallized) were tested and the results were consistently l to 2% high. The absorption of co 2 was evidently satisfactory but the efficiency of the scrubbers for incoming ai~ was questionable . Heating t he Baoo 3 precipitate before centrifugation was definitely an improvement and the centrif'u.gates were clear. 11. To serve as a compact and efficient CO 2 scrubber, a K~aissl tube packed with Asoarite was placed in the position shown in Figure 6 (page 91) . After the combustion tube :M had baan heated sufficiently long (5- 6 minutes) to decompose the sample and the excess Cl'o ; the oombu.iption tu.be was neai-ly fil 3 led with conoantrated H so • w1i~ less d9ed space in tube Ma 2 4 shorter period of air flow would be necessary to sweep over all the CO • 2 Four analyses of adipio acid gave results less than 1% ( about 0.05 mg of_ carb on) high. Baco 3 A blank run produced no visible prec_ipitate in the absorption tube. 12,. Inasmuch as the Ba co 3 precipitate ;was washed for the last time in a solution which had baen adjusted to the phenolphthalein transition point,. the final back titration with standard NaQH was made to the phenolphthalein end point for several samples which had been titrBte d to the methyl o~ange end point. The average difference in the t wo titrationa amounted to ab out 0.01 mg of carbon. 13 • .A lO•ml portion of water which ha d been used fol' wash- ing the dispersers and the walls of the centrifuge tube wa s tested with BaCl and a p raoi pit a te wa s obta ine d. The preci pi2 tate was estimated to correspond to 0.01 mg of carbon; thei-efore, freshly bo iled distille:d wa ter should be used for washing. 14 . When several milligrams of NaF weTe placed in the combustion tub e with a sample. the disperser became clogged during the process of running a determination. Small pieces of porous porcelain were :placed in the rubb er tubing conneotin_g the combustion and absorption tubes; however, after samples containing fluoride were run, etching was visible on the upper portion of the disperser inlet tube. With a plug of moist glass wool in the connecting tuba, no signs of IDi1 attack ware visible beyond the glass wool . Four samples of sodium oxal at e with added NaF were analyzed. One pair of samples was tested when 0.05 F HCl was placed on the glass wool, and the results were 0.05 to 0.11 mg o~ carbon high . The other two samples· were analyzed with distilled water on the glass wool and the results were only 0.02 mg o:f carbon high: similar results were obtaine d with samples containing no fluor ide and with distil led water on the glass wool~ 15. The cement on the dispersing unit deteriorated rapidly because of the basic solution in which it was immersed and it was not possible to make a glass s eal between the Al:f:rax and glass tubing. After a lengthy correspondence with the Oar.,. borundum Company, a sample of :porous Zircofrax tubing was obtained . This tubing was 6- mm o . a. and 4-mm 1.d. and coul d be sealed to Pyrex glass . A small dis:persing unit was then con• structed from a 3•mm length o:f Zir cofrax tubing which was seale d to a 3•mm Pyrex inlet tube and the opposite end was olosad uvi th a glass plug.. This disperser was very satisfa c• tory in all respects. 16. The vacuum siphon shown in Jligure 7 {page 91) was developed to minimize ab sor ption of CO was removed from the Baco while the excess NaOH 2 pre cipitate . Tests showed that 3 small positive errors were not materially reduced by the use of the vacuum siphon; however, its use should eliminate posit ive errors which might arise if the time Tequired to make the determination wore unduly prolonged ·• 17. The · oombustion-tuba head D was designed to have the minimum number of ring seals: seals. earlier models required 3 ring Final Procec1u.re The procedure given below and the apparatus shown in Figures 6 and 7 (page 91) are the results of the experiments described above and some observations which were not included there. Assetmbl;y . of th;e ,Appara~u~~ .Figure 6 (pa ge 91). Assemble the apparatus as shown in Clean, dry and reserve a gl ass- stoppered 15-ml centrifuge tube ( Tuba M) ( Note l}. PaQk the K:raissl I absorption tube A on both sides with 20- 30 mesh Ascarite or soda lime (!t ote 2). Add sufficient concentrated H so to the 2 4 bub'bla counter B to immerse the inlet tuba about 2 mm. Connect micro stopcock C (Note 3) to tube A and head D with sho:rt l engths of olee.n, dry rubber tubing; leave stopcock C in the open position . Lubrioate stopcock E with eonoent~ated H3Po4 and leave it in the closed position~ (Note 4), Connect tube F to the outlet of head D with a 2 to 3-inch length of rubber tubing. Fill tube F with glass wool moistened with 3- 4 drops of distilled water (Note 5) . Bora three holes in a No. l ~ubber stoppe~. Through one hole, pass a short length of 4- mm glass tubing which has a 1 - mm tip . Con..nect this tube to the delive~y tube of the stoTaga bottle H (which contains sodium hydroxide-barium chloride reagent» 0~5 F in NaOH and 0.004 F in BaCl 2 ), and attach pinch clamp J to the connecting tube. By a pplying a sli ght pressure through the soda- lime tube of bottle H, foroe solution from H until it fills the delivery tube, then close clamp J., Moisten ' , f ·-- -+TO VAC U UM .i TO VACUUM+- the gla ss t ubing which cal'l'ias the disperser K (Not e 6) Wi th a dro p of glyoel'ine and insert_ the tubing in the rubber s·topper (Note 7) ; Qonn ect the disperser inlet tu.be to tube :B' with a 4 to 5- inoh length 0f rubber tubing. Inser t in the ·third hole of the stopper a 4• mm glass outlet tube and connect it to mioro stopcock L (Note 3) with clean, dri rubber tubing (Mote 8). Pass the other arm of the stopcock (L) through a No• 3 rubber stopper , then fit the stopper i nt o a 125-ml suction flask and connect the side aim of the suction flask to a water aspirator . Just before attaching the 15-ml centrifuge tube (G) to tho 3hole stopp er assembly , open clamp J and allow 2 to 3, ·ml of the NaOH-BaCl reagent to flush out the connecting tube. 2 Finally, assemble the vacuum siphon (Figure 7, page 9l) e Place tube R (Note 9) and s p ecial soda-lime tube (I11ote 10) in a two - hole No. 1 rubber stopper. Oonnact tube R to the 50-ml flask with about 2 feet of rubbar tubing, When the vacuum siphon is used. disconne ct tho rubber tubing from the outlet of tube G and connect that tubing to the outlet of the 50-ml flask. Notes . 1. The tube , Which receives the sample, s hould be cleaned with hot Cro 3- H so solution p :rinsed, dried, and protected from 2 4 dus t by i nverting i n a clean beaker or by stoppe:ring with a glass s to pper. 2. Pl ace a plug of gl ass wool in each compartment of the Ki~a issl tube to protect the inlet and outlet tubas . Another plug 0f gl ass wool should be pl aced between the 1:illing and the stoppel" . 3. Stopcocks C and L shoul d be lubri cated with a thin :film of greas e, 4, '.fhe s t opcock should t ut1n easily. Do not :fo:rce the stop- per as it may look! 5, If :tt; is known that the sample does not eo:q.ts,i n fluo rine, tube F may be omitted from the apparatus ,. This tube is insel'..!ted to remove HF whi oh would attack the inside of the disperser an4 so clog the pores that the di sperser oould not be f lushed out . A clean tube and filling should be used for each run be- cause the aci d produced by the absorption of so 3 lowers the efficiency of HF abso?:ption, 6. The disperser can be ma de by closing, wit h a pi ece of gl ass , one end of a sho r t length o:f Zircofrax tubing, 12- 7P. s9; 1/4 x 1/8 , and joining the othe~ end to a lengt h of standard wall Py-~ex tubing. 79 The joints are glass . Since the di spersel::' is moved i n the st opper du.ring each Tun , the i nl ~t tube should be kept movable i n the stoppare Tm dispe:rser should be ttinsed out with dilute HCl end water and shoul d be drie d before being used . 8. Solid or l i quid Which nay have been deposited in the connect ing tube by spray from previous runs should be washed out a.nd the -tu.be should be dr ied. 9o Tube R is me,da fl'om 4- mm tubing dravm to a 1 .. mm tip a t one end . Allow a 15- cm l ength from the tip , than make a right e.ngl a bend Which will s,id i n moving the tube in the stoppe1' , 10 . Spacial soda- lime tube S consists of a 3• i neh length of 10--mm t _ubing at tache d to 2 i n ches of' 4• mm tu.bin g bent as shown i n Figure 7 ( page 91} e The 4- mm t-u.bing ca:n. be sealeo. to 10- mm tubing or a s eotion of the 10-mm tubing oan be collapsed a.nd drawn to 4- mm dis.met er ~ Place a plug of glass wool at the narrowed and of t he 10- mm section a.nd fill the tub e with Asoarite or soda lime. Cap the tube with a plug of glass wool and a cork which ha s been grooved on the side wi th a file . !_rocedure f._Q_r th~ . .Anal:y~i s ~:f ..§.21.id Substance a Weighing the SamQl~, Oa:r efv.lly wei gh (Note 11) a 10 t o 12-mg sample of the substanc e :for analysis into a clean , dry 15- ml glass- stoppered oent~i fuge tube (M). Moisten the g:round s urfa ce on head D w•i th concentrated H Po ; then attach tube M0 turning the tube until a fi rm seal 3 4 is obtained . Fl ushing a nd Fill,ing the Appa:ratus. Tu:rn on the water aspi1~a-- tor and slowly open stopcock L until the rate of a ir flow ·through bubble counter B is 3... 4 bubbles pell' second . Aftez, 4-5 minutes , carefully open clamp J and admit NaOH to tube G t o wi thin 3 to 4 cm of the top of the tube. Do not admit the NaOH so rapidly as to cause liqui d to baok up into the disperser tube. Pour 5 to 6 ml of co mbustion solution into cup N, and ad just the :rate of ,air flow to 1 to 2 bubbles par second. Close stopcock O (Note 12) and ~edi a.tal:z op en stop coclt E • . When most of .the liquid has ente:reo. tuba M close stopcock E (No te 13) and immedia- tell open stopcock o• .Q.Q.!nbustion of the Sapmle and Abso_ t >ption of C_axb,on Dioxide, By means of stopcock L ad just tha r ate of air flow thTough counter B {Mote 14) to 3 to 4 bubble s p er second,. \"'latch the &,ction of bubble counter B ( Mote 15) and fir st :wat'm the upper poi·tion of the oombustion mixture wi tb a micro burne:r , then ·the l ower part. When gas is evolved , control the heating so that bubble countel' B indicates positive flow or ia at equilibrium {Mote 16) . If liquid starts to back up in B (Note 17) , qui ckly close stopoook 0 and cease heating for a few seconds until the evolution subsides. Open stopcock C and oonti.nue to boil the combustion mixture . Heat the mixture so as to obtain vi gorous gas evolu- tion (Note 18) • .Aft er 5 t o t3 minutes of heating (Rote 19) , remove the flame and fi ll cup N with concentxated H2so 4 • open stopcook E. Close stopcock O and When tho liquid level in tube Mis 2 to 3· mm below the in.let tube of head D, clo se stopcock E {No1ta 13) and open stopcock o. Ascertain that the ~ate of a ir flow thr ough B is stil l about 3 bubble s per second $ Removal. of the Ab sorbing 1.rub e. When 3 to 4 minutes have elapsed since filling tube M with acid , slide the disperser inlet tube up through t he stopper until the dispe rse,:.- i s 1 cm o:r more above. the sol ution. Close stopcock L, then di sconnect the rubber t ul)i ng from t he disp ai· E- er i n let t ube . By means of a long oapi - llary d:roppe1" fill the disp arse:r i nlet ·cube\ with 00 - free dia2 tilled water (Note 20). Oa.refv.J.l y op en stop cock L an<.l draw the water through the disperser. perse:r .} {Do not draw a ir th~ough th@ dis- Close L and. :temove the stopp er f l'om t he oen.trif~ge tube (Note 21}. Rinse off the bottom of the stopper and the disperser with a few drops of co - free w~te~ . then rinse down 2 the walls of tube G with a few drops of the water. Praoipit a t1011 of the Carbonate. Wtth a dro pper Ol' pipet add 1- 1 . 5 ml of barium chloride z-ea gent (1 Fin BaCl , and 0 . 001 F i:n 2 HOl) (Mote 22) to tube B and !.:m!I!ediatel.y stopper the ·tube with a JJ'o o 1, one- hole l'Ubber s t opper ·~vhich is fitted with a sho:rt piece of o.5-mm capillary tubing. Place the centrifuge tube in a bath of boiling wate-r f or 2 to 3 minutes. Replace t he :rubbe:r stopper with a rubber centrifuge tube cap , then centrifuge until the solution is oleal' . Washil'}_g ~he Precipit.ate . in a bu~et olamp . Mount the centrifuge tuba (G) :firmly Moisten tube R in the vacuum siphon appaTatus so that it moves freely through the :rubber stopper . Adjust tuba R so that the tip projects only 3 to 4 , cm below the rubber stopper . Remove the cap f~om t he centrifuge tube and immedia tely insert the rubber stop11er, Ca i'efully open stopcock L until s olu- tion is drawn slowly out through Re With a twisting motion push R down as the solution level 1.•eoedes until the tip of R is ju.st above the preci pitate (Note 23); no :eracipitate should pass into li• Remove the stopper f?'o m the oentttifuge tube s.nd :rinse down the W8.lls of the tube 1rnth 1 to 2 ml of 00 - fttee water . 2 Add 1 drop of 0 92% :r;,henolphthalein to the nixtu'.!e a.nd stir up the precipitate . By means of a d:roppe~ . add o.o5 F HOl, stirring ~he mixture inte:rmi ttently to break up the pr ecipitate (Nota ~4} , until the phenolphthalein eolo:r. baaomes li ght pink; then add 8 to 10 dJ:ops o:f l F BaC1 (Note 25) and 1 drop of 1% 2 Whil e stirring, add o.o5 F HOl dropwise Aerosol OT solution. unti l the mixtul'e is white , then add just l itJ;"OE of th.a a cid in excess., Rinse any pt-ecipi tate • from the stirring :eod into the centr:tfuge tube. Oantrj_fuge until the sol ution is olea:t•; and remove and discard the solution withs dropper . T.j. trat ion Qf the Cal'bonat ~• l inse down the wa1.ls of the tube with 1 to 2 ·ml of boiled wator •. Add l dTop of 0 .2% phenol• ph·thalein and 3 to 4 drops of l F BaC1 precipitate (Note 26) . and then stir up the 2 If necessary add 0 . 05 F HCl d.~opwise with stirri ng , until the mixture is white . then add 1 drop of 0 . 1% methyl orange . F~om a microburet , add standaTd 0.2 F HOl and sti:'t~ in·t ermi ·tt ently u.:..ntil the solution is pink . Place the tube in a bath of boiling water for a.bout l minute , stirring to aid evolution of co • If necessary add more standard acid 2 until the pink color remains, then add about 0,1 ml of the standard acid in exoaas . Transfa:r the solution :to a. 50- ml conical flask and :rinse the centrifuge tu.be with 2 to 3 ml of 00 • free water. While swirling the flask continuously, boil 2 the solution gently for about l minute . Cool the solution ta room temJJeratul'e and titrate with standard 0,05 F NaOH. After . .. . the methyl orange becomes yellow , continue the titration (2 to 3 dro ps) until the pink of' the phenolphthalein produces a tight orange colo:r; ·this last color is the end•point of t he titration . From the amounts of acid and base us ed, calculate the percent of carbon in the compound. Notes . ----- 11 . Solids should be handled in weighing tu.bes, eomr:ro.ati on tube Hold the in a horizontal posit ion and insert the weigh- ing tube inside it so that the sample is o.alivered into the tip of the oombustion tube. Do not scatter partic~es of the sampl e along the combustion tube as the wei ghi.ng tuba is w:l thdrawn. Sines the combustion :fluid occupies only a 1:ihird of tu.be A, parti cles of sample in the u:ppal' part of the tuba may not be decomposed . The weighings shoul d be estimated to 0.02 mg. In order to obtain this accursey, the sensitivity of the balance should be 8 to 10 divisions per milli g~am . Al though this p:rocedul' e has ·been designe d primarily :for I solids, ~ , ~ dichloroethyl sulfide (mustard gas) has been satis- factorily analyzed . Liquids o:f similar or l owe1~ volat :i.lity may be s.nalyzed by this pro ced.u:re , usi. ng a pycnometer for weighi ng out the sample 8 I n order to decrease the possibility of loss b y evaporation ~ the sample should be delive~ed int;o the very tip of tha combustion tube o 12 . Stopcock C mus ·!: be closed duxing the admission of -99- li quid into tuba M; otherwise bubbles of the Viscous combus tion mixture will be formed in the hea d (D) and combustion fluid will be blown up into the outlat tuba . 13 . Out off the flow of liquids before the i ast few drops reach the stopcock; no air should be admitted to the apparatus . 14~ The l'ste of gas flow through disperser K should be watched in making the rate adjUB tment as the main object ive of the now re gulation is to keep the bubbles from K relatively It may be necessary to acljust the rat e of flow occasion- small . ally if the flow through K i s sae:n to vary app:rec:i.ab l y ~ 15. 'vTh en li quid ba o,.ts up in bubb l e count err • B 1 gas is passing f:rom tube M to tube A . any CO 2 This shou ld ba prevented since Which reaches tube A will ba absorbed and cause error; f ox this reason it is necessa~y to observe counte~ Bat all times. 16 . Tube JJI should not be hea ted in one spot a s the oxi- clat ion r eaction may p:rooeed suddenly and force CO t u.be A. back into 2 AJ. though it is sattsfaotory t o o:pe:t'ate ·the s ystem so that the~e i s no actual positive flow through B; there must be ~ J'~Vezt~~-:.tl..Q.Y:! through B. 17 • When the t em:p ers.ture of the combusti on mixture is well over 150° C, t he chTomi c acid decomposes to give off oxygen. The oxygen evolution i s not smooth Ol' easily controlleo. ; therefo:r.e , the hea"<i shoul d be a · :p lie d c&refully. 18111 Actual boiling o f the rnixture is needed ·to oompletely sweep out the co • 2 Occasi onal heat l ng at the t i p oi:' tub e M is - 100- desirable because the bubbles formed there will sweep thl'ough tha entire s olution. 19 . If the solution in tube M has not turned green by this time , beeausa of the decomposition of chromic acid to chromic ion, the mixture has not been heated sufficiently . 20 . All distilled water used for rinsing or washing in this determination should be boiled before being used in order to eliminate the error due to dissolved co • Boil the water 2 for several minutes and store it in a glass- stoppered PyTex bottle from which portions may be removed to dropping bottles. 21 . To prevent absorption of 00 , avoid breathing into 2 the open centrifuge tuba . 22. BaCl 2 wise . Since diffusion in the solution may be slow, the reagent should be squirted in instead of being added dropStirring is not necessary if the reagent is injected for c efully. 23 . Avoid jarring the appar~tus during the removal of solution; otherwise particles of precipitate may ba freed and drawn out. 24 . Expecially in the case of large amounts of precipi - tate , as much as 3 to 4 ml of the 0 . 05 F HCl may be required . Therefore , do not start by adding the acid dropwise but deliver it in jets through the surface of the solution by means of a fine-tipped dropper. Use a dropwise addition when the base is nearly neutralized. It is necessary ·to keep the time of washing -101- down to 3 to 4 minutes to avoid absorbing an appreciable amount of 00 • 2 On the other hand, large local excesses of aoid must be avoided~ expeoially on the surface of the solution. 25. Beoause of the acid in the Bac1 reagent, the mixture 2 In that case do not add may turn white after adding the Bac1 • 2 any more o.o5 F HCl but proceed directly to the cantri :f'a.gation. An excess of barium is provided to prevent solution of the precipitate during the addition of acid. 26. All precipitate adhering to the walls must be scraped off and thoroughly stirred up . Discussion This procedure was designed primarily fo-r the analysis of solid compounds, especially Chemical Warfare agents and their derivatives. The apparatus is less complicated and can be assembled much more rapidly than the dry combustion train: in additiop, single analyses can be made much more quickly with the apparatus described here. However, this method has not been tested on the wide variety of compounds which the dry combustion method is known to handle; therefore, difficulty may be encountered with some oompo11!1-ds. For example, soma compounds containing a ON group have given off HCN when treated with the combustion mixture; hence, quantitat ive carbon values for such compounds could not be expected if this method is used. Halo gens, nitrogen, and sulfur do not interfere. ~• ~ I Dichlo:roethyl sulfide has been satisfactorily analyzed by this procedure by weighing the liquid fr om a pycnometer into the combustion tube. Although it seems ?'easonable that other 11- q_uids of- similar or loWel' volatilities can be satisfactorily analyzed by this method, the maximum volatility of substances which can be adequately handled by this proeadu~e has not been , dete:i:-mined . Compounds more volatile than ~ , ~' diohlo:roethyl sulfide can probably be analyzed by cooling the combustion tuba and sample during the flushing operation. Combust ions performed wit~ ammonium chloride and with biuret proceeded smoothly and pr oduced no unusual or sudden gas evolution " Blank analyses gave less than 0.01 mg of carbon. Some typical analytical results are given in Table x. It is apparent that the method is accurate to ± 0.05 mg of carbon. A determina tion requires about 45 minutes. A brief summary of this work has been published in Analytical Chemistry (1). TABLE X Results of Carbon Analysis by Viet Combustion Method Millig-i-ams of Carbon Found Present Sodium oxalate Potassium hydrophthalate ~ , ~ 1 .. Dichloroethyl sulfide Adipio acid 2.92 2.95 2.15 2.16 1.87 l.84 l.86 1.88 1.69 1.68 1.88 1.92 4 1151 4.54 6 .29 '6.26 5 . 33 5 .33 5.61 5.61 2.73 2.76 5.84 5.85 5 . 38 5.43 5.20 5.20 5 .47 5.51 1. . Fa:i-rington. _P., s. ,· Niemann; e. • and Swift,. Ohem • ., _ll, 1~23 (1949) • 2. E4 H.; Anal. Lindenbaum; A., . Sohub.eit,. J.,. and A?"mstrong. w. D., Anal. Oham •• .!Q., 1120 (1948 ) • . z•., Ind. Eng. Ohem.; 3. MeQi-ee.dy, R. Mt, and Hassid, W. Anal. Ed., li, 52.5 (lf42). 4. Van Slyke, D. D., and Foloh, J., J. Biol. Ohem. 0 136, 509 (1940). -106- Propositions Submitted. by Paul s. Farrington l. (a). Ethers have been used almost exclusively for the ex- traction of complex metal acids (for example, HFeC1 } from 4 aqueous solution (l,2). It is suggested that banzotrifluoride (CF C H }, trifluoromethyl~cyclohexana {CF 3C H11 ) or other 6 3 6 5 fluorinated compounds be substit;uted for ethers. (a). The use of HF in place of HCl should be considered for the extraction of complex metal acids. 2. Salts of alginic acid have received little attention as stabilizing colloids for volumetric precipitation procedures; investigation of their applic.abili ty is recommended. 3. Substanees which redu.c e hydrogen ion cannot be titr@.ted coulometrically by the ordinary oxidation procedures (3). The use of I Cl or CuC1 2 solutions is suggested for such titrations. 4. An amperometric method for determining 'b.alogen or halide concentrations in a turbule:ntly flowing stream should be useful for studying the transfer of material under such conditions. 5. A method for the determination of micro quantities of water might be developed by using a modif.ied Karl Fischer reagent and electrolytically generated iodine. 6. The design of water treatment equipment based on the lime- soda process has been largely empirical and often has been made on the a.s sumption that st1persa.tu:rattan effecrbs (4) wete the mcu3t impol'tant aspeot the ptoblem.,. @f Recently 1 t has btHm noted that the :following rate equation is important (5) :· No thorough study of the effect on reaetion :rate of oaoo other sui,faees has been nm.de. or 3 Such en investigation with data snowing tbe effeots of' temperature should provide a mo:re sound basis for the design of lime-eoda units. 7. !©rt exohange materials hav& been used ss catalysts for acid eataJ.yzed organic reaotions (6). hehange matel'ials whieh are resistant to chemical attack shauld be eonside~ed as possible eata.lysts for ~xida.t1Qn•.l 'eduction i-aaetions; fol! example. acid washed alumina (7) mi.gb.t catalyze th~ oxidation of hypophos• pho'tO.US B. aeid by halogen$ . Indust~ial speo1£ieations of density and viseosi ty ai-e often • given as functions of the particula~ apparatus US$d to determine the physical plJope:rty. Calculations involving those physical prope·: rt1es would be simplified if speeificatio:ns welt'e in tei-ms of absolute units. 9. (a). Meth$nol is suggested as a solvent fol;' the dete~mins• tion of unsatul'ation by means of electrolytieally generated bromine. (b) •. Coulomet:t'io and ampe:r:ometfio methods oan be uaed to p:rovt«e u.eeful in:fol"mation eoncemtng rates of addit:ton and substi tutian wb.en the halogen.$ :react ,with Ol'gar'lic o.om:pounds. 10. (a)• The number of pb1$1ca1 education classes requ!l'ed. of seni0i students should be a'edueed to two p&:t" week. (b) • Seniol!' students should 1l'8Qe1ve more, enQoui-agement to consult with the faculty of the Ohemistl;'y Depa:rtment oon• ceining gfsduate study. Refe.l 'enees 1. Mylius and Hutt ne:t, Bet . 44, 1315 (1911). 2. Swift. E. n., J. Am. Ohern. soe., !§., 2375 (1924). 3.. Thesis, page 66 • 4. Behrman and Green; Ind. Eng. Ohem., 31~ 128 (1939). Sussman. s., Ind. Eng . Chem., ,a, 1228 (1946}. X:ubl1, H., Rel V11 Qhim. Acta• !Q, 453 (1947) •