Ml Tl ‘ 8—— ___’_— l WWW .THS HiGH FREQUEf‘éCY TETRATEOAS TN ACETEC ACéD MEDIA av CONRAD M. JANKOWSK! MASTER OF SCEENGE DEPARTMENT OF CHEMISTRY MICHIGAN STATE COLLEGE 1954 {I . . ‘V 44 111 J a . . 4 . y ‘ e r . . I . . . . . . . v . . . A n . . . . o . T . . . . l. : . n n c In . T : . T . ....a t.JH .. . r O ‘. 1.4 v ,,. y .5 w c . .VosT.-.ok4 o. ...T...»... ... .4 . 4... a u T a 0.4 1.. 4.... . VT: .4. , o...o.“.v.,l. ..‘cuv F-nu-yuu4rn-nnun.“u-nvm-¢Dnarnahihun.ouruvbrhvhtnth».nurnn’hwnFnNhuruhhrnh V rvnrfiufburrhh lawnfibn.pwaFhVnwbrbluuwrnvuu.r-ruvb..nrvwvstw'.puv...u.... . . ,‘ ,.. “A ..,T .. T . . ruwmwu.mowvm-Trurrw.mh:.nm.huuumfl.‘ . . . . . . 4-nwuV-Phrurberbf». FNWNM ‘ HIGH FREQUuNCY TITRATIQRS IN ACETIC ACID mfiDIA By Conrad M. Jankowski A TriESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of EASTER OF SCIENCE ‘‘‘‘‘‘ Department of Chemistry 195% Conrad fl. Jankowski High frequency titration is an electrometric technique of end point detection. It possesses the advantage over existing techniques in that it eliminates the necessity of internal electrodes. L The purpose of this investigation was the stady of the advantages and disadvahta es inherent in the method and a com- parison of this method to conventional methods of end point detection. The instrument used was a conventional tuned plate, tuned grid, 3600 KC crystal oscillator utilizing a 6E5 tried . Beas- urements were made by a capacitive loading of the plate cir- cuit. The effect measured was the maximum biased grid voltage of the loaded circuit while the instrument was in oscillation. The end point detected by the instrument as used is exPlained in terms of conventional conductometric titrations which may be related to this technique by ordinary methods of electronic circuit reduction. or the factors evaluated most important are the effects of solvent system, dissociation of the substance titrated, and the concentration. The effect of the solvent systeu is con- ventional and in accordance with the Breasted theory of acids and bases. It was found that weak oases such as dimethyleni- line, pyridine, hexamethylenedianine, 8~hydroxyquinoline, ani- line, p-toluidine, and p-clanine, when titrated with a strong acid such as perchloric acid gave better on points in acetic acid media than in water or dioxane. The degree of dissocia- [1"‘r‘1r'926 a“; "V§.,'- . I .k.~‘ Conrad h. Janhowshi tion of the SGLSJ5306 titrated in a particular solvent s;stem had a profound effect. The more highly ionized substances gave better end points taah the less dissociated or "weak" substances. The effect of concentration is perLaps the nest unconventional aspect of the high frequency te' diode in t319.t sensitivity of response is not a linear fare tio:1 of concen- tration. ihe meL: -Lois utilized gave two re,ious of maximum sensitivity separated by a region of low sen“: Ht Vitj. It was found taat ea ch s,stem titxeteu had a unique region of highest sensitivity. The location of ti 118 maximum region of sensitivity doperids upon.t11e kind 01 ions present. The reg- ions of maximum sensitivity varied from 0.053-0.002A for ani- line to 0.1H—0.0lm for dicthylaniline. In acetic acid media the precision of himh frequ hey and U C!” *‘3 I.» O m r J CL ‘0‘ 0 law: y... I a c+ Cl fl CD *3 (9 points compares favorably with potem ions the concentration of naxinum sensitivity is utilized. In most cases of weak b- see the Lifih frequency region of fialefifl sensi~ tivity extends to lower concentrations than does the potentio- netric concentration range.‘ iwe potentiorm trio method in ace- tic acid svstems extends to weaker bases tLan does the hi h frequezm 3 method studied here. Thus 1d .h IreqaeicJ teei;.ni- ease and potentiometric nothoos are conploncntary to the study of acetic acid systems.’ A C KN OW 12.11.53 33. *Lfll T3 The author wishes to gratefully thank Dr. Andrew Tinnick than. old, supervision, and inspiration made thin work ponciblo. The writer is indebted to Hr. Arthur H. Johnson for hi: lid in assembling tho oloctronic equipment and his in- voluoblc odvioo in too phases of the work involving the knowledge or cloctronicl. Dru. E. Loiningor and K.G. Stone for thoir helpful suggestions and advice. and Dr. J.L. Hall or Wont Virginio Univorlity for the pro-publication circuit of tho titrimetor used in thin study. TfiBLE 014‘ GUN TENTS LIST OF FIGURES LIST OF TABLES INTRODUCTION HISTORICAL BACKGROUND Applications Instrument Classification Frequency measurement Power loss measurement Dual response measurement Circuit Characteristics Oscillators The equivalent circuit The response curve Cell Types THE TITRATION CURVE EXPERIMENTAL NETHOD AND DISCUSSION OF RESULTS Preparation of Reagents Incidental Instrumentation Apparatus Instrument characteristics Cell types and Characteristics General Experimental Procedure Graphing Procedure Sources of Error 111 CD ‘3 0‘ U1 #r \» u: r4 fig 47' u) x» u: to n: to A; g» F‘ t4 P3 Id r1 co 1» hi U1 \n s» u: <3 ~o to h‘ r4 Discussion of Results Poser loss measurements in aqueous media Power loss measurements in dioxane media Power loss measurements in acetic acid media Cut in voltage measurements Cut in capacitance measurements Loading measurements Peak voltage measurements Effect of concentration Effect of base strength Other methods of end point detection Summary and Conclusions APPENDIX LITERATURE CITED ’ 1+3 ha in 1+? so 52 55 57 53 77 83 83 91 93 Table Table Table Table Table Table Table Table I. II. III. IV. V. VI. VII. VIII. LIST OF TABLES Loading measurement titrations of aqueous hydrochloric acid with 1.5936 N potassium hydroxide Titrations in acetic acid media by differ- ent measurement techniques with .1132 H perchloric acid 8 hydroxyquinoline titrations in acetic acid media with .1132 N.perchloric acid Aniline titrations in acetic acid media with .1132 N perchloric acid Hexamethylenediamine titrations in acetic acid media with .1132 N perchloric acid Pyridine titrations in acetic acid media with .1132 H perchloric acid Diethylaniline titrations in acetic acid media with .1132 N perchloric acid Comparison of titrations of,bases of dif- ferent strengths in acetic acid media with .1132 N perchloric acid 5h 60 63 69 72 76 79 Figure Figure Figure Figure 1. 28s 3. Figure h. F1811“ 5e Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 6. 7. 8. 10. 11. 13. In. 15. 16. 17. 18. LIST OF FIGURES Frequency change instrument response Frequency change instrument response Sensitivity curves for frequency change instruments 9 Power loss instrument response 9 Equivalent admittance network 13 Equivalent circuit 13 Circuit diagram of titrimeter and power supply 29 Response curve of a tuned plate tuned grid instrument 30 Finger cell 35 Dip electrodes 35 Cylinder cell 35 Band cell 35 Polyethylene base electrode cell 37 Aqueous titration of hydrochloric acid sol- ution with 1.5936 N potassium hydroxide us Titrations in various media 48 Cut in voltage in acetic acid media with 0.1132 N perchloric acid 51 Cut in capacitance in acetic acid media with 0.1132 N perchloric acid 53 Loading voltage in acetic acid media with 0.1132 H perchloric acid 56 Peak voltage in acetic acid media with 0.1132 N perchloric acid 59 .Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 19. 1910 20. 21. 21.0 22. 2280 23. 2330 23b. 23Ce 2ua. Peak voltage measurement: of 8 hydroxy- quinoline in acetic acid media with 0.1132 R perchloric acid Peak voltage measurements of 8 hydroxy- quinoline in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of aniline in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of hexamethylene- diamine in acetic acid media with 0.1132 N perchloric acid. Peak voltage measurement or hexamethylene- diamine in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of pyridine in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of pyridine in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of diethylaniline in acetic acid media with 0.1132 R per- chloric acid. Peak voltage measurement of dietnylaniline in acetic acid media with 0.1132 N per- chloric acid Peak voltage measurement of diethylaniline in acetic acid media with 0.1132 N per- chloric acid Peak voltage measurement of 0.000898 N diethylaniline with neutral salt added in acetic acid media with 0.1132 N perchloric acid Peak voltage measurement of base strength in acetic acid media with 0.1132 N per- chloric acid Peak voltage measurement of base strength in acetic acid media with 0.1132 N per- chloric acid 61 62 61+ 67 68 70 71 73 7+ 75 78 80 81 Figure 25. F1 gure 26 e £01811” 27 0 Comparison of peak voltage and potential measurements of 0.01765 N aniline in acetic acid media with 0.1132 N perchloric acid 8h Couparison of peak voltage and potential measurement of 0.0096J3 N B-nydroxy- quinoline in acetic acid media with 0.1132 N perchloric acid 85 Comparison of peak voltage and potential measurement of 0.003915 N diethylaniline in acetic acid media with 0.1132 N per. chloric acid 86 INTHQD33TIGI high frequency titration is a relative newcomer to the field of electrometric analysis, consequently most studies have been primarily concerned with instrumentation and the theoretical basis for measurement. Only few papers have dealt with problems of construction and Operation. Although high frequency methods possess unique proper- ties only a few of its possibilities have been realized. Po- tentially such an instrument could be widely useful wherever a measure of solution concentration is desired. most studies to date have used high frequency methods on classical deter- minations for instrumental and theoretical background, rather than for the purpose of improving the determination. This is to be sapected, since instrumentation and theoretical rami- fications must be understood, prior to application. Several excellent review articles have appeared (9.11.33, 57) and reviewers agree concerning the nature of the measure- ments, however instrumentation seems to be a matter of indiv- dual preference. In light of the foregoing, this study stresses applica- tion of high frequency techniques to analytical problems, rather than theoretical aspects. An investigation was made of various Operating techniques on a simple high frequency titrimeter of the type described by hall (39). Various cells were evaluated from the Operational stanfipcint. finally s study was made of hi n frequency spnlicsticns to nonscceous acid-base titrations. tines this is not an ennsustivs re- port on the subject of high frequency analysis in nonsqceous medic no study is specifically linitei to some acetic acid systems. HIT”: ‘IJEiI CA.) LEA (1123 2:4) J'T'ID High frequency analysis, a method in the general field or cloctromctric analysis has gained widesorcad attention in the past few years. This methoi possesses certain advantabos over standard methods of clcctro-analysis. Uonductomotric, potantionetric and ailici mctqodc do not give satisfactory results with a great many innortant systems, an: many systems which give adeqcato results are contaminatcd by tno immersion of oloctrocca or deactivate tnc electrodes unless elaborate precautions or. taken. Applications The high frequency methods, even though tnc_ have inher- ent limitations, have been used for many diverse determina- tions. Successful high fr6qnancy analytical procedures havc been reported for: classical acidinctry, (2,n,ll,22,36,39, h0,h5,53,66): nonaqncouc acidimctry, (£5,65): oxidimetry, (uS): general precipitation reactions, (21,30,n0,§5,62,6o)3 adaptation to continuous reading and constant rocoriing appa- ratus, (21,6n); kinetic studies, (25,30,32,u6); beryllium de- termination, (3); norcurimotric determination of chloride, (12); micro titrations (16); volunctric thorium determination, (1h); detcrmination of calcium and magnesium ions, (13,47); analysis of mixtures, (67); dctcction of chromatographic zonae, (52); sulfate determination, (51,62); determination of water in alcohols, (53); moisture tester, (6)3 indicator for pipe- line liquid separation, (26); measurement of dielectric con- stants, {1,5,25,31,33,37,58)3 dipole moment measurement, (2?): a comparator for solutions, (19): capacity measurement, (24); argentometric titrations, (B): versenats titrations, (lo); and dimothylglyoximo cholation, (53). Instrument Classification A brief survey of the litsraturo on this subject, would seem to indicate that tnese mctnods, each using different techniques, instruments, and measuring different electrical properties, are only related by their use of a high frequency alternating current. This observation givos rise to the gen- eral characteristic common to all high frequency measuring de- vices. When a solution is placed in the tank circuit or an oscillator, circuit parameters due to the solution alter the characteristics of that oscillator to a degree determined by the nature and concentration of the solution, and as the con- centration of the solution changes, these oscillator charac- teristics are furtncr altered. Ideally, for analytical pur- poses, a change of solution concentration should cause a lin- ear response of a single variable characteristic. In prac- tice the oscillator changes that are noted in conjunction with solution changes are change in frequency and change in power loss. In the langua;o of current litcrature the solution changes measured are capacitance and high frequency conduc- tones or a combination of tho two (57). This dual nature of the change offers a convenient dis- tinction between types of instruments for purposes of descrip- tion. They can be classified as instruments which measure change of frequency (capacitance), those which measure a trunction of power loss (conductance), and those which meas- ure a combination of the two. ‘Erequency neasurcncnt,_ The frequency measuring instru- ment has a response of the type, shown in Figures 1, 2 and 2a, and can be further broken down into two subnroups depending upon actual operation of the instrument. One subgroup util‘ izee two oscillators to measure the change of frequency, an indicator oscillator and a reference oscillator. This is the so called beat frequency technique Cll-lu,37,52,5u,54,66,68). The indicator oscillator contains the unknown to be analyzed in its tank circuit. During the determination the character- istic frequency of tno indicator oscillator is changed. The degree to which the frequency changes is primarily a function of the change of capacitance of the cell containing the un- known. "to alternating current of different frequencies from the two oscillators is fed into a mixer amplifier circuit and the chance in the frequency of the indicator oscillator is recorded as a beat frequency change. Tnis beat frequency change can be determined with a frequency meter or by deter- mining the unprepriate lissajoua pattern of an oscilloscope. The reference oscillator may be an unloaded indicator oscil- lator or any stable oscillator Operating at an appropriate frequency. It has been suggested (as) that by using a stable Beat Frequency (Kc/sec.) l l l Log of Molarity at 30 Megacycles Fig. 1. Frequency Change Instrument Response 5- A e; 514'». C o D A 1M0. 2° mBL. B 3M0. ‘3' E 0 g c 10 MC. 52f D 20 MC. 4.) ”3 E 30 MC. an a O 1000 2000 3000 hooo 5000 Specific Conductance K0 Fig. 2. Frequency Change Instrument Response receiver s best can be taken on the tone of fictional Bureau of Standards station W.W.V. In the second subgroup of frequency measuring metnods an indicator oscillator is used, which by suitable tuning devices in its circuit is returned to its original frequency (8,25,27, 31,3?,h3,57,61,62,53). Usually the tuning is achieved by re- moving capacitance from the tank circuit although it has been suggested that slug tuning of the inductance may serve the same purpose. In this method s reference oscillator may or It: not be used. The tuning usually is to the original reso- nant frequency which may be determined by means of a frequen- cy meter or s suitable detector with a ”magic eye” indicating device. If a reference oscillator is used, tuning is carried out on the indicator oscillator until a zero beat results on the recorder, which may‘be a null-point galvanometer connected to e rectifier circuit, an oscilloscope, by earphones or some audio method when suitably amplified. Power lees measurenent. fine second general technique of high frequency analysis is the method using power loss meas- urements which is a function of conductance. Changing the concentration of a solution, which in cart of a tank circuit of an oscillator, changes the characteristics of that circuit through change of the resistance of the solution and its sub- sequent absorption of real power free the oscillator circuit. This power loss when the frequency remains constant. manifests itself by changes in the electrical preperties of the oscilla- tor tube in the tank circuit. The changes and measurements which have been reported are plate voltage or current (lh,27, 31,51,65), and grid voltage or current, (2,3,h,29,3),53,58). The characteristic responses for this tyre of measurement are shown in Figures 3 and h. Measurements of the real component- of impedance in which the amount of radio frequency current transmitted through a cell is rectified and measured have been described by Blake (16-20, 22-25). Dual reggcnce neceurenentlp A third type of instrument gives measurements decendent upon changes of both frequency and conductance. These are the instruments which measure vol- tage or current change in a tuned circuit which may (2,3) or may not (uh-h?) be Operated at resonant frequency. Some de- vices measuring impedance directly have been reported, (36, 39ohao57)o Circuit Characteristics The specific response obtained is dependent upon the relationship and values of the circuit parameters. Tie para- meters may be espreeeed as complex vector Operators arising from instantaneous values and operating upon the network quan— tities of current, voltage and frequency. The vector opera- tors derived fron the parameters are variable in two senses, directly and indirectly (not to be construed as linear and non-linear relationships). Directly variable are those vec- tors which change due to changed parameter values such as changed conductance or capacitance caused by solution change. A Methanol B Water Capesitanfig Change (mmf.) Spe399ic Conductance K0600 Fig. 2a. Sensitivity Curves for Frequency Change Instruments A HCl ’ B NaCl a; E A B ~I F .p c a) I- L. n a O L— Log of Molarity at 30 Megacycles Fig. 3. Power Loss Instrument Response 11 Changes due to a volitional change of one or more of the mul- tivalued parameters installed in the circuit are also direct variables. Indirectly variable are those vectors which are changed not by a change in circuit parameter values but rather a vector readjusted by a changed circuit quantity, chiefly frequency change due to a directly changed vector. Individual circuit components give rise to one, both or neither directly or indirectly variable vector quantities. Oscillators._ A discussion of oscillator network param- eters must of necessity recognize the inherent differences in different types of oscillators. To date a wide v.riety of dierrent oscillators have been used. Workers have reported successful use of; tuned plate oscillator (Bu,h5,5h), Colpitts oscillator (ll,l£,lh,lS,u7,66), tuned plate and tuned grid 03- cillator (h,32,hh,h6,h7,§l,52,&u), and the crystal oscillatcr, (25.27.31.39) and others less frequently reported include cathode coupled, magnetic feedback and quarter wave length concentric oscillators. The or terion for choosing an oscil- lator is its stability and reproducibility. Equivalent circuit. High frequenc; measuring devices may be resolved into an equivalent admittance network system. The admittance network of Figure h composed of non directly var- iable vector quantities is a general representation of any high frequency instrument. The values of the admittances Y1~Y6 de- termine the characteristics of the instrument. Tue difference between instruments arises from the characteristic measured, the manner of measureaent (A—A') and the experimental set up of the solution measured across B - 3'. Thus the manner of measurement and the circuit quantity account for differ- ences between frequency change, power loss. and dual response instruments. in» network is the equivalent circuit of the specific oscillator used. The solution-cell equivalent cir- cuit is the representation of the experimental set up. Fig- ure 5 is the equivalent circuit of a solution in a capacitance type cell in parallel with e variable restoring capacitor. with the aid or specific equivalent circuits the response of an instrument may be calculated as a function of solution concentration. Response curve. Tue response obtained or measurement is e function of current, voltage, and frequency, operated upon by couple; Operators, he directly (Ix) and indirectly variable (2.) admittances. Considering the cell-solution equi- valent circuit analysis the expression Ex O GP+J39 is obtain- ed where GP 3 (eekcil/[&Z+I2{CS+CC)ZJ and EDS [{ICOkZ+I3CcC’3(Cc+Ca)} / {k2+$2(33+3°)23)+['czl Y the net admittance of equivalent circuit Figure A Gp high frequency conductance term Bp imaginary part of tne admittance I 21f, where t is the frequency k low frequency conductance Cc capacitance due to walls of container capacitance due to solution C2 the variable capacitor for circuit adjustment I 13 Y N E T: Solution 3 _ ‘ Characteristic A! —J Yel - m B' * LZI' L_é.J Figo no Eduivalent admittance network I 02 Rs \/W\r fl) 00 Ca II #II If ll Figo 50° Equftnlcnt Circuit Instrument control Czoecitwncc tnroufin the container Solution capocitC1ce Reefntcflcc of colution 1h J is an operator. If 02 is set equal to zero the expression reduces to that of Reilly and thurdy (5?). Circuit reductions of other cXperimcntcl set ups have been reported (39,57). Approximations can and have been made of the function, f(YI,YaY,K,f) knowing the appropriate cir- cuit parameters of Y: with the attendont assumptions as to the nature of the changes or constancy of Ya,V,I,f (39). Because of the nature of the assumptions and the ffiut that the equiva- lent nctwork is also a simplified approximation, perfect agree- ment was not expected between calculated and exocrimeutel re- sults. The equivalent circuit of the solution (Figure 5) and its calculated admittance can yield a great deal of informa- tion concerning the response of a given type instrument and the correlation of this response to earlier eXperimental roe cults. In the frequency.meceuring methods, frequcn y'which is a function of capacitance, or beat frequency itself, is plot- ted ag inst concentration change of the solution. The plots give rise to curves similar to Figures 1, 2, and 2a. To ex- plain the shape of these curves it is necessary to refer back to the equivalent circuit, Figure 5, aesmuing 02 to have in- finite impcdonce. when the impedance of H3 becomes small com- pared to Co (at high electrolyte concentration) C8 is almost shorted out and the frequency approaches 1/(27735) asynptotlo cally; In dilute solutions in which the impedance of R3 is large compared to that of C5 the frequency approaches (1/2 w)C/('s,+cc)/(Lc,co)] asymptotically. In the interme- diate concentration regions where tee impedanccs of H3 and C, are of the same order, has frequency varies monotonically with concentration between two two limits. For conductance (power loss methods) an explanation of the sense of Figure 3 is based on the assumption that tne re- sistance Rs or the solution is changing while the other prop- erties remain approximately constant. At high concentrations Rs is small and passes the current with little absorption, how- ever, absorption increases with increasing H3. Above a cer- tain limit tus capacitance of tee solution 6, begins to shunt greater portions of the current and when R, is large enough, absorption is small because the current in R, is Scull. Thus the power absorption passes through a maximum as‘does tne plate current, etc. From Figure 3 it can be seen that there are two regions of maximum sensitivity (on either side of the hutp). Consequently the concentration limitations are much less sc- vere than those for frequency measuring instruments. ihe dual response instrument which measures the effect of both dielec- tric chon e and power loss change has given excellent results (t4,h5), however, in some titrations a rapid conductance change near the and point may work in cepositicn to the capacitance change and make the end point determination less scarp than would othersidc be cxyected (33). In the preceding discussion little has been said concern- ing the effect of conductance on frequency measurement.1 In most cases it affects tne frequency only slightly, however in case. of larger conductance change (overloeiing tne occille- tor) it in poacible to dampen oscillation completely. Whenever conductance is sufficient to change the frequency acpreciebly the oscillator must be brought back to maximum resonance by adjustment of a variable capacitor in the tank circuit. These conclusions are in agreement with the conclusions drcvn by earlier workers. Forman and Crisp (33) chewed that change in frequency was due to change in.dielectric constant of the solution (or change in capacitance of the cell) and power loss III primarily a function of the conductance'or the solution, remembering that in cases of measurement with rapid- ly chenging polarity, with the electrodes isolated from solu- tion, the ordinary ohmic conception of conductance is invalid. The op tcnm ie the reel component of the complex admittance function. ibis real conductance component in ionic solution is the result of the movement of molecules relative to their2 V illn electronic terminology, change of frcqccncy due to loading. Elna absorption of power manifests itself by a fleeting effect. A rise in taupereture or change in viscosity of c Iolution in a tank circuit, other thinaa remaining uncnangcd, will cause a change in the cheracteristicc of that tank cir- cuit. This could conceivably be e source of error in titra- tions, however, the voltages used are low, consequently the omount of power dicsipcted is low enough so that tenpercture effects can be ignored. 17 neighbors. The power loss (conductance) effect is concerned with the storage of energy and its subsequent return to the signal source.' If the source and toe system did not have their inherent energy losses it would be possible to transfer energy from source to system without attenuation. The nature of the' conductance change has been investigated by Formsn and Crisp (33) and Richards end Loomis (53), who have derived mathema- tical expressions relating power loss to specific conductivity and dielectric constant of the solvent. An empirical relation- ship of frequency corresponding to maximum energy loss has been shown to be AV: Kgby For-man and Crisp (33), in which Xis the wave length, for maximum power loss, Y is the concentration of the solution and K3 is a particular constant for each simple electrolyte.‘ “his relationship was determined by means of the temperature rise of a solution tested in s calorimeter during a definite time interval. Txis relationship is indicative of the concentration'limitctions of the method, since outside of a certain concentration.rsgion, the degree to wnich power loss or wave length (frequency) changes witn respect to concentra- tion falls of sharply. Consequently outside of this region the sensitivity is not great enough to satisfactorily indicate an and point in s titration. Cell Types Earlier workers used inductance type cells, (31,4n,45). Unknowns were placed in a glass container union in turn was placed in the inductance coil of the oscillator circuit. 13 Several disadvantages are said to be attendant upon coil type cells Union are said to be overcome by using the capacitance typo cells (11). Reported disadvantages are a lack of repro- ducibility and s lower order of sensitivity than that of the capacitance coll. Recently Fujiwara end Rajeshi (36) have re- ported overcoming tness difficulties by modifying the oscil- lator circuit, snd measuring technique. The capacitance typo cell is a container, usually glass, uitn external condenser plates which are part of tne oscillator circuit. The rooms- try of the cells reported has been diverse and greatly de- pendent upon the system analyzed and the particular use. dost employ s pair of metal bands about s glass cylinder. Several studies of cell geometry hgve been made (Su.57) and it has been found that plate size effects the landing of the oscillator. Cells have been designed for special purposes, flow cells for constant recorcing apparatus, covered cells for inert atmosphere studies, finger type cells, and cells for studying chroeatogrsphic zones. 1:, TI -411.” 0‘5;ch The shape of a titration curve can be explained pictor- nlly by considering the titration graph to be the locus of a point moving on the appropriate response curve. For a speci- fic example relating concentration to grid current by moans of the response curve of Figure 7, consider a titration occur- ing in 3 concentration range where the initial and end point concentrations arc on oppositc sides of the "hump". In this caac the grid current increases at the beginning of the titra- tion until the "hump” is passed. after which the current dc- crccsco as the and point in approacncd. After the and point "hump" is passed the grid current increases passing over the and decreases again. The resultant titration graoh has n "pip" chased ccpcarancc. Eeponding upon the starting point and the concentration changes which occur, a largo variety of complicated titration curves can be obtained. For the case of an aqueous titration the graph is similar to that of a conductomctric titration, providei the titration is curried out on the linear portion of tho concentration-fro- qucncy change or concentration-power loco cnangc relationship. Up to the end point the change is due to the sum of the chang- cc from decrease in concentration or substance titrated, the incrcacc of reaction product, and oilution factor due to inc creased volumo upon addition of titrant. After tnc end point, the change is due to tea cum of changes from ciiition of ex- ceea titrcnt, and the dilution factor. In cruer to obtain a sharp break at the end point and linear response during the titration it is necessary that the substance being titrated and the titrent both be in the concentration reagel or great- est sensitivity (greatest chance) for tact particular frequen- cy. Assuming linearity of the green, 8 second oetimue canal- ‘tion exists. For sharper and point creche, the Greece titrcnt should change the direction of the response, or a least cnange the elepe of the graph sharply. Eon-linear curves enoulu have a cusp or inflection at the end point. Ike considerations men- tioned above hold true for both power measuring instruments and frequency measuring instruments. The concentration and non-linearity limitations can be overcome somewhat by the use of hinner working frequencies (Figures 2 and 23). If an ideal high frequency titration can be postulated, a green of this titration will poeaecs certain attributes againnt which the prepertiee of actual high freeiency titre- tion greens can be contrasted. the comparicon will show the efiventegee, disadvantages, ontinum.worxing coniitians, uni the greatest sources of error inherent in tuo metnole used here. The ideality ctriVed for in tic graphical representation is primarily based upon two conditions: a sharp break at tne ““1i33£*t£§ end point tneflerecenee of reaction creuuct, if I highly ioniecd substance, will influence tne Optinue concen- tration for the greatest cnange. Bl eni point, and a linear rceoonae of the depencent variablo.2 Thus the ideal high frequency gregh would resemble a typical conductcmetric titration graph. A high deireo of curvature indicates a #rent departure fP03 linearity col acifie free oth- er consioeraticnc a more inferior graph. ineee otger coccid- craticne, wgion are often mitigating factors jnetifyin; con- clncionc, are: symmetry, nueber of cointa, one angle of break. The most ieportcnt or the turee is the angle of tee break at . me and point. Ideally the break at toe and point should be as sharp no possible, thus the egaller the angle between the tangents at the end point, the sweeper the and point, an& con- versely no break at all would result in an angle between the tangents of 133°. Unfortunately two opposing tcoéenciea are present here, each limiting the angle of the end point break. the smaller the break 1.6.. as too tangents approech 130°, the greater is toe effect of random errors in chnoging toe lo~ cation of the break. The larger toe brook, i.e., no the ten- :enta approach 00, too greater tJfl range neoeeeery for tnc / 2A more linear rec once for a particular concentration range can be achieved by increasing toe coexing fr quency of the instrument and/or within lieite, clenged cell geonetry and plate cite. Heilley and occurjy (5?), have snoen tact for a parti- cular ooncentreticn and a pooticeleo system, the cosine pro- jection of tee aimittcncc of tue parallel equivalent circuit for the instrument used under specified ccnfiltione when mode a function of specific conductivity will yield a transfer :lot by means of which toe eelivclen* coniuctonotric titration graph may be prepared. Shin ic matnauetically echVelent to allowing the response to become a parameter of a function of specific cocductance union is in turn plotted against the in~ d coolant variable yielding a lineer green. P») N measuring instrument, Which in this casa is a vacuum tube voltmeter (V.E.V.ao). Thus the minim“u angla of tho break is limited by the maximum vcltaie cbtainabie dgfi to tha na- ture of the oscillating triodo and also tne maximum ran5@ of the vacuum tube voltaater. ihe Optlhuu angle of baa tantnnta at the bruak dao to txeae considaratlona is in tJu vicinity Of 90°. It can be said from a qgalitative standpoint, that if a curve passenges axes of symm try £43 swallsr is £46 probabil- ity of randon errors preaeut causing a distcrtion of the and point. It can also ha said qual titively, for a greater num- bar of points the probability In larger taat in a snoothed curve the positive and negative rafiloa errors cancel. EXPLRIdLKTAL METHOD AhD DISCUSSION OF PUSULTS Preparation of Reagents All chemicals used were of reagent grade. to special purification of reagents was carried out except in the case of dicxane as mentioned and the aluminum oxinate which was reprecipitated from a dilute acetic acid solution several times. Melting points were taken on the organic solids to confirm their purity, agreeing to 3:10, except glycine which decomposed at a point 10° lower than the theoretical decom- position point. Solids used were weighed out to the nearest 0.1 of a milligram employing weights calibrated against Bureau of Standard weights. All solutions were compared against stand- ard solutions using potentiometric, conductometric or visual indicators, as a means of detecting titration and points, wherever these means were applicable. Liquids were measured in calibrated glassware. Aqueous acids and bases were prepared to the approximate desired normality and standardized against known acids or bas- es. Primary standard grade potassium acid phthalate was used as the primary standard. End points were detected conducto- metrically or visually with methyl red or phenolphtualein in- dicator. The reagents used for aqueous studies were reagent grade, acetic acid, hydrochloric acid, oxalic acid, ortho- phosphoric acid, phenol, potassium hrdroxide, sodium chloride, sodium hydroxide, and sulfuric acid. Acils and bases in dioxane media were prepared with dioxane rediatilled over sodium. The bases were compared against a standard perchloric acid solution in dioxane us- ing methyl violet in chlorobenzene as the indicator. The perchloric acid solution was standardized Against reagent grade diphenylauanidine using methyl violet indicator (3%). The reagents used for studies in dioxane media Herc, aniline diphenylguanidine, perchlorie acid 70-72fi, pyriiine, and p-toluifiine. The acetic acid media acids and bases were prepared from du?ont C.P. glacial acetic acid and C.B. acetic annydride. The reagents used were,;9-elcnine, aluminum oxinate, p-amino acctoyhenone, ui-amino-isobuteric acid, aniline, bcnzidenc, bencoic acid, diethylanilinc, glycine, hexamctnglenediamine, fi-nspthylamine, 8-hydrcxyquinoline, perchlsvic acid, pnenol. phthalic acid, potaecium acid phthalatc,;wyggiua, sodium acetate, sodium perchlorate, p-toluidinc, and urea. '30 glacial acetic acid was assumed to contain 0.5; or water. Enough acetic anhydride was adficfi to react with 50} of the water present. Fifty percent was chosen so that the final solution would contain loss than 0.3% of water and no excess of acetic anhydride. The 0.3; or less water level was shown by earlier workers not to affect the solution measure- ments (34,65). In tie preparation of perchloric acid solu- tions enough acetic anhydride was added to react with age of the water present. Perchloric acid solutions were stanlard- iyed against primary etaoaard potassium acid phtnalate die- solved in acetic acid (c1). Comparisons of solutions were made conductoeetriceliy, potentiometrically and visually using methyl violet indicator. Incldoxtal Instrueentction Aside from high frequency noaenreuents on toe instru- ment, which will be conceited below, comparieone and inci- dental measurements were carried out on other instruments. A Beckten model “-2 line Operated p3 meter using type {ijo-SO rod label glass electrode and a colonel tyne #49?) fiber elec- trole, was used for titrations in acetic acid media, in the manner rcportei by previous workers (3A,S5,So,59,63). An in- duetrial Instruments conicctivitj bridge model kc-la at 1003 cycles per second with Beckmnn clatinized platinum lencreion electrodes, was used for titrations in aqueous media and for some acetic acid media titrations in tae cecal manner. A Heatnkit vacuum tube voltmeter model V-EA union was used to measure the grid potential of the high frequency titeimeter. Line voltage '88 controlled by a lZD volt cola constant vol- tage transformer Serial D-73537. A 3.3. Army Signal Corps surplus frequency meter type BG-afllu was used to determine frequency reproducibility of the excorimontal instrument. A Fischer electric block melting point apparatus was need to de- termine the purity of the organic collie. Apparatus The titrimeter used with few modifications, chiefly in the power supply, has a circuit (Figure 6) wnicn has been re- N} 5“ corded by hall uni others (1.5.39). A 3599 kilocyclo cry- stal with too appr0pricts coil was used ratocr than too two megscyclo crystal used by flail, the variable capacitor uti- lized 3 national tesrei drive, and the 20 volt direct cur- rent was taken from a lino Operated, half wave voltage rc- gulator power supply, employing a selenium rectifier, (Pig- urc 6). A short description will be given hcrc of too function and Operation of the various congoucnts of the circuit. The 623 "magic cyc" tuba functions both as an oscillator triodc Ind a resonance indicator. Inc triooc oscillator circuit is of the tonal plate tonal grid typo with a piazo-activo quartz crystal as the grid portion of the rasonant circuit and as the frooucncy controlling dcvica. The circuit oscillatcs when- Ovar the plate parallel resonant circuit is tuncd to the same freqncncy as the fixed quartz crystal. The transfer of ener- gy between plate ani grid circuits takes place through too plate-grid cap o tones of too tube. Oscillation exists than F = [ZHMCflfl Ullfil‘fi 3: is too crystal {.m::1uoncy, j: is the inductance and g the capacitance of the parallel resoncnt circuit. Toe state of oscillation is marked by a oecrcaas in plats correct, grid current, and {rid voltago. Graph 5, Fig~ urs 7 is c representation of baa grid cip responsc for a tuned plate tuned grid oscillator as described by J.L. Hall (39). The instruocnt used hero being ccccntially the cane in“tru- mcnt with a few moiificationc has a similar response. 3‘0 .4 In tho grii circlit the R.F. choke provcntc too high frequency alternating current from short circcltlnfi tho cry- stal through tnc grid leak resistor. fhc 639 "maria eye" also functions as a cathode tube in- dicator. When there is no oscillation the plate current is high, thus the tuba target is maintained at its lowest poten- tial with respect to the cathode. A control electrode is connected through a resistor to too plate of the triodo and is therefore at a higher potential than the place, due to the potential drop ccroca the plate r cictor. ihc shadow unglo is the widest at this point. When the circuit is in oscilla¢ tion there is a redistribution of potential around too plate circuit. The potential drop from too cathoac to two plots increases becaose the cffoctlvc resistance of this unit in- creases while one plate loci resistor remains constant. Inc plate current decreases uni hence toe voltage drop across too plate loafl resistor dccrcasea col toe voltcgc of too plato ao- proachcs the voltnho on the control clectrod. and tac chadow angle becomes smaller. If the capacity in too plate tuning circuit is incroacel or dccrccced beyond certain limits it will no longer be in resonance with the quart; crystal and oscillation will stop. At this point there will be a cuflicn gum; in plate current, the potential of the slots w on rcspcct to the cathode fic- creascc uni as a conscqacnce the snaloc angle will alien abruptly. This point is rcgr ducrblo cnfl can be used as a basis for measurements. 28 In parallel with tho capacitance type cell 13's series of capacitors, Cl thPGUgh 05 inclusive, Figure 6. The 01 variable capacitor is the rougn aajustmcnu, tne G2 variable capacitor, with tnc national reducing drive, is aha fine ad- Justmcnt for the propcr capacitancc to produce oscillation. The fixed capacitors 03 tnrough C6 allow the circuit network capacitance to be roughly adjustefi to prupcr range fur oscil- lation to occur depending on cwihcn pocition. She fi3 variabla resistor acts as a potcnticmctcr wnich given an opposing voltage, balancing the measured grid vol- tags. The opposing voltage aervca to bias the vaccum tube voltmeter (V.T.V.fi.) by a predatcrmined constant pctential. The rasponse curve of graph E, Figure 7 is given in terms of the absolute grid voltage. In the titrations of Luis study I bias was placed on tgc vacuum tube voitmcuar. 1: a b1&l whose value 13 larger than tuc maximum value of hue grid vol- tage is imposed, an examyle of which might be a twenty volt bill. on the grid dip reeponcc curve of graph 3 figure 7, a rovaraal of tno responac 13 obtained as depicted in graph A Figure 7. It can bu scan tact a twenty volt bias will result in voltaao reacinga such tnat a twenty volt grid voltage given a reading of 0 volts on the V.E.€.i., nineteen grid volts. givaa a one volt reacing etc. 136 purpcao of biasing the vol- tage was no that by chcosing tac procer bias voltage tne range measured could cover tne entire voltmeter scale and so in- creaao the pointer deflection thereby dccrcasing baa rcading 29 £22qu x \\\\\V 1 ,A: 5' - ~A 3- ‘ ; n r-w, .v -‘ 3‘ ‘- -( ‘ . », -'— ~. . . - - ~' \ ‘ y - F244: 60 0-1.‘3 1.1. J 143‘ 5' “7 Oi L1.c.:“i.v1,{: cu- 5:.11 DOJO -‘ 34130le- *A parts list :11.” g; ‘09 104.11.}. ;..n she gypencixo l 16- ltg E; H O V.T.V.M. Reading V0 30 8 N G) A V.T.V.M. Bias B No V.T.V.M. Bias l J 0.1; 0.8 1.2 1.6 2.0. 2.1; 2.8 Concentration (moles X 102) Fig. 7. Response Curve of a Tuned Plate Tuned Grid Instrument 31 error.1 Thus tne response of this instrnmsnt can ha dis- cussed in terms of two dirfcront curvss. Titrations for tnis study wore carried out using both the ”bowl” shaped response and tho "nump“ anspsi response. No difference was expected or found in tne observations or conclusions from tocso observations except tno inversion of the relationship between conductance can rosyonse. Two in- clusion of curves from both types of rcsxonso (which are moro- 1y horizontal mirror imsgsa of coon other) was for tao pur- pose of showing the complete convertsbility of one idontity to the otnsr. dnloss, otherwise specified, tua voltage moss- uremcnts are of tno ”hump" shaped rssponse. Fo any related set of cuers on any one figure tho use of a particular res. posse mcascromsnt in consistent. Instrnmont;chsrnotaristics; daproduoibility of the in- strumont was determined using an Army surplus frequency motor. Out in and out out points as detsrminsl by suddco ooening and "magic eye", were exactly rooroiuoiolc wusn closing of the detsrminci as the null-point of beat frequency by using car- phonos. Ehc variablo conifinssr dial rosiing was rsprodncibio to‘: 0.3 dial units. This was tested on various portions of the Variable csoaoitor range by txs excellent of nsind differ— ent concentrations of hyirncnloric acid solutions in tho cell. ins variability of the dial roafiin s was found to be greater line reading error of the v.i.V.a. in the bursa volt range estimated as being I ;;milivolts. 32 It the firct 5% and thc last 5; of the capacitor range. Peak voltagez measurements wore exactly reproducible with rcspoct to voltcgo and frequency for ccvercl different coll loads. At the pack voltage relative cacccitor readings were reproducible to;:’2 dial units. At the peak voltc:c point the frequency was found to be conctant even under changing load conditions in the cell. At otner grid voltages not on the flat top portion of the characteristic grid voltage cc- pacitancc curve the frequcncy changcd different amounts. Reproducibility of grid voltagc during loading measurements was possible but only it great care vac tckcn with repeating tho prearranged capacitor setting. Stability measurements wcrc made at various grid volt- ages and it was founi tact at all grid voltages tccro was a slow steady rise which caused after a pcrioé of 90 minutes. The rise at the pcak voltage was slower than that for lower grid potentials. The cause of this, although not definitely ascertained can logically be attributed to heating of the crystal curing occillation aac cooling of tnc triodc during oscillation and tccir acbccqucnt attainment of thcrmal aqui- librium with the surrouncingc. Etacility with rccpoct to stirring was moacurod using both majnctic stirrer and motor driven paddle stirrer. It~ fifl 2A ficmp goaocd_ccrvc sxcwod to the right witn a flat top resulted when capacitance removed froc circuit is plot- ted against grid voltage. lhc grid voltage on tic flattened top portion in the "peak voltatc". 3) was found that 'hc glass paddlc stirrer coco motionless in any position or rotated at any soocd had no effect on tho grid potential, unless air was beaten into tho solution. At a speed watch air was beaten into tno solution, violent fluc- tuations of grid potential occurred. fiith the magnetic stir- rer the some effect was notcd. Also noted was generally low- ered aonaitlvity (identical solutions t tr to} under idonti- cal coniitlons had a less sharp onj point brook). A motion- less magnetic stirrcr gave varioi roadings, dooonding noon the orientation of the stirrer clement. too two latter of- fectc were attributed to the presence of a mass of metal in the stirrer element. Coll Typos and Characteristics Several different tygo cells wore toatod for apoiicobil- ity to tho pocacnt project. Toe cells described hero arc ty- pical high frequency titratlan cells in that tnc oloctrodea are electrically insulated tron the solution by the container walls. The electrodes are suitably coupled, by means of co- laxial cablo or by direct oonnection3.to toe plate circuit of the oscillator (Figure 6). In the finger typo coll (Figure 3) a {1853 test tube or "fingo " was fused over a circular Opening in the bottom of gulrcct connection woo achieved by fastening a male coaxial connector to tho coil and to: facclo coaxial con- nector to the instrument in such a manner that two; could be connected wifhout an intervenigg cabic. 33+ t beaker-liko container. do the outside surface: of the con- tainer copper foil was cemented resulting in two cylindrical concentric condcnccr plates, with too solution concentrically contained between the two plates. flcchanical difficultloa in construction and inofficicnt stirring made the use of other cells described here more feasible. 'Dip clectrocea (Figure 9) of different since, shape: and modes of shielding taro devised. loose consisted of too con. dancer plates coated with on insulating material, immersed in the solution, cllowiofi tho solution to flow between and around the plates. The chief difficulty incurred in such arrango- meat was that the level or the liquid changed with respect to tho clactrodcc, consequently a very large part of the cocoa. measured was due to volume change. The cylinder typo coll (Figure 10) consiatcd or I beaker typo container on the outside of which two homi-cylindcrc or copper foil were ccmontcd as condenser platac. This cell proved to be entirely catinfcctory in use, providcd suffi- cient colution was used to minimize the volume cnangc effect. One dafficicncy of call design for all of the above mantioned cello ohich.vac not overcome, but witi careful titration tech- nique. could be minimized, was chiclcing the cell from stray external capacitance and radiations. Tho band typo coils consisted of a series or glass con- tainers on_uoicn two metal bands were placed, one above the othcr (Figure 11). Bands of brass, aluminum, coopcr and gal- 35 tfl: _.J Figo 8° Finger Cell Figo 9. Din Electrodes * % § - __J # F—u‘ '——-w Jfl ___J1 _ I i Figo 10o Cylinder Cell F1 go 110 Bani Cell vanized iron of different widths on; with different coupling arrangements on the outside of the container, were tried. Una cell with tne plates inside tne container, insulated from the solution by laquer was tested. factore onion efiected the sensitivity of the response for this tyne cell were found to be, size of plates, thickness and kind of insulation, and cap- arction of plates. These factors were noted qualitatively rather than quantitatively and further investigation along this line was precludei by other unsuitable features of band type cells. The chief difficulty was to deeign adequate ehielding for a cell on which coupling protuberancee extended. A second difficulty was present in the form of response change due to volume change. Otner studies of this type cell have been reported previously (54,37). no last type cell tested consiated of e polyethylene container vita electrodes imbeded in the bottom (Figure 12). Thin was the cell subsequently used for all quantitative non- aqueous high frequency titrations. ihe advantages observed in this cell compared to the others were, a high degree of sensitivity, probably due to the use of polyethylene rather than glass as an insulating material, a smaller change due to volume change by placing the electrodes at tne bottom of the cell, eaee of construction and enioluin , union was achieved by fastening the entire assembly to the bottom of a heavy 00'913381' cup 0 37 \ I..- ,ucus..uuun...n "ll" ||'l'l|'."l' ""'|"|-'\. \ x x .......... cum .. - ....H..H..-\ e or . Polyethylene Cell and Insulation Copper Electrodes and Shield Polyethylene Base Electrode Cell F180 12. General fixoerimental Hroooduv Solutions of known concentrations were pinata; into tho experimental cell. Volumes of solution from 03‘ ml. to 50 m1. wars used. Adlitioual solvent was tnan pipotol into on titration vessel to increase the voluno so that tno elootrofloa were well below the solution surfaoo. Eolvent was aidofl in 25 ml.. 53ml., or 100 ml. yortions aspending upon the partic- ular cell used and too vollmo of solution already preaont. The instrument '98 then adjusted to proper workln; ran;a for the concentration of solution used. This was don: by a trial and error choice of a series fixed condenser and adjustment of the 01 Variable capacitor {Figure 6). Upon final aljust- ment thono controls remained fixed throubxout tho titratioo. The voltage bias as dobermined by the R3 variable roaiator up to this point has boon zero. The V.?.V.fi. is aijaated to the 30 volt range. The R3 potentiometer is than adjuaooa, while the inatA meat is in oscillation, so taat the voltage meas- . L“ ured is within has proper range“ and 13 positlve or negative) I! #8 depending upon the cnoico of "bowl" or hump Bhopal roaponae. W ‘T “Tue proper rears cnoice is an instance of a priori knowledne, howavor tuis knowledgo is of a genoral sort noich can be detorminel by a preliminary titration carried out with no bias on the'30 volt soalc. in general it was found for tie aqueous titration: too range was rarely greater tqan 3 Volta and for non aqueous titrations the raayo wax raroly graater than 1.5 volts. SSince 0.8. measurements are being male on two V.#.V.7. it is necessary to reverse the polarity each time too re- sponse curve is inverted. 39 ?hus for aqueous titrations the E3 was adjusted so tnat the V.T.V.n. registered 0 or 3 volts, with the instrument in use cillction at too initial point of the titration, and with the oppropriatc polarity for the response desired. For non- aqueous titrations R3 is adjusted so that tne V.T.V.£. re- gisters 1.5 volts with too appropriate polarity. For any one titration twe 33 setting romaine constant. Measurements were made by recording grid toltsgc or CE variable capacitor reading versus ml. of titrant edded. neth- ods of measurement will be discussefl below. During the titre- ticn tne burnt tip was below the surface of too solution tic trated except where otherwise noted, the stirrer was adjusted to the maximum speed at which no solution swirling occurred. fiessurements were taken between as and 60 seconds after addi- tion of the titrant except an otherwise noted. ais was to permit complete stirring of the solution to take place. The size of titrent increments was such that a greater number readings was taken just before and just after the and point. End points were detected by graphing the oscillator charac- teristic versus ml. or titrant. At the end point a break in the graph was observed. Several different techniques were used for obtaining high frequency measurements. One group of related techniques were based on power loco measurements, indicated by change in grid voltage when tnc circuit was in oscillation. In the other group the change in capacitance necessary to return the in- no strumcnt to its resonant frequency was used as a basis for measurement. The power loss measurements are made by observing the following characteristics at constant frequency; "out in vol- tage", the grid potential at which oscillation just begins, I! I! when the eye snapped shut as capacitance is removed from the circuit by means of the variable capacitor; ”cat out vol- tage”, the grid potential just before oscillation ceases when the “eye” snapped open as capacitance is sdoed to too circuit by means or the variable capacitor; ”peak volta;e measurement". the maximum potontial by adjusting the variable capacitor so that s change in either direction will cause s lowering of the potential as indicated by the V.T.V.H. Another power loss measurement records the grid potential se increments of titrant are added during titration without changing the variable capacitor from some pro-arranged setting. In this tyne of measuronsnt the frequency does not remain con- stant. This is essentially a loading measurencnt because cs. penitence and conductenzc sro lumped into a single measurement. Excellont results have previously been reported for this type of measurement (kh,h5). Capacitance change measurements were made by two teconi- ques corresponding to out in and out out voltage measurements described above. Instead of recording grid potential at the points mentioned, the relative variable capacitor settings in dial units to the nearest tenth of a dial unit were recorded. kl These methods will be referred to as cut-in end cut-out capa- citance change methods, respectively. Graphing Procedure Figures 13 throth 2? ere selected greghs of the titra- tions carried out with the instrument described. The graphs represented here are depicted es oscillator characteristic versus percentage of the theoretical amount reutrelized. The choices 13 represerated in unit of percent of the stoichiome- tric end point with the ceziter mar k equel to 1005. A few of the larger scale graphs start with 60% and 80%. The ordinate is represented in units of millivolts except for Figure 16. Both the independent and dependent variables are recorded as relative units. Host of the individual graphs are soothed to 24111113126 6 end point errors. The individual point protetle readinv er~ rors, in keeping with the este listed convention of showing the error es teiug coo)letely contei.ned in t.3e depen;‘1ent vari« able, are shown by circle size in %‘ igtzres lu-l7, 22s-2i, tim cele of the other figures preclude the use of this convention. The uncertainty sh we is the reading umcerteilzty, other sour— ces of error are discussed below. 6Figure 15 shows Loth smoothed and n ed rrephs for the same titrations. Smoothing grs 3h 5 L s is not uncormon for titrations and consists of dleuinh as snooth a curve as possible ttr ough the probable location of the great- est nuznzier oi points. on~sm mot! ill‘e‘u Sources of Error The error of the method studied, is the error of the lo— cation or the end point break with resyect to tee ceprOpriste percentage values located on the choices. Fictorally this error may be reoresented as the algebraic sum of two errors, displacement of the break on a stationary graph and toe dis- placement of the choices values with reeoect to $18 break. iho random sources of error can be summarized as errors of standardization, the drawing error, and the error of the individual points. The errors of the individual coints are reading errors and instrueent errors, such as too change of thermal equilibrium, dilution during the titration, fluctuam tion of tie supply voltage and fluctuation of tee circuit pa- rameters due to stray radiation, capacitance, or otner unde- termined causes. Individual points displaced by random errors did not effect the location of tie end point greatly eince the graph was recorded as a smoothed curve. Determinant errors eere keot at a minimum by a method of cross checks, such as double readings, and grachin; in a man- ner to mfinicize prejudxetcnt. Predictable condition errore such as the cubical expen- nion of acetic acid solutions (3.1L percent per 00) standard- ized at one temperature titrated at anotcer, were either min- imized by making the conditioce constant or makine 1-9’ tee ap- propriate corrections in too calculetione. L3 Eiscucaion of Results ?owcr loss moaenwe ants in &@JEOWS moflia Tao power loss w measured hero was the loaifln; ficacurencnt technique utilizing the undo?" shaped rcgponce. Tao coll usoi vac the polyethy- lene cell of Figure 12, with a paddle stirrhr. rho aoluti 3 consisted of a measured volume of the standardizoi base dilut- od with a measured volunc of colvcnt so toot the total voluoc was greater than 50 ml. The parallcl ca acitJrc more cijostod I0 that too V.?.v.i. rcgistcrco in tic region of too first 20 per cent of the frequency roconcncc curvc {an absolute grid voltage of approximately 5 volts). Inc appropriate V.E.V.A. range was nolcctod by aijuctin; too fl} rcsictor. moo titra- tion was carried out using a standard acid solution with the burot tip below toe surface. Juring tuc Litrction the instru- mcnt controls were fixod. The rosclt was a graph of too vol- ume of titrant vorcua too scale reading of too biased V.T.'.fl. Tao purpose of the aqueous titrations was to tact the sta- bility Ina sensitivity of tgc inctruoont and to dotormino the effect of call geometry and design. These results havc been summarizei under inctruaont uni call charactorictico. fiho cocoons hyfiroculoric acid, sodium hriroxido titrations of Figuro 13 cni iablc I show the effect of working on diffe- rent portions of the ruoooncc curve of Figure 7. In Loncrnl the curves of Figure 13 giro a cynogsis of all prcvloucly rc- portod curvc snapcs for this type of titration (4,23,QS;, from the almost linoar (A) through variooa degrees of curvature (a, 0,9) to the inverted one (3.5). The explanation is tact the a. response voltage 18 a function of condactance. Ina titra- tion curve 18 the locus of a point following the coudactaaoa changes on the response curve of F151Pe 7. Canalaer use ini- tial point of curve A to be on tue left side of the "hump". as tne titration pregresaes and cunflqctance dscrensea, tna locus of the point moves dawn $19 graph and the resgonse vol- tago docreasss until the end point is reacnad. Aftar the and point the conductance inoreasaa and £16 point moves back up tne response curve to the right. Conaequenbly when titrant uddad is plotted against V.T.Y.M. readin a, or in tarms of Figure 7 the vertical position or the moving point, a V shap- ed curve results. TREE-.15 I 12::;.;~.;'§1‘ ‘i‘I'Z‘JATIuIJIi’. 03} ;3.1.: iifai’iifiibu'filc ACID 1113 1.5936 R EGTASSISA 523591133 Figure 13 Normality moq. Takan &eq. Found Dagigtéon L 0.3175 1.767 1.767 0 a 0.3361 8.632 3.632 a 0 0.1927 19.23 19.31 11.5 a 0.3668 36.61 37.00 -13.5 .3 0.7163 33.76 35.85 2.1 F 1.19% 37-h3 37.51 2.0 Similar exvlanationa may be advanced to oxalain the chapel of the curves B,C,D,E,F, of Figure 13. fans all tbs curves of Figure 13 can be considered plota of a moving point to the MS 6.0‘ . n. A O°O}+7SS N z 5.5 ' ~ ’ 5.0+ L h-St ' B 0.08617 N 0.3668 N b) o O N 0 \fl V.T.V.M. Reading Volts N o O 1.5 1.0 0.5 L L l A 20 to 60 80 100% ‘Percent Neutralized Fig. 13. Aqueous titration of hydrochloric acid solution with 1.5936 N potassium hydroxide 1+6 left up to the end point, to too right after toe eni point, on tn "hump“ curve of Eigire 7. the curved gregna (3,0,9 are a result of fine point moving over toe curved top portion of the “burp". The inverted grqnhs (£,F) are a result or the starting point being on the right side of 13 ”hump”. The greatest sensitivity, that is toe greatest change of power loss per unit change in concentration col consequently the sharpest and point break, occurs on too straignt line aide portions of tie reapouae curve. hiemi etion of iigure 13 con- firms this. Curve A is the only one that doen not conform to this general tendency ani tnis is very likely due to he very dilute state of the solution since the change of conductance is snail due to the reduced number of ions involved. The conciderationo given above for aqueoue titrations are analogous to factors present in non-aqueous titretione.7 flow- evor the effects noted are considerably reduced since the sub- stances involved are lees ionizable and coneequontly have a amaller conductance. The eolvent systome taemaelvec are in- herently lens conducive to high conductivities. In general aqueous titrations gave the most satisfactory results for strong acids titrated with strong bases (Figure 13) Ind vice versa. Acids such as hydrochloric, sulfuric, and {Acid base titrations in tueir widest cenco do not neo- easarily involve interaction or hydroxgl and hydronium ions (34,35,ul,u2,60). Since nigh conductivities in tne aqueoua titrations of Figure 13 are due chiefly to trece two ions, it is reasonable to expect tint in toe absence or one or the other or both, he conductive character of too solution is lowered and considerably altered. #7 tribasic phosphoric, and bacon such as sodium hydroxide, and potassium hydroxide gave excellent results. flock acids- such as dibasic phosphoric, acetic, and monobasic oxalic, when titrated with a strong base gave fair results, and points war. not as sharp as those for strong acids. Very weak bascs and acids, aniline, phenol, and moncbaaic phosphoric hitharto unreported in tho literature of high trcquancy titration, when titrctcc with strong acids or baaos respectively, showed no and point break. _?eccr lcsa;§caaurcrccfc in dicgcnc moiia. A series of preliminary titrations were made in dicxanc to dctcrminc its suitability for adaptation to high frequency techniques. Sons comparisons of titrations in each of the respective media arc shown in Figure 14. Uioxana, ano aprotic solvent, gavc a sharp end point where both acid and base were strong, such as pyridine (Figure la curve B) titrated with perchloric acid. Hook and intermediate basos, such as aniline, when titrated with parchloric acid (Figure 1% curve 1) gave irregular graphs. The titrations represented in Figure 1% curves A and B were carried out at the optimum concentration (approximately 0.01M) for these compounds in dioxane. Otnar high frequency titra- tions in diozsnc (not shown hero) verified these results 1.9., diohonylguanidino a strong base gave a relatively sharp and point br ck, p-toluidino a weak base gave no end point. Power loss moacurcmcnts in acetic_scid modia. Titration- of aniline and pyridine at approximately 0.013 concentration were carried out in acetic acid media (Figure 14 curve C ani- w as 3.0 0 a 205 N O l I 0 0 0 0 V.T.V.M. Reading Volts H . 1:1 o‘ 2* $7 . O.5~ l L B 0.00330-N 0.002h1 N Aniline in Dioxane \‘ 4‘ \‘ Pyridine in Dioxane 0 O Aniline 1n Acetic Acid '1 c é. o 0 0 0.008915 N Pyridine in Acetic Acid j (I 3 3‘5 0': 0 23 6 ¢ ‘ I) J .3, ’l ‘ ‘ G 0 \ ‘t‘ \ ‘a o k‘ x‘ A 20 #0 Fig. 1%. 60 80 100% Percent Neutralized Titrations in Various Media 1&9 lino, curvo D pyridine) for purposes of comparison with dio- xano as a solvent. It will be noted tat the magnitude of the and point break for pyridine in both media is approximately the same even though dioxano measuronent was taken in the 0pm timum concentration range and acetic acid measurement was not. Comparison with titrations carried out in tno ootitum conoone tration Ponies for aniline {Figure 23 curve A) and pyridine (Figure 22 curve 0) show much sharper and points in the acetic noid media. Conventional theories of solvency state that acetic acid as a solvsnt causes a greater degros of dissociation of weak bases than éoos voter or dioxsno as a solvont (3%). This acetic acid systems allow greater conductance changes during the course of a titration, giving I sharper and point. ‘hoo- rotioally and oXporimontslly soetio eoid systems showed a wi- der range of applicability for weak bases than did water or dioxano systems. For this reason it was decided to confino txo invostigation of the titration of weak bases to too soon tio acid media. Une limitation of acetic acid was its power of solvency. Some substances such as aluminum oxinato could not be titrated because of limited solubility. Conoupvont with the study of solvent systems was an in- vastigstion of the various possible meaairomont techniques. Some techniques worn superior to other: and altnoagh each sol» vent ayotem was tested by each possible technique, tfie pre- vious fiisoussion was condensed uni confined to loading meas- uromonta for rsasans which will become Apparent in the dis- cussion of measurement toohniquos. Various operating procedures are feasible as mentioned in the section on exocrimontal methods. The previously das- cribcd methods of out in voltage, cut in capacitance, peak voltage, and locfiin; measureficnts were utilized for the ti~ tration of sofiium acetate, B-hydroxyquinolino, exile-alanine, with percnloric acid as the titrant in acetic acid media. (A,B, sad C cloves rosgectivoly for figure 13,13,17 and 13). Cut in voltagg_ocssurencnts., iho general proccdcro for this measurement involves the sacs preliminary steps used for loading measurements. The rosoouso measured, was too scale reading of the V.T.V.fi. at tgo onset of oscillation as indi- cated by the nodden closure or the 6L3 eye. This measurement was obtainod by increasing toe capacitance, using too 02 var- iablo capacitor until the instrument was out of oscillation, then carefully aijusting the 02 capacitor until oscillation just startcdg when the voltage rcafling was taken. Curves of too out in voltage (Figure 15) etc: tie oz- trome irregularity and a complete lack of on& point for tnis Operating procedure. Th6 curves are sicwn by a cotted line connecting tie points, but a more accurate representation comonsursting with too uncertainty and irreprodcolbility of the incividusl points might be g’ven by tQB solii curves of Figure 13. The cosngo of voltage at t.o and point break was of the sccc mstnitude as the uncertainty of too voltage read- Bihc stoppin; of oscillation for each reading is an un- desirable characteristic of this methoi and alliod netbals, since two instrument is not allowci to reach equilibrium. 51 1.8 1.5 1.2' V.T.V.M. Reading Volts '0 in o o 0‘ 0.3' (:> A 0.02863 N Sodium acetate B 0.01776 N 8-Hydroxy. quinolins \ I Q 7" 6“ 0 0.01702»: \\ V fl-Alanine 6 \ 20 no 60 80 100% Percent Neutralized Fig. 15. Cut in Voltage in Acetic Acid Media With 0.1132 N Perchloric Acid in; (on the 39339 of 250 m.v.)- At tha 3:393 3f oscillation for the first 0.3 unit of '93 Oz dial (333933. 0.02 mm?) 1 o Volta e incrc 3:3 was on the 09339 of 2;5 valia. 333 to the relatively coarse dial cont9317 cowgrred to the vulta 3 339356 at the ouset of 039 cillation, it was inpassiblc to t3 3 3 vol- ta. 3 reading uni Spool f3 it as the 33 t in voltage. A farther defect was an up: Jerd 0933p C1 the Volta o w3ich levulo& out after 90 minut38.13 Cat 13_09390,*9990 inns: 9W jl_ fi;. """ '6 13 3:13 Tubhe II show the results of tlw 9323 tit Ha ions usin: tha cut in capa- citv: 03 tecanfnxc. 1‘0 n9ficeq 96 1:."0.Wed for tiis measurou maz3u £3 idea 33.331 to £2123 3990:3193 of tha cut in Vultfibe mans- *1. u939.nts 093:3: t at V .3 (2 V391313 39 393-339 sot Lin ,13 3L0 3333335.: :t variable rather than the measured voltage. Lnficr ideal conditions of concefl'rwtiafl 3:31 points were observed using thia Op‘m ntin" procedure, how3999 many of Lie defects 2‘9 prasent in out in voltage 1323390103t are 813 32233333 L393. 7328 reactinn time of tha cparatur tarni _ the 3131 at different 333333 H33 9 139"3 f9,tar L393. Pz3931313 3 CL «taut apead noc“"=Lcnl d: ive 37; 3 aicail, 0,1991 3 by txe iuszru- 9.3.9:: may have 3039:9909" tr; 3 93331:: 916332. 61159334319 9, .i..o.:3..3r 1t 13 tHOu"L t at SWTfIc r 3% L9 {1933 11 tla mc99 L9aiu exists t3 make even this 3:333:9170. fincthar 3:33.1L16 solution to this problaw of reiining a on 933 u41u3u::.t, 13 to use n 3391139 C 639863 tar, howeve9 tLLfl 933111 elfactively out dawn tic range of tLe 33Lrunrut and titratiota involv.iag 13933 cLangea could not be maue. LJ ;'~:' \. 10 ;';e upwrrd cli b of 93934933 90119 e la equivale m1t to a daunwurd 3:399 of “rid volt9ge. ';.3 £130 (3 ) 01. t.’ 3 climb 0? vanamgwsd Voiltalgfi 8 2-: 9331990331 1.L.L.‘C:}.S;.Dl'§ of L313 :6 unmade of the voltage. .C2 Capicitor Reading (0.0? mmf) 53 A 0.02363 N Sodium Acetate B 0.01702 N fl-Alanine 0 0.01776 N . 8- Hydroxyquinoline 1 20 no 60 30 100% Percent Neutralized Fig. 16. Cut in Capacitance in Acetic Acid Media with 0.1132 N Perchloric Acid. 54 ".5 .3 «v YT 1-; .11... ~4L¢b v-Tn-:. ' Hr ’“fil .- wrman ya P‘éQflVMJ'UW ‘1. an 11031.3 110 AC .: ‘4 U R3324] «Add, A Bf U53 1 -..-.J:..:..>z. .1 .“'»A..._~.'-'1...,v..'-..\s.;'19" 3. Ticafil.bws a: T1.30510210 A110 .1132 u 1; Figure Subatanco floraality Bag. Taken fieq. Found Deviation Titrated Popoto 15 A 30:11.13! , AOBtatQ 0002353 10713 1072; 60% B 8-Hyfiroxy~ ’ “ quinolin 0.01770 1.001 1.07% 9.3 C dd-Alanlnu 0.317;)2 10019 1.110 8303 17 A Sodium _ , Acetate 0.01702 1.01) 1.021 1.9 O ‘A-Alanine 0.01031 1.131 1.131 0.0 13 A Sodium _ , _ . Acetate 0.03357 3.§2;D 0.0303 2.3 B 5-Hyfiroxy- , quinolina 0.1003 5.30? 5.332 2.3 C ‘fl-Alanine 0.01702 1.019 1.019 0.0 It will be notad from Figure 15 and Table II that tns and point breaks in terms of dial unita wera very small (10.3). since the uncertainty of rewrofllcin a dial settin,; was-+0.1 the end points obtained were unreliable. Ike time required (B/Q to 1 hour) to carry out a tit1~ation using has method of cut in vol- tage or cut in capacitance is a disadvantage of the metgod. fihe techniques (not raoordci have} of cut out voltage and cut out capacitance, in which the procedures are ifientical with those above excM t ti‘xat tne goiut r60021ded was 311. before 03- cillation 003101, suitor t.a aama éefects and gave less repro~ duciblo results. \ 11 \n Loading measurcxeuts. At constant conductance the grid voltage passes through a typical parallel resonance curve with respect to changing capacitance. Any portion of the re- sonance curve could be used for the titration rcSponse if the frequency were kept constant. In practice adjustnent of the capacitance is not fine enough to keep the frequency constant except on the flattened Toxins of the resonance curve. Due to this instrument characteristic two types of measurement are possible. The first is to allow both the not cell caps- citence and conductance to change. The measured voltage is then a function of the lumped parsoetcrs, he so-called load- ing measurement. The second is the so called peak voltage reesureucnt, 1.e., the measurement of the maximum grid voltsge after each increment of titrant is added, keeping the fre- quency constant by adjusting the CZ capacitor. The capacitor adjustment is not critical since the frequency at the maximum grid voltage remains constant over a range of 3.0 capacitor dial units. The resultant graphs of the titrations of both methods are reproducible and in most cases gave good cud points. Since the instrument was allowed to warn up for over an hour and was not out of oscillation at any time, thermal equili- brium was maintained. Figure 17 represents txc results obtained for loading measurenents, using the procedure of page u3. The V01t$¢9 range utilized was the first 20 percent of the resonance curve. Curve A of Figure 17 is one of a series of reverse titrations carried out in this study. In this example the titrsnt was f" c> t 0; v1 Reading Vol 3 u: C) 0.5 56 T 1 ‘f a A 0.01702 N Sodium acetate \\ 0 \‘D \\ o \ 0 0.01031 N \\ fl-Alanine 0 e \ Q ‘3 9 \\ 0 G G \\ fi \‘ 6 l L I I l _L _ F180 17. no 60 80 100% Percent Neutralized Loading Voltage in Acetic Acid Media with 0.1132 N Perchloric Acid sodium acetate and the substance titrated perchlorio acid. The onohaaizod break mentioned above war not noted in the con- ventional titration of sodium acetate with perchlorio acid. In gonoral not all reverse titration: showed thia cognaaized and point break. From the standpoint of conventional titra- tions with perchlorio acid the reverse titrations for both loading end peak voltago measurements were less satisfactory. Peak voltago moasggereqfigg Tgis is a conductance moan- urement technique utilizing both "homo" and "bowl” shaped roa- ponoe curves. The titrations listed here have all been oar- riod out using the cell of Figure 12 utilizing a paddle stirrer as previously described. iho solution titrated consisted of a measured volume of the standardized base diluted with a measured volume of solvent so that tie total volume was great- or than 50ml. The parallel capacitors were adjusted so that the V.T.V.M. registered the maximum (about 13 v) or minimum voltage possible depending upon the response used. The appro- priate V.T.V.fi. range was selectoi by aigustins too 33 resis- tor. The titrations were carried out using standard perchlo- r10 acid as ti rantoll Baring the titration the Cg cagacltor was adjusted after each increment of titrant so that the max- imum or minimum voltago, as the case may be, was registered. The resultant series is plotted as a grayh of the percont ti- trated versus the scale reading of the biauod V.T.V.%. illsevoral titrations,AaoThotgd;Aaro so;called reverse ti- trationa, that is, titration of perchlorio acid «in the appro~ priatc base as n titrant. \n O) The graphs of 11; are 13 are titrations of (A) nodium acetate U ) 8-h3droxyquinoline, and (CJ‘fl-alaninc using the pod: voltage motLod of moaocrcncnt. Acco1tcble 021d points are present except for curve C and this is probably due to tl'ze very low ionization of,a-alc1:ine, (on the order of emporison of a large number of titrations carried out using both peak voltage flAd loading techniques {BATS co ..1par- able results with reference to accu'acy and reproducibility iowevcr in zost ca cos tho and point break was noticeably less sharp for he loading measurements. Comparison of curves C of Figure l? and 13 shows this for the case ofli-alanine. This leveling effect was caused by allowing the capacitacce to very during the titration. Several exceptions to this leveling effect were noted in some loadiig measureuout titra- tions. The harpcr t:'7 n ordinrqr break 01‘ Figure 1? curve A is an example. Peak voltage moasurcucnt was c11059n in favor of loading moasurcucnts for this study. The reasons are, slithly sharp- er and points obtained by pea 1: voltage measurements, and the fact Lat the: core concentrated solutions tend to load the tan? circuit out of oscillation. 1 ffect oi concentratiOh. series of titrations on a variety of weak bases was carried out in acetic acid solution, perchloric acid being tgc titrcnt. a purpose was to check the characteristic curves obtained for any one substance and tle chcnbes that occur in tLeso curves when the solution con- centration is varied. V.T§V.M. Reading Volts 1.00 O ‘3 U1 0 \n O 0.25 59 F A 0.00857»! . Sodium acetate. B 0.1065 N '8 Hydroxya quinoline C 0.01702 N flsAlanine 80' 100% Percent Neutralized Fig. 18. Peak Voltage Measurement in Acetic Acid Media with 0.1132 N Perchloric Acid The first acriaa of Titration: were carried out on the base 8-hydroxyquinoiinc as represented in Figures 19, 1): ‘nd Tabla III. 8 0203;1fiszqulnn Tlrnswleua IN 00 TIC 1010 06211 WITH YABLE III .1132 H PxfiflfiLORIC ACID Figure normality Koq. Taken fioq. Found Deviation P.p.t. 19 A 0.1065 5.307 5.292 “92.8 “‘ B 0.05261 2.655 2.655 0.0 c 0.03.633 0.5314 0.5314 0.0 l?a 1 0.002453 0.159 0.1602 5.0 3 0.001770 0.1063 0.1053 ~9-h The most concontratcd solutions, curve. A and B Figaro 19 show a characteristic dip before tha and point, most pro- nounced for curve A. This is caused by the formation of I precipitate which cftcctivoly removes the participating ions from sciatica. Curve 0 further indicaaaa this tendency al- though only in I very.mild form. During the titration repro- sented by curve C the prficipitatc tormod was evident by only l alight cpnlcsccnco whereas in titrations A and B large cry- stals formed. The dip was caused by the auflmation of two non~linclr voltages, onc increasing the other decreasing. The dupth of tho dip depends upon the rate of change or the two voltages. The tendency observed for Sonydroxyquinoiinu is tnat tno and point breaks become more obtuse as tno concentration de- creases. The titration: of Figures 1? and 19 a were carried 61 A 0.1065 N Q 103’ CD 102’ ID . 4.) H O > cm - 5 B 0.05261 N (3309" 0 CI: 55 :; s: > 0.66 0.3» 0 0.009683 N 80 90 100% Percent Neutralized Fig. 19. Peak Voltage Measurements of 8 Hydroxyquinoline in Acetic Acid Media with 0.1132 N Perchloric Acid 62 A 0.002058 N |-' O N 1 g Veits O o \0 j V.T.V.M. Headin O o 0‘ 0.3 L l 1 20 ,00 60 80 100% . "' EPercent Neutralized Fig.519a. Peak Voltage Measurements of 8 Hydroxy- ' quinolinc in Acetic Acid Media with 041132 N Perchloric Acid 63 out on the right side of toe hump”. in curve C of Figure 19 {Sufi-hp” wan and curves A and B of Figure 190 the top of the approached from the riQnt as is indicated by tne snaps and direction of the curvature. Tnc decreasing cwarpnees of the curve at the lower concentrations on the right side of the 'hump' corresponds to the higher error level of these titra- tions shown in Table III. "cc left side of too "hump” with the corrccpcnding increase of sensitivity evidently occurs at concentrations too low for practical analysis. The more re- liable results occur at higher concentrations, howevsr a spec- ific optimum concentration cannot be predicted mince the chcrp~ non. of the end point break due to sensitivity is obscured by the precipitate dip. Figure 20 and Table IV are titrations of various concen- trations of aniline in acetic acid media using perchlorio acid titrant. 133131.23 I ‘1' c 101L102 1103141003 IN 102110 10:0 1011 11:3 .1132 N PARCHLOVIC ACID . g ‘ . i , Figure 23 Hormality moq. Taken 00.. Found Deviation ' PQPOTO A 0.05200 2.603 2.669 8.0 C 0.00525 0.2630 . 0.2593 -lu.2 D 0.00175 0.1009 0.1033 ~10.0 fiagncr and Kaufman have reported results for tnia parti- cular system (65) using the technique of frequency cuan;o 6h 0.05288 :92 1000 "’ \\ 0* B 0.01765 H \ ”0°75 " , \\\\ :3. v 0 .x L3 ‘ . é no ‘ I G H "U :5 G) a: $ 530 50 .. : 3 «.Q E4 . > c \\ § 0 0.30525 N 0.25, _ D 0.00175 M I 1 I 60 80 100% , percent Heutralized Fig. 20. Peak Voltage Measurement of Aniline in Acetic Acid Hedia with 0.1132 fl Perchloric Acid measurement. ihe relative sharpness of the and point breaks are approximately the came for the two methods. The greens of wagner and Kaufman show a more lin er relationship of the response tncn do the 5P3 he on Figure 20. She relative accur~ cc: of the two methods will be diecueeed in the final section. In the curves of Figure 23 titrations were made on both sides of the "huep" checcd response curve (figure 7). The curves may be explainec in terms of changing conductance:- The K3 of aniline is small and the ccniactcnco change duo to he removal of the base by titration is neglibible. The con- ductance change due to the salt which is appreciably hignly dissociated is relatively large. Baring a titration, up to he end point. the conductance for dilute solutions increas- es due to he formation of a salt more icnizable than the orig- inal beec. After the end point the ccnductance increased at a greater rate hen before due to adfiition of the excess acid Iiich has c higher conductance teen the salt. In terms of voltage from the right side of the "hump" (Figure 12) the vol- tage decrease. up to the end point at a certain rate and de- creases at a faster rate after the end point. cince tce over- all conductance change depends upon overall concentration change, the more concentrated Iolutione will have a larger total vol- tage change. Ehis can be seen by cornering curVea 8 ani C of Figure 20. The size of tce end point break depends upon the difference or tne coniuctance caange before uni utter the and point. 6o Curve 0, show: even less or an overall voltage cnange in diition to having a V snare caused by a crossing over tno ”hump” at the and point. Titration A departs from tnc above considerations of shape and overall slope in that the concen- tration of the solution is high enough to suprcss tie ionisa- tion of the salt formed. In this respect it is essentially a non aqueous buffer solution with respect to the salt. These observations verify the fact that aniline is a weak base in acetic acid, and postulates that aniline porchlorate though not a “weak” salt is not completely ionized in acetic acid. Aniline in acetic acid is a relatively weak bass, tnererora an Optimum concentration for titration as such does not exist, however a greater overall coniuctance change occurs at too higher concentrations. A series of titrations were made of various concentra- tions of hexametnylcnedisnine in acetic acid meiia vita por- chloric acid as the titrsnt. the results of these titrations are shown in Figures 21 and 21s and Table V. The hcxsmetny- lencdiamine titrations are straignt forwaro in every respect, the salt is strongly ionized, the base modorstoly so, and there is no precipitate formation. There is but a single and point break, this indicates that the basic groups are indistinguish- able sni that hexamcthylcnedismino acts as a mono—acid bass. The overall grid voltage cnange in the dilute curves is toward increase in conductsncs on tnc loft side of tn. ”bowl" for Figure 21 curve C and D, sni Figaro 213 curves A,B, and C. This indicates that the conductance increase prior to tno 67 1.00 O 0 .KJ U1 f V;T.V.M. Reading Volts c> {fl C) T 0.25 B 0.02707 n 0 0.009170 n ‘ D 0.00500 H. A l 90 100% Percent Neutralized Fig. 21. Peak voltage measurement of hexa- methylenediamine in acetic acid media with 0.1132 N perchloric acido 68 1.00 V) 30.75 0 > 00 I: H 'U 6 0 m 57 EOOSO 5" ‘.> 0.25" A O . 001980 ll 0 0.000360 u ‘ j A l 1 20 .hO 60 80 100% Percent Neutralized Fig. 21a. Peak voltage measurements of hexa- methylenediamine in acetic acid media with 0.1132 N perchloric acid. 69 end point, due to salt fonnation is greater than the conduc- tance decrease due to base titration. The end point of curve B Figire 21 occurs at tne minima of tne “bowl” and conscquent- ly displays no break. Titration A rigors 2i occurs on tne right side of tnc ”bowl" and consequently displays a V snags. It can also be noted tnat the degree of curvature remains rec latively constant wnile the sharpness of too breaks decrease with decreased concentration. Optimum concentration for ti- tration occurs in the more concentrated region. TABLfi V. HEXAfi-T?YLEHsJTA.IHJ TIIRATIOQS 13 AC£EIG ACID EEDIA 01TH .1132 H ?EHJJLDRIC ACID Figurc molsrity Mmo. Taken HmogiFound Deviation P.p.T. 21 A 0.0509 2.736 2.730 -.8 0 0.02707 1.375 1.370 0.2 0 0.00917 0.5034 0.5463 -3.6 0 0.00500 0.2700 0.2700 0.0 21a A 0.00198 0.00930 0.09970 5.0 0 0.000661 0.03956 0.03050 0.0 0 0.000360 0.01976 0.02016 20.0 Figurcs 22 and 221 and Table VI represent a series of ti— trations of various concentrations of pyridine in acetic acid media with perchloric acid as the titrsnt. The curves of tnese Figarcc snow the typical dip prior to the and point doc to the formation of a precisitate (65). Inc dip is most 071- dcnt in curve C of Firuro 22, less so in curve U and vanianes (g. ‘ 70 63:“ . 0 . 0' 0' ' A 0.09800 N Loci 0 «v c a: q g Q {a 0) 490.7 - '3 5 B 0.00896 N ." > to r: .r—q '6 w l a) D: . c E. 3.0.50 '- E4 :- 0 0.01636 N 0.25r D 0.008915 N, 80 90 100% Percent Neutralized Fig. 22. Peak voltage measurement of pyridine . in acetic acid media with 0.1132 N perchloric acid. 0 B 0.0016u6 N ~‘ 0 / 8 G 0 0.000898 N_ {g 0 0 l l A_ 1.007 00 4.) H O b 00 50075 *' 'U a 0 0: >1 5-! >¢0¢SOF 0.25’ 60 80 100% Percent Neutralized Fig. 22a. Peak voltage measurement of pyridine in acetic acid media with 0.1132 N perchloric acid. 72 completely in the graphs of Figure 22a. The precipitate dip, of curves A and E, Figure 22, are not shown since the abcissa coordinate starts at 80 percent. All the titrations except C of Figure 22a had their initial inception on the right side of the "hump” shaped reaponee curve. As the concentration de- creases for this series of titrations so does the sharpness of the and point. Optimum concentration here is in the more cone centrated regions as it was for 8~hydroxyquinoline. TAILL VI. PYRIDIEE TITRATIOfi IN ACETIC ACID HEBIA'HITH .1132 N PERCHLORIC ACID Fmguro Normality Meq. Taken fieq. Found Deviation P.p.‘l‘_L 22 A 0.09800 n.883 b.9126 0.6 B 0.00896 2.010 2.0300 1.1 C 0.01636 0.9790 0.9790 0.0 D 0.008915 0.0880 0.0978 2.0 22s A 0.000930 0.2022 0.2022 0.0 0 0.001006 0.09850 0.09358 ~5.0 0 0.000898 0.0h919 7 0.00722 «0.0 ‘_ The graphs of Figures 23,~23a, and 23b represented in Table VII are the titrations of various concentrations of di- ethylaniline in acetic acid media with pcrchloric acid as the titrant. The titrations of Figures 23 through 230 show the results which would have been predicted except for curve E Figure 23a. These curves, "tent" or V shaped depending upon whether the left or the right side of the "bowl" curve was used to show typical titration curves resulting from the con- 73 A 0.2397 N V.T.V.M. Reading Volts C) Ci. 0.2 0.1 ,a x 80 90 100% Percent Neutralized Fig. 23. Peak voltage measurement of diethyl- aniline in acetic acid media with 0.1132 N perchloric eeid. 7h V.T.V.M. Reading Volts 002 h 0.1 B 0.00896 N -0 0.01636 N D 0.008915 N l 1 l ' 80 90 100% Percent Neutralized Fig. 23a. Peak voltage measurement of diethyl- aniline in acetic acid media with 0.1132 N perchloric acid. 75 0.5 - 0.3 " 0.2' ‘ A 0.000930 N O‘ 0‘ B 0.001606 N 1¢ 0 0.000398 N 1 a l 60 80 100% Percent Neutralized Fig. 23b. Peak voltage measurement of diethyl- aniline in acetic acid media with 0.1132 N oerchloric acid. 76 ductsnce changes when a medium strength base is titrated with a strong acid and the salt tossed is soluble and highly ion- ized. The titrations of Figures 23 through 230 could be term- ed the titrations of a well behaved system. TABLE VII. Didififtfifiibisa Tifaatldflfi IS ACsflC AXIJ e_DIA £123 .1132 H ?§fidflL&RIC ASID figurermeoreslity seq. Taken Meq. Found 0.91;;10n P.0.f. 23 A 0.27)? 9.099 9.099 0.0 0 0.09100 5.20.10 5.2.2.0 0.0 0 0.01073 0.5405 0.5005 0.0 233 A 0.09300 0.833 n.893 2.0 0 0.00336 2.055 2.t53 0.0 0 0.01636 0.9790 0.9701 ~5.0 D 0.003915 0.4300 0.0390 0.0 23b A 0.000930 0.2070 0.2050 98.1 8 0.001606 0.09350 0.09830 0.0 0 0.000393 0.00920 -.0.0037 ~1o.1 23s 1 0.000393 0.00920 0.00959 0.0 8 0.000893 0.00920 to E.P. so E.P. The graphs of Figure 230 were made for the expressed pur. pose of showing that e perfectly linear response can be obtain- ed with s well behaved system provided test the response range is so chosen that the straight side of the "bowl” curve is utilized. In Figure 230, as in Figures 23 and 23s, a decrease of sensitivity results sits a decrease in concentration. Ti- tration B of Figure 23a displays a snail dip before the and point, the cause of which is not known. Inis 010 could not be duplicate}, and all subsequent titrations of tnia system at and near this concentration, yielioi graphs whose appear. once is a graph inflio ted by the dotted lino replncing the dip. A plausible explanation of tnis effect is, tact at the tine of macsuronont there was s variation in cupgly voltage. The curves of Figure 230 arc of titrations idcntical‘ ' sith tnc C graph of Figure 23b except for the adiition of too neutral salt, cofliun perchlorate, which was sddod for the puro pose of determining the rcsult of lowering tne total rosin- tcnce of tn. solution. In curve A Figaro 23c tno solution in 0.022? N eofiium perchlornto and curve B in 0.0455 R sodium perchlorate. The addition of a small amount or s neutral salt to the solution titrated in curve C Figure 23b raauitcd in s decrease of sensitivity (curve A Figure 230). This lesstning or the and point break was generally noted whore the concen- tration of the neutral salt was not sufficient to cannon os- cillation completely. The more concentratud the neutral salt in the titrated solution the greater the uncertainty and ran- com apreaiing of individual points (Curvc 8 Figure 230). ngect of bn50 strongtg; Titration: in acetic acid medis using conventionsl methods or end point actuation have been in- vestigated by s great number of workers. A for of tneae in~ vestigctors have made an attempt to put acetic 0010 systems on the 88mg metdamatical rooting as aqueous ayntefiio (hi to #3 78 ( rm. tin T‘ C *‘L J. v1 o V;T.V.Mo 000455 N A 0°0227 N NaClOu 80 100% Percent Neutralized -nk volta e ncasareneit of O vita-001132 N D 0 . -. ~\ r. ~ I " ‘ '7 1 S .r’_ l t {)1 3.918 ‘. .L ' L t «arcnlorj 3 7401' 79 and 63). Tnis 0006001601 clunlfioation of acetic told-walk but. Iyltonn though nut oomplotn 0nd or quontionnblo ubaolnto vain. (S9). 0110': In ordering of the halo: titratod in this Itudy, from strongest ta weakest. With tun blIOI tans ordor- 0d according to strength, titrations and-r identical 00nd:- tionl should show charncteristio differences due to tho but. 13:01! and dittoronoou 10.005roe of ionization. Tho ordonod “nations of burn are shown in Figurau 21;. Zip and T0010 VIII. Thea. titrntions havo boon carriod out using tn. pout voltage technique in nootio meld madia with porohlcrle 1014 " the titrant. CI=5’~313AI‘1IE’~3H 05-“ TITRATIOHS 09‘ 3.3.3135? 0F DIFKffl-LNT 33.43-200'13-{3 IE AGETIC ACID QEDIA 31TH .1132 N PEHCfiLURIC ACID Baa. Normality fioq.Takon fioq.Found Deviation Titratod yOEQT. an A Pyridine 0.01636 0.9790 0.9790 0.0 B Uiathyl« .3111“. 0.008915 0.h390 0.#890 000 C fiexnms- thylanac diandno 0.009170 0.5686 0.5680 0.0 an. A mun. 0.01765 0.056 1.066 9.6 B p-Alanlno 0.009290 0.5090 0.5905 1.0 O Urfi‘ 0.0091?5 005030 No 30?. lfiogopt‘ Campnriaon or the titration ourvoo than u progresalvuly poorer and point brunt no the bnalaity docreauoa until no and point at all 1- ovidont for tn. urea titration (ourvo 0 Fig. urn 26.). Curv. A of Figure an that: :06 typical tent chap. V.T.V.M. Reading Volts 80 ’1 C) A 0.01636 N 1.0_ Pyridine (3 0.8L > 0.6» B 0.008915 N Diethylaniline 0.6 0.2 0 0.009170 N Hexamethylenediamine 00_. 80 100% Percent Neutralizcd Fig. 20. Peak voltage measurement'of base strength in acetic acid media with 0.1132 N perchloric acid. 81 100 "' 098 F” V.T.V.M. Reading Volts C) 0‘ 0 ;r 0.2 * 0 \\ ° .A 0.01765 Aniline 0 \‘ 0 °\ \ ° \ é \‘ o B 0.009290 N grAlanine 0 0.009175 N- Urea 60 80 100% a _ Percent Neutralized Fig. Bun. Peak voltage measurement of base strength in acetic acid media with 0.1132 N perchloric acid. 62 curves of the right side of the "hump” while curve B and 0 char the tent shape of the left side of the "bowl". ihic in- dicates a decrease of conductance up to the end point and in- crcase after. A: the bane becomes progressively ueaxer the change of conductance is no longer dependent upon tne change of concentration or tuo base but ratnor upon two concentration of the salt formed. This trend in its extreme is 00000 in Curve A of Figure Zha onerc tno right side or too ”bowl” shaped curve producee a typical V shaped curve. Curve 8 of FiédPO at. shows the V shape of tnc left side of the "hump”. Titra- tion A Figure 2% is an extreme example or the precipitation formation dip. The titrations although done at Optimum con- centrations, show progressively poorer and point breaks as the base becomes weaker. Other methods of anq_point detection. iho titrations of the acetic acid systems discussed from the standpoint of high frequency and point detection were alao carried out using con- vantional methods, some were carrioi out using a beat fro.uency method. These analyses by oticr metnoda were made for tne pur- pose of determining tno accuracy of the high frequency.metnod studio; here and for tne purpoae of comparing the quality of the end point obtained against tee and points of conventional metwods. Conventional conductomatric titrations vita a bridge were attempted on acetic acid systamc. In a for cases detectable and points were obtained, however the breaks were small. Ehe bulk of the non-high frequency titrations were pot- entiometric titrations carried out in the usual manner (30, h1,k2,o3,50,56,63). It was found that s grunt majority of ti- trations cited hers gave satisfactory titration curves poten- tiometriosily. (Figures 25,26, and 27 are selected ssscpies.) Several titrations in acetic acid meaia were carried out wits a thirty megacycle beat frequency, high frequoncy device (A?). The results were inconclusive, in comparison to the pov~ or loss titrimetcr, some titrations had better breaks, others poorer. A plausible explanation for tuis, drawn from the gen- eral characteristics of too two types of instruments, is that within a certain ran;e, too best rrcquency instrument at 30 megscycles is more sensitive than too power loss instrument at 3.6 mogaoycloe. Sinco toe range of too power loss instru. mcnt is larger than the range of the beat frequency instru- mont, the above Apparently anomalous results are exoiained. Summary uni Conclusion osny factors are capable of affecting to: quality or an end point. Some factors and their effects are quite obvious cad corrections can be made to minimise the errors toey may cause. Expansion of the solutions with increased temperature, and predictable instrument changes, are correctable factors. Possible random errors have boon minimized by the copro- priete treatoent of data, however the conventional metgods us~ ed as refsrcnoo methods sro relatively unprooisslz tuereforc lZToe precision of acetic acid media titoations.nss been variously reoorted as +0.23 for too potuntiomotric method (32, im am 10.» for tile Indicator methods (32). 8L: \\ 500 f . a fi ’00 F . - 4 A Potentiometric 0 . ; c 0 H -. a. :1300 \ C “ :2 o _' ‘ a . “ r3 “ 200 . . ‘y B Peak Voltage 100 r ‘“ l L ‘ 60 80 100%: Percent Neu3rulizei tho 250 Comparison of paxk volt gs an: potential measurem-nts of 003170; N aniline in asetic iCid fledifl With 001132 N puruthPiC 3C1; 85 & é - 1 o o - é - -- ‘a , 1§ ’ I ‘. ..... . . .. A rocontiomon 10 G. ooah ‘ \ In . Q. 1| or“ O o O‘ r . B Peak Voltsbe ‘ Q .Nqag ‘e - '0 I. “ Potential move 0 0 .p I a. tfi .. *' I. \‘ ‘3 \‘ O o 2 " I. D. a. \‘ - - ‘. l l 1 60 80 100% ' Percent Neutr;lized Figo 26° Comparison of 'eufi'roitwge an; potential aegsusements of 00009633 N Seagdroxy;ainoline in acetic geld ueuig nltu 001132 N percnloric acid 86 1.2 . ' W 1.0 p ..A Potentiometric 0.8 :- :>° - t - 2 = : .. I g 6 - b E? f ' t‘ H . “0.6' B Peak Voltage I; ,- . \‘ C‘. \ ‘3 ;- , e C s i ’ K‘ . CL i, . Q 1 .. I/ If. " a. “ ./' / I 94 . a I . . ‘ .1“ I. O.” .. ' \ 2". ,7 ' f/ If ‘A f ,_ , 0 . .. ' 002 b. at - C " _1 1 J 60 80 100% Percent Neutralized Fig. 27. Comparison of peak voltage and potential measurements of 0.008915 N diethylaniline in acetic acid media with 0.1132 N perchloric acid. :37 it should be oXpectcd that this be reflected in the precision of the data of the present stidy. A calculation of average and standard doviationa13 on tno incorrect aasumption that the reference methods are precise yields a result (average devia- tion is 0.2hfi, a'ia 0.7%) which compares favorably with the standard methods. Two factors found not to affect the and point in any way, are immersion of the burst tip, and the magnitude of the bias of tho V.I.V.E.. Though no effect was noticei by changing these factors, must titrations were carried out wifii tie burst tip immersed and a bias sufficiently great to give the "hump" shaped response for reasons of manipulative convenience. Stirring speed was not a critical factor if boating did not occur. iho glass pedals stirrer was found to be suoerior to the magnetic stirrer which had a tendency to decrease sen- sitivity. The method or meacurement is a factor affecting the qual- ity of the and point and can be discussed by comparing the ad- vantages and disadvantages of the six methods studied. ?ho loading measurements and peak voltage meaaureaonts presented a greater number or aflvantagea in the acetic acid system than did the other four methods. ihe cut in and cut out voltage mrtnojs, though satisfac- tory in aqueous media for strong acids and base: where tufi 13Calculations were made on the A? peak voltage titrations rather than the total of 213 titrations performed, since a great many of these titrations were of an exploratory nature. 38 overall voltage change was large co pared to tnc uncertainty or the indiviiuei points, could not be utilized for weak ba- lce in the acetic acid system since he uncertainty of tne readings were of tne same order of magnituae as tno and point break. Tn. cut in and out Oct capacitance methods gave eaticfuc- tory results in aqueous manic for strong uni intermediate acids and bases. In the acetic acid medic these two methods showed snail and point break: for strong and intermediate bases, none for week bases. The uncertainty of the reciings, though of considerably ices magnitude than the and point breaks, are still a large accuracy reducing factor. A furtner'cncountar- ed difficulty, where tie oscillator is constantly taken in and out of oscillation, is the matter of the upward crcep or vol- tage cc crystal and triodc reach tucrnal equilibrium. Two factors which have a profound influence on the quel- ity of tne end point are tue order of titration and tne pron eencc of a neutral electrolyte. there both ccic uni base are Itrong the order of titration has no effect, where one or the otner is weak the neck member must be titrated with the strong to avoid adverse effects. This observation is analogous to the same effect noted in conventional conductomctric titra- tions. Tne desensitization due to a neutral electrolyte is also analogous to the effect notec in conventional conducto- metrio titrations. Ho significant change in response was noted when the or- der of aolvent-ccmpie introduction on: reversed. Peak vol~ tags timed reaiingn indicated that for toe systems studied hero the interval botwoen successive inoreoonta or titrant, is not a significant factor. Back titration in too usual sense Inero a weak base was over titrated With a strong standard acid than bacx titrated to an and point with a strong standard base, was found to Have no advantagas over his oonvontional filtration to the direct 0nd point. In I for cases back titration gave a lowor sensi- tivity than the conventional titration. The importance of to: aolvont nan be extrapolated from too results obtained for aqueous, dioxane, and acetic acid media. In general the prOpartiaa of a solvent union allows it to inoraaoe the dissociation of a weak acid or base and tharaby inoroaee the conductance, are aivautageoua to high frequuncy and points. High frequency methods of analysis in non-aqueoua media possess definite concentration limitations. lbs more versa- tile instrumonts, such as the ono discussed in this study, having a more extended concentration ramps tends to have other off settin$ detects. Extension of hue concentration ran$e un- ually results in a lowered sensitivity. Increasing frequency usually extends the concentration range witnoat a deorvase in sensitivity, however in moat cases tne instrument booomos more complex with uttendent undesirable complications. The effect of concentration in oeak voltage measurement is predictable from tho general characteristics of high fro- quenoy inatrumonta. Since caoaoitanoo is adjusted to renain constant during a series of moasuvonunta the response notad in olmoot oooplotoly due to change in coniuctanoo. Conduc- tanco of a solution dapondc upon the number and mobility of the ions present, tnat is, the concentration ani kind of ions present. Sonaideving a response for a particular system the concentration defines the magnitude of too r.aponse. The particular systdm defines the functional raiationohip botwoon concentration and rosponao, which is in all probability uni- quo for each system and in a function of tn. dissociation of the particular aobstanco in toot oystom. Thus too optimum concentration for greatest sensitivity cannot be calculated from the present development of the mothod it must be fiotor- mined oxporiocntally. A ganorol observation in this study was tnat too lower the degree of ionization the higher too concentration requirud for maximum aensitivity. A comparison of the potontiometrio method of analysis to the high frequency peak voltago mothod of analysis will serve to show the relative advantages and disaivantages of the method. In the course of $113 study it was found that to. potentiometric mothoi had a moro severe concentration limitation tuna did too hign fro- quency method, toot is. in the very dilute solutions where a potentiometrio and point was unsatisfaatory too high frequency method gave a batter and point. The limitation of base strengths use less severe for potentiomutric titrations than for nigh frequency, that 13, bases which were too weak to show a high frequonoy end point, have reportedly boon successfully titra- ted potentiomotrioally. APPEHUIX Instrument Parts List (Figure 6) Capacitors Cl, 0-100 mmf midget air padder 02, 0-35 mm! midget air padder 03, to mmf mica Cu, 25 mmf mice 05’ 10 mmf mica c7, 03, cg, Clo, 0.001 mfd uso volt paper 011, 012, to mfd electrolytic can Resistors R1, 150K 1 watt Ra, th 1 watt R3, 10K wire wound pot. Ru, 15K 1/2 watt R5, 2K 1/2 watt 36: 3900 ohms 1/2 watt R7, 1200 ohms 1/2 watt R3, 1/2 megohme 1/2 watt R9, 600 ohms 1/2 watt M GA 3. 80 Diode rectifier 635 "Magic Eye" figscellgneous parts cont'd power transformer S.P.S.T. switch 6 terminal rotary switch crystal Tuning coil pilot light 6 volt dry plate selenium rectifier (1) (2) (3) (A) (5) .(6) (7) (8) <9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) LITLRATJRE CITE Alexander, F.C., Jr., Electronics, gg, No. a, 116(19u5), Anderson, K. Bettie, E.S., and Revinson, D., Anal. Chem., gg. 7h3(1950$. Anderson, K., and Revinson, D., lg;g.. p. 1272. Arditti, 3., and Heitzman, P., Compt, rend., gag, hh(1949). Bender, p., J. Chem. Education, g1, 179(1946). Berry, A.L., U.S. Patent 2076Q41, April 6(1937). Bever, R.J., Croutnamel, 5.0., and Dienl, H., Iowa State Coll. J. 501., g1, 239(1949). Bien, G.S., Anal. Chem.,‘g§, 909(195h). Blaedel, W.J., Burknalter, T.S., Flom, D.G., Hare, G., and Jensen, F.W., Ibid., gg, 198(1952). Blaedel, W.J., and Knight, H.T.,‘£2;§.,‘§§. 7Q3(195fl). Bleedel, W.J., and Malmstadt, H.V., £§i§., gg, 73h(1950). .£QLQ., p. 1&10. Lg;g., p. 1&1}. ‘ggig., g1, u7l(1951). 318.0361, ‘WOJO, alfil'flfitfldt, I-IOVO’ Petitjean, DOLO. and Anderson. WOKO’ Ibldc, 21, 1214.0(1952). Blake, G.G., Analyst, 1g, 32(1950). gggg., p. 639. 6.3. Australian J. Sc1.,‘;2, d0(l9u7). ll. 59(19h8). .égi 32(19h9). 6.6., Chemistry and Industry, 23(1946). 741(1949). Blake, ELL-g." EElEND Blake, Ibid., (23) (2h) (25) (26) (27) (28) (29) (30) (31) (32) (33) (3h) (35) (36) (37) (38) (39) (AG) (A1) (A2) (A3) (nu) (AS) Blake, G.G., J. Sci. Instruments, gg, 17A(19h5). 23;g,. g5. 77(19u7). l§i3., p. 101. Broadhurst, J.W., l§£§.,‘gl, 103(l9nh). Chien, J.Y., J. Chem. Education, gg, h9n(1947). Conant, J.B., and Werner, T.d., J. Am. Cnem. Soc., 2g, hABbil930). Bowling, J.J., Sci. Proc. Roy. Dubline Soc.,‘lé, 175(1921). Duke, F.R., Bever, R.J., and Diehl, H., Iowa State Coll. J. 8010, .2}, 297(19):}.9). Fischer, R.B., Anal. Chem., l2, 835(19h7). Flom, D.G., and Elving, P.J., Ibid., 2 , 5a1(1953). Forman, J., and Crisp, D., Trans. Faraday Soc., 42A, 186(1946). Fritz, J.S., "Acid-Base Titration: in Nonaqueous Solvents," lst cd., Columbus, Ohio, ihe G. Frederick Smitn Chemical Co., 1952. . Fritz, J.S., Anal. Chem.,.gg, 1023(1950). Fujiwara, 8., and Hayashi, S., gy;g.,‘g§. 239(195u). Gent, w.L.G., Trans. Faraday $00., 3;, 758(19u9). Hall, J.L., Anal. Chem.,‘gg, 1236(1952). ‘£QLQ., p. 124A. Hall, J.L., and Gibson, J.A. Jr., £2;3., Q}, 966(1951). Hall, N.F., J. Am. Chem. 800., 5g, Sll5(l930). Hall, N.F., and Conant, J.B., £3;g., 53, 3ou7(1927). Hall, N.F., and fiernor, T.H., 3233,, 29, 2367(1928). Jensené F.W., and Parrack, A.u., Ind. Eng. Chem. Anal. ‘ 1 Ed.. 1 595(19u6). Jensen, F.W., and Parrack, A.L., Bull. Agr. Mecn. Coll. Texas, h92(l9h6). (1.6) (A7) (A3) (49) (50) (51) (52) (53) (5h) (55) (56) (S7) (53) (59) (60) (61) (62) (63) (61+) (65) (66) Jensen, F.W., Watson, G.M. and Beckham, J.B., Anal. Chem., alt 1770(1951) 0 Jensen, F.W., Watson, G.fi. and Vela, L.G., Ibid., p. 1327. Johnson, A.u., oral communication. Johnson, A.H. Unpublished Masters Ihesis, micnigan State College (leags. markunas, P.C., and Riddick, J.A., Anal. Chem.,‘g}, 337(1951). nilner, 0.1., Ibid., gg, 12u7(1952). fionafihan, P.H.. Moseley, P030: Burknalter, T.S., and Nance, Ovo' Ibid., p. 1)}. Nakano, K., Hara, 3., and Yasniro, K., Ibid.,‘§é, 636(1954). Nance, O.A., Burknalter, T.S., and fionaghan, P.H., Ibid., Pifer, c.w., and Wollish, 3.6., Ibid., g5, 310(1953). Pifer, C.W., Wollish, E.G., and Schmall, M., Ibia., p. 310. Reilley, C.N., and McCurdy, W.H., Jr., Ibid., p. 86. Richards, W.T., and Lonnie, A.L., Proc. Natl. Acad. Sci., U080, .42, 5'37(1929)0 HidjiCk, JOAO' Anal. Cllefllo, 3&3 all-1952). Ridiick, Jr., and J.A., Fritz, J.S., Davis, a.a., fiillenbrand, E.F., alarmafl, P000, Ibido' p. 310. Sargent, E.H., and 00., Scientific Apparatus and aetaods, Le 39(1951). Sargent, E.H., and Co., technical bulletin describing Sar- gent iodel V oscilloneter, (1)52). . Seaman, W., and Allen, E., Anal. Chem., g2, 592(1951). Thom-as, Bel-4o, Faegln, FoJc, and Wilson, Claws, Ibido, p0 1750. Wagner, W.F., And Kauffmen, W.B., gg;3., 35, 538(1953). West P.W., Burkhalter, T.S., and Broussard, L., Ibid., _2___2_. 1169(1950) . - (67) West, PM}, fiobicnaux, T., and Bark'nalter, rJ.’..'il., Ibid., 2.2. 1625(1951). """‘" (63) West, P.W., Senise, P., and Burknalter, T.S., Ibid., g2}, 1250(1952) . HIGH rHiZ....,UuI.,FCI TI’i'xLA‘IIu;{S III ACLTIG ACID .uiDlA By Conrad M. Jankowskl AN A 38 Tit-'1. C T Submitted to tne Scnool of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for tne degree of .‘xiAS T2348 OF S'JI;:.=.;'3E Department of Cnemistry 195% Approved m... NH ”2 CLAl b1 The e +U n‘ be re Conrar fl. Jankowski High frequency titration is an electronetric technique of sad point detection. It possesses the advantage over existing techniques in that it eliminates the necessity of internal electruies. The purpose of this investigation was the study of the advantages and disadvantages inherent in the method and a com- parison of this method to conventional methods of end point detection. The instrument used was a conventional tuned plate, tuned grid, 3600 KC crystal oscillator utilizing a 6L5 triode. fleas- urenents were made by a capacitive loading of the plate cir- cuit. The effect measured was the maximum biased grid voltage of the loaded circuit Kile the instrument was in oscillation. The end point detected by the instrument as used is exelained in terms of conventional conductoaetrio titrations which may be related to this technique by ordinary methods of electronic circuit reduction. Of the factors evaluated most important are the effects D 0L solvent system, dissociation of the substance titrated, and the concentration. The effect of the solvent system is con- ventional and in accordance with he irflnsted theory of acids and bases. It was found that weak bases such as dimethylani- line, pyridine, hex methylenediamine, 3-hydroxyquinoline, ani- line, p-toluidine, and valanine, when titrated with a strong acid such as perchloric acid gave better end points in acetic acid media than in water or dioxane. ”he degree of dissocia- 4 Conrad fl, denkowski “tion of the substance titrated in a.particulsr solvent system load a proiound e feet. The in ore highly ionized substances gjave better ear points than the less di deleted or "wesL" substances. The effect of concentration is perhaps the most luaconventional aspect of the high frequency technique in that sensitivity of response is not a linear function of concen- tration. The we fit eds utilized gave two regions of maximum sensitivity separated by a region of low sensitivity. It 'was found that each system titrated had a unique rehion of Mgnest sinsitivity The location of this maximum r;gion of sensitivity dope ends upon the kind of i011s present. The raga ions of maximum sensitivity varied from 0.0§fl-0.00ZH for ani- line to O. L “-0 01m for dieth31a11iline. In acetic acid media the prec 131011 of hi 3h fresue:1cy end points compares favorably with potentiomotric end points where the concentration of maximum sensitivity is utilized. Ir most cases of weak bases the higj1 frequency region of maxinun sensi~ tivity extends to lower concentrations than does tLe potentio- metric concoztrstion range. The potentiometric method in ace- tic acid s"stems exte11us to weaker tags cs than does the high frequ -e:.cy method studied e. ihus high frequency techni- ques and potentiometric ”QCIqus are c moi scentsry to th study of acetic acid systens. MICHIGAN STATE UNIVERSITY L It lllll lullluumflfiji 3 1293 3062 03 A RIES