Analyst, August, 1979, Vol. 104, pp. 739-749 739 High-frequency Microtitrimetric Determination of Acidic and Basic Constituents in Lubricating Oils Part ll.* Determination of Total Base Number T. Fernhdez, J. M. Rocha, N. Rufino and A. Garcia Luis Compafiia Espafiola de Petrdleos, S.A . , Research Laboratory, Santa Cruz de Tenerife, Canary Islands and F. Garcia Montelongo Department of Analytical Chemistry, University of La Laguna, Tenerife, Canary Islands A high-frequency microtitration method is described for the determination of total base number in new and used lubricating oils. The sample, dissolved in toluene - propan-2-01 - water (5 + 4.95 + 0.05), is titrated with standard alcoholic hydrochloric acid solution. A study of solvents, titrants, sample size and other experimental parameters is reported.This method provides sharp breaks a t the end-points of the titration graphs and more repeatable and faster results than those obtained with the standard potentiometric method. The method can also be used for macrotitrations, with a com- mercially available oscillator, and can be suggested as an alternative to ASTM and IP methods for the determination of total base number. Keywords Total base number determination ; lubricating oils ; high-frequency titration In a previous paper the determination of the total acid number of lubricating oils by use of a high-frequency titrimetric method was described.1 In this paper we report the application of the high-frequency titration method to the determination of the total base number of these oils.The basicity of an oil is caused by nitrogen compounds, salts of weak acids (soaps), heavy metal salts and additives, such as inhibitors and dispersants. Basicity is an important property of lubricating oils as bases neutralise acidity, and therefore offset corrosion, which can be caused by oxidation of the oil during use. The total base number (TBN) of an oil is defined as “the amount of acid, expressed in milligrams of potassium hydroxide, that is needed to neutralise all basic constituents present in one gram of ~arnple.”~,~ The ASTM and IP method^^,^ of measuring TBN describe the potentiometric titration of a solution of the oil in toluene - propan-2-01 with a standardised solution of hydrochloric acid in propan-2-01. If a manual instrument is used these methods are rather slow, because of the large number of points that must be recorded and the slow indication of the changing pH.At present, this situation is improved because most laboratories working with these methods use automatic titrators that do not require too much operator time. However, the graph obtained does not always show a satisfactory end-point, particularly with used oils. The use of a standard buffer solution, 2,4,6-trimethylpyridine, and hydrochloric acid in propan-2-01, improves the results somewhat, but it is not certain that these results are meaningful and actually measure TBN as previously defined. Recently, potentiometric methods using glacial acetic acid as solvent have been introd~ced.~.~ The acceptance of the results obtained is wider than for those obtained by other methods.However, they are not completely satisfactory for used oils. In service, the replacement of a lubricating oil is determined by several factors, amongst which is the TBN. For heavily used oils the time for replacement may be difficult to deter- mine because of errors in the measurement of TBN, if this is the only factor considered. High-frequency techniques have been applied successfully to the titration of bases in non-aqueous media6-9 and can help to solve the difficulties mentioned previously. In this connection a method for TBN determination has been described,’ using a high-frequency * For details of Part I of this series, see reference list, p. 749.740 FERNANDEZ et al. : HIGH-FREQUENCY MICROTITRIMETRIC Analyst, VoZ.104 titrator adapted for use with a specially designed cell, but its specific design has created difficulties in its practical application. This paper shows that the high-frequency titrators that are commercially available can be applied to macro- and micro-determinations of TBN. Procedures have been developed by using solvent and titrants similar to those used in the ASTM methods.2-5 The proposed method reduces the time of analysis to about 10 min per sample, eliminates the contami- nation of electrodes, always shows sharp breaks at the end-point and provides greater precision. In addition, more meaningful results are obtained. Experimental Apparatus High-frequency oscillator for microtitrations, digital ammeter, pH meter, electrodes and microburet te have been described previous1y.l An Oscimhometer OK-105 (Metallurgical Research Institute, Budapest),lo equipped with a 500-ml (350 ml effective) inductive cell, was used for macro-titrations. This instrument was obtained from Metrimpex, Budapest 62 (P.O.Box 202) and the cost in 1977 of the manual instrument was about fT450. According to reference 9 the Oscimhometer can be classified as being Group 1 response type, and although it is equipped with its own response meter, a Hewlett-Packard digital multimeter, Model 3465A, (1-100 pA & 0.2%) measuring instru- ment was used. Samples In order to prepare synthetic samples a paraffinic base oil [diO 0.866 g ml-l, kinematic viscosity 10 cSt (10-5 mez s-l) at 100 "C], zero TBN value is used. The additives and trade lubricating oils are listed in each table. Reagents In the following paragraphs volumes referring to macrotitrations are given in parentheses.Analy tical-reagent grade chemicals were used throughout without further purification. Titrants Hydrochloric acid, 10-1 M solution in propan-2-01. This solution is made up according to ASTM D-664.2 Standardise it frequently by use of the high-frequency technique, using 50 p1 (3 ml) of 10-1 M standard potassium hydroxide solution in propan-2-01 or 10-1 M Tris [2-amino-2-(hydroxymethyl)propane-l,3-diol] standard solution, and toluene - propan- 2-01- water as the solvent. The maximum deviation allowed must be less than &0.2% ( 5 2 X 10-'M). Perchloric acid, 10-1 M solution in glacial acetic acid. This is made up according to ASTM D-2896.4 Standardise the solution frequently by use of the high-frequency technique, using 100 pl (3 ml) of standard potassium hydrogen phthalate solution in glacial acetic acid and toluene - propan-2-01 - acetic acid as the solvent.Sodium acetate solution in glacial acetic acid, 10-l M. This is made up according to ASTM D-2896.4 Standardise the solution by using 1OOpl (3 ml) of standard perchloric acid solution and toluene - propan-2-01 - acetic acid as the solvent. Standard 10-1 M Tris aqueous solution. Standard 10-l M potassium hydrogen phthalate solution in glacial acetic acid. Dissolve the phthalate according to ASTM D-2896,4 but diluting only with glacial acetic acid. Titration solvents prepared according to ASTM D-664.2 Toluene - propan-2-01 - water (5 + 4.95 + 0.05).This is the main solvent, which is Toluene - propan-2-01 - glacial acetic acid (3.5 + 3.5 + 3). Toluene - ethanol - glacial acetic acid (4 + 3 -t 3). Procedure Method 1. Samples with TBN more than 5. The test sample, 5-50 mg (1-5 g), is weighedAugust, 1979 DETERMINATION OF CONSTITUENTS IN LUBRICATING OILS. PART 11 741 into the titration cell, dissolved with 6 ml (350 ml) of the toluene - propan-2-01 - water solvent mixture (5 + 4.95 + 0.05) and titrated with the 10-1 M alcoholic hydrochloric acid standard solution. Method 2. Samples with TBN less than 5. The test sample, 50mg ( 2 4 g), is weighed into the titration cell, 50 pl (1-2 ml) of standard 10-1 M sodium acetate solution are added, and the sample is dissolved with 6 ml (350 ml) of toluene - propan-2-01 - acetic acid solvent mixture (3.5 + 3.5 + 3).The solution is then titrated with the 1 0 - l ~ perchloric acid standard solution. Volumes and masses given in parentheses are for macrotitration with the Oscimhometer. The working technique with this instrument is similar to that already reported for micro- titrati0ns.l The titrant was added in 0.5-ml increments from a 10-ml burette, using about 3 ml of titrant to reach the end-point. Calculation The TBN value can be calculated from the equation TBN (A & B) x M x 56.1 m where A is the volume, in microlitres (millilitres), of standard acid titrant used to reach the end-point minus the volume of acid titrant corresponding to the sodium acetate solution added (Method 2); B is the volume of basic titrantl or acid titrant, molarity M , used to titrate 6ml (350ml) of solvent mixture and is to be added in Method 1 and subtracted in Method 2 (the solvent of Method 1 may become slightly acidic and that of Method 2 slightly basic, see under Sample size below); M is the molarity of the acid standard solution; and m the mass of the oil sample in milligrams (grams).Effect of the Experimental Parameters on the High-frequency Titration In order to avoid duplication and presenting a study similar to that previously reported on the determination of TAN,l we prefer to discuss only the experiments carried out by use of microtitration. However, the main studies of TEN determination were also carried out by macrotitration. As will be seen later, the shape of experimental graphs, the results and the conclusions agreed well with those obtained by use of the micro-method.Titrant standard solution In order to use, as far as possible, the reagents specified in ASTM methods D-664 and D-2896, we used only solutions of perchloric acid in glacial acetic acid and hydrochloric acid in propan-2-01 as titrants for direct titrations and sodium acetate in glacial acetic acid for back-titrations. Titration solvents Solvents used as media for the titration of weak bases in petroleum products must dissolve the sample completely, enhance the basicity of the species to be titrated and have a low solvating power for cations. Further, the relative permittivity of the solvents used in conductimetric titrations should be less than 20,* but not too small as they must ionise compounds to be titrated, at least partially.Solvents such as nitromethane, 4-methyl- pentan-2-one or trifluoroacetic acid, which are widely used in the potentiometric determina- tion of weak bases, cannot be used in conductimetric titrations because they do not satisfy some of the stated requirements. Acetic acid or toluene - alcohol - acetic acid mixtures give better results, but these solvents do not always improve the titration of weak bases. This can be related to the fact that the conductimetric end-points are generally located at the intersection of two straight, ascending lines. In order to enhance the localisation of the end-point, we tried to alter the slope of the graphs by using mainly toluene - alcohol mixtures as solvents, with or without the addition of acetic acid.The behaviour of the selected solvents was studied by reference to the response graphs of the oscillator with the acidic titrants and verified later by use of the titration graphs.742 FERN~NDEZ et at. : HIGH-FREQUENCY MICROTITRIMETRIC AaaZyst, VoZ. 104 Non-acidic solvent mixtures In our first trials a number of binary mixtures, 1 + 1 toluene-X (X being ethanol, propan-2-01, acetone, 1,4-dioxan, butan-%one or 4-methylpentan-2-one), were used as representative solvents and tested in the titration of a synthetic sample with hydrochloric acid solution in propan-2-01. Toluene - ethanol and toluene - propan-2-01 mixtures give the greatest slope change at the end-point. Further experience with such alcoholic solvents showed that a solvent with a relative permittivity of approximately 11 and a 40-60% alcoholic component should be the most satisfactory.The work was continued with toluene - ethanol (6.5 + 3 4 , (6 + 4) and (5.5 + 4.5) and the solvent described in the ASTM method D-664, toluene - propan-2-01- water (5 + 4.95 + 0.05), was chosen for comparative purposes. Response graphs of the oscillator (Fig. 1) show that toluene - ethanol (6.5 + 3.5) is the best solvent because of its high sensitivity, but that the ASTM D-664 solvent, which has less sensitivity, can be recommended as it originates a more linear response graph over a wide range of titrant concentrations. Both solvents can dissolve up to 200mg of sample under the experimental conditions. Volume of lo-' M acid reagent solution/pl Fig. 1.Instrument response graphs for 10-1 M hydrochloric acid and several solvent mixtures: A, toluene - ethanol (6.5 + 3.5); B, toluene - ethanol (6 + 4) ; C, toluene - ethanol (5 + 5) ; D, toluene - ethanol (5.5 + 4.5) ; E, toluene - propan-2-01- water (5 + 4.95 + 0.05), M perchloric acid as titrant; and F, toluene .. propan-2-01- water (5 + 4.95 + 0.05). Acidic solvent mixtwes Acetic acid was selected as the acidic component, because of its well known behaviour, as the solvent and species to be titrated are similar to those described in ASTM D-2896. From the mixtures tried, after reference to their response graphs, toluene - ethanol - acetic acid (4 + 3 + 3) and toluene - propan-2-01- acetic acid (3.5 + 3.5 + 3) were selected. Toluene - acetic acid (1 + l), chlorobenzene - acetic acid (6.7 + 3.3) and the solvent specified in ASTM method D-2896 were also studied for comparative purposes.These solvents were further tested in the titration of a synthetic sample by direct and back-titration (Table I). From the experimental graphs shown in Fig. 2, it can be deduced that toluene - propan-2- ol- acetic acid mixture is the best solvent for the direct titration and toluene-ethanol- acetic acid for the back-titration. Fig. 2 also shows that the solvent in the ASTM D-2896 method and toluene - acetic acid are not suitable for either titration method. Table I summarises the recommended solvent - titrant pairs.Aztgzl.St, 1979 DETERMINATION OF CONSTITUENTS I N LUBRICATING OILS. PART I1 TABLE I 743 SOLVENTS AND TITRANTS RECOMMENDED FOR THE DETERMINATION OF TBN VALUES BY HIGH-FREQUENCY TITRATION Solvent Titrant ( -10-1 M) Toluene - propan-2-01- water (5 + 4.95 + 0.05) Toluene - propan-2-01- acetic acid (3.5 + 3.5 + 3) Toluene - ethanol - acetic acid (4 + 3 + 3) * Total volume added less than 60 pl.t Perchloric acid previously added, back-titration. Hydrochloric acid - propan-2-01; perchloric acid - Perchloric acid - glacial acetic acid Sodium acetate - glacial acetic acidf glacial acetic acid* .llO1 (a' A A 100 - 90 - 80 - 0 20 40 60 80 100 120 140 Volume of lo-' M acid reagent so I u t ion /p I I , I 0 30 60 90 120 150 Volume of lo--' M Cti,COONa sol u t ion/;r 1 Fig. 2. Titration graphs for a synthetic sample of ECA-5830, TBN value 74, in several solvent mixtures.(a) Direct titration with 10-1 M perchloric acid; ( b ) back- titration with 10-1 M sodium acetate solution (150 pl of 10-1 M perchloric acid pre- viously added). A, Toluene - propan-2-01- acetic acid (3.5 + 3.5 + 3) ; B, toluene - propan-2-01- water (5 + 4.95 + 0.05), 10-1 M hydrochloric acid as titrant; C, chloro- benzene - acetic acid (6.7 + 3.3) ; D, toluene - acetic acid (5 + 5) ; and E, toluene - ethanol - acetic acid (4 + 3 + 3). Saqble size The response graphs showed that the total amount of titrant added should be limited to about 1OOpl. The sample size chosen was adequate for the expected TBN value and the deviation of the results was studied. Table I1 shows the sample sizes recommended for different expected TBN values in micro- and macrotitration, and Table I11 shows the deviation obtained in titrating solutions of pure additives.The so-called pure additives are not pure substances but mixtures of different chemical species (e.g., zinc dialkylphosphoro- TABLE I1 SAMPLE SIZE RECOMMENDED FOR VARIOUS EXPECTED TBN VALUES Microtitration A r > I Expected Approximate Sensitivity TBN value sample masslmg of weighinglmg 0-5 100 0.2 5-10 50 0.1 10-20 25 0.05 20-50* 20 0.02 50-100* 15 0.02 Macrotitration Approximate Sensitivity sample mass/g of weighing/g 5-3 0.01 3-2 0.005 2- 1 0.005 1-0.5 0.002 0.5-0.2 0.001 A > * Alternatively use a preparation of the lubricating oil in a base oil, which allows an amount in one of the first three ranges to be weighed out.744 FERNANDEZ et aZ. : HIGH-FREQUENCY MICROTITRIMETRIC Analyst, Vd.104 dithioate - basic barium dinonylnaphthalene sulphonate, polyamino monoalkenylsuccinimide - alkaline calcium petroleum sulphonate, barium thiophosphonate) . Further experiences with a synthetic sample made from ECA-5830 and the base oil, TBN value 7.4, and amounts of sample ranging from 20 to 100 mg, showed that maximum deviation from the mean is 0.2 TBN unit (Table IV). This result is the same as was found for pure Lubrizol 864 solution, which has the same TBN value but a very different composition. TABLE I11 TBN VALUES DETERMINED FOR SAMPLES OF ADDITIVES TBN value Lubrizol 864 Anglamol 99 Lubrizol 6610 Determined . . . . . . . . .. 8.95 19.2 61.4 8.91 19.3 62.3 8.69 18.7 62.6 Mean . . .. . . . . . . .. 8.85 19.1 62.1 Maximum deviation (from mean) ..0.2 0.4 0.7 Maximum deviation (among results) . . 0.3 0.6 1.2 TABLE IV INFLUENCE OF SAMPLE SIZE ON THE DETERMINATION OF THE TBN VALUE FOR A SYNTHETIC SAMPLE OF ECA-5830 DISSOLVED IN BASE OIL Sample mass/mg 24.13 40.09 57.3 67.5 76.7 92.7 111.6 10-1 M HCl added/pl 33 54 77 90 101 120 143 TBN value 7.67 7.56 7.64 7.48 7.39 7.26 7.19 Mean .. . . . . .. .. 7.44 Mean deviation (among results) . . .. 0.5 Mean deviation (from mean) . . . . 0.2 Standard deviation .. .. . . 0.2 Relative standard deviation . . . . 2.3% Relative error11 . . . . . . . . 2.1% Similar experiments were carried out with the same synthetic sample, toluene - propan- Fig. 3 shows the experimental graphs and Table V the results obtained. Fig. 4 relates 2-01 - acetic acid solvent and perchloric acid in acetic acid as the titrant.100 2 90- W 80- f! 70- - a m C .- L s 50b ;o 40 $0 E;o lA0 1;o 1 2 Volume of l O - ’ M perchloric acid sol u t i on/p I Fig. 3. Titration graphs for the deter- mination of synthetic samples of ECA-5830 dissolved in base oil, according to Table V (graphs A, B, C, D and E correspond to titrations 1, 2, 3, 4 and 5 in Table V).Augast, 1979 DETERMINATION OF CONSTITUENTS IN LUBRICATING OILS. PART 11 745 TABLE V INFLUENCE OF SAMPLE SIZE ON THE DETERMINATION OF THE TBN VALUE FOR A SYNTHETIC SAMPLE OF ECA-5830 DISSOLVED IN BASE OIL, IN TOLUENE -PROPAN- 2-OL - ACETIC ACID AS SOLVENT AND 10-l M PERCHLORIC ACID AS TITRANT Volume of 10-1 M perchloric acid Test Sample masslmg added/$ 1 20.36 31 2 32.49 49 3 48.03 69 4 71.3 101 5 89.2 125 TBN value r Uncorrected Corrected* 8.54 7.60 8.46 7.88 8.06 7.65 7.95 7.67 7.86 7.64 A \ Mean value .. . . .. .. .. .. 8.17 7.69 Maximum deviation (among results) . . . . 0.7 0.3 Standard deviation . . .. .. . . . . 0.3 0.1 Relative standard deviation . . .. . . 3.8% 1.4% Relative error1' .. .. .. . . . . 4.7% 1.8% Maximum deviation (from mean) . . .. , . 0.4 0.2 * 3 pl blank correction. titrant consumptions to sample mass, and extrapolation reveals the need for a blank correction for the solvent, which must be taken into account in calculations. In this way the standard deviation and relative error are improved and the maximum deviation among results becomes less than 0.3 TBN unit. This correction can be related to the basicity of the acetate ions in the solvent.Sample mass/mg Fig. 4. Relationship between titrant consumption and sample mass according to Table V (on the magnified scale the blank value is shown). Addition of sodium acetate or Tris As increasing sample size above 200 mg leads to a loss in sensitivity and disturbs readings, the addition of a pure, typical, weak base to the sample was studied, in order to allow titration of those with a very low TBN value. Sodium acetate and potassium hydrogen phthalate were tested in the acidic solvent with perchloric acid as titrant. The results shown in Table VI suggested that the use of sodium acetate is better than potassium hydrogen phthalate. However, according to the results in Table VI, in a medium of toluene - propan-2-01- water with hydrochloric acid in propan-2-01 as titrant, the addition of Tris can be convenient; this base must be used cautiously, keeping the amount added to a limit of 2.5 pmol.In all of the experiments the consumption of the 10-1 M acid titrant was kept below 150 pl so that the net consumption of titrant was about 30 p1.746 FERNANDEZ et al. : HIGH-FREQUENCY MICROTITRIMETRIC Analyst, VoZ. 104 TABLE V I INFLUENCE OF THE ADDITION OF SODIUM ACETATE, POTASSIUM HYDROGEN PHTHALATE OR TRIS ON THE DETERMINATION OF THE TBN VALUE FOR A SYNTHETIC SAMPLE OF ECA-5830 DISSOLVED IN BASE OIL TBN value Volume of 10-l M base added/pI* 0 25 25 50 50 75 100 Acetate? 7.72 7.79 7.78 7.24 7.81 7.87 7.97 Phthalatet 7.72 8.07 8.10 8.73 8.38 7.83 7.28 TrisS 7.67 7.69 7.62 8.02 8.12 8.70 8.72 * Sample size -20 mg.t In toluene - propan-2-01 - acetic acid as solvent with 10-l M $ In toluene - propan-2-01 - water as solvent with 10-1 M hydro- perchloric acid as titrant (3 pl blank solvent, Table V) . chloric acid as titrant. Results and Discussion In order to check the proposed methods, synthetic samples with compositions corres- ponding approximately to trade samples were prepared from a base oil and a set of additives widely used in the lubricating oil industry. The TBN values of the additives were known from the manufacturer or could be determined by the standard ASTM potentiometric method. Table VII summarises the TBN values obtained in the microtitration of several synthetic samples. TABLE VII TBN VALUES DETERMINED FOR SEVERAL SYNTHETIC SAMPLES OF ADDITIVES DISSOLVED I N BASE OIL Lubrizol TBN value Oloa , A \ 851 4266 4084 6610 3882 3882 56 56 219 4084 1360 1360 3715 Determined .... 14.5 5.48 6.85 7.02 9.35 11.1 15.8 64.7 14.6 5.46 6.81 6.96 9.38 11.4 15.7 63.8 14.5 5.66 6.83 7.16 9.44 11.6 16.2 64.1 Mean . . 14.5 5.53 6.83 7.05 9.39 11.4 15.9 64.2 Expected’;alue;. . . 13.9 5.4 6.8 7.1 9.4 11.4 16.0 64.3 Potentiometrict . . . . 5.25 6.91 10.7 15.1 (5.11-5.47) (6.63-7.20) (10.3-11.3) (14.8-15.4) Difference$ .. .. $0.6 -0.13 $0.03 -0.05 -0.01 0.0 -0.1 -0.1 * Calculated from % m/m and TBN of additive added as known from the manufacturer. -f Determined by ASTM D-664, mean of three determinations; range given in parentheses. 3 Difference: mean - expected value. TABLE VIII DETERMINATION OF TBN VALUES ON SAMPLES OF NEW CEPSA TRADE LUBRICATING OILS Sample trade name ‘ Rodaje Extra Multigrado Teseo S-3 Serie 3 Troncoil RCgulo Chrysler HD-10 20-40W SAE-30 SAE-30 1530 HD-50 Determined .. . . . . 5.94 6.27 7.10 9.72 11.3 15.1 16.8 5.72 6.21 6.96 9.68 11.4 15.2 15.9 5.78 6.20 7.13 9.83 11.2 15.1 15.9 Mean . . . . . . 5.81 6.23 7.06 9.74 11.3 15.1 16.9 Expected Val&* . . . . 5.4 6.8 7.1 9.4 11.4 15.0 16.0 Potentiometrict . . . . 6.10 9.14 14.9 (5.92-6.27) (8.99-9.34) (14.6-1 5.3) Difference$ . . . . . . +0.41 -0.57 -0.04 +0.34 - 0.1 + 0.1 - 0.1 *, t and 2 as in Table VII. Rdgulo Super 650 63.2 64.6 64.4 64.1 64.3 60.4 - 0.2 (59.8-61.6) Other titrations were performed on two used oils from a Chrysler C-24s plant motor afterA$tgust, 1979 DETERMINATION OF CONSTITUENTS IN LUBRICATING OILS.PART I1 747 Fig. 5. Typical graphs obtained in the deter- mination of TBN values of two used lubricating oils removed from a Chrysler C-24s plant motor. (a) Motor oil MIL-L-2104B after working time: A, 48 h (TBN 4.01); B, 96 h (TBN 2.79) ; C, 192 h (TBN 2.24); D, 312 h (TBN 1.26); E, 408 h (TBN 0.30); and F, 0 h (TBN 6.23). (b) Motor oil MIL-L-46152 after working time: A, 50h (TBN 5.07); B, 100 h (TBN 3.97); C, 225 h (TBN 1.80) ; D, 300 h (TBN 1.36) ; and E, 0 h (TBN 7.06). various periods of working (Figs. 5 and 6). on samples of new CEPSA trade lubricating oils [Table VIII and Fig. 7(a)]. X; the latter table summarises TBN values determined on used oils from marine engines. Replicate determinations were also carried out Finally, the results obtained by macrotitration are shown in Fig.7 ( b ) and Tables IX and L 100 200 300 400 500 600 Timelh Fig. 6. Variation of TBN value with working time for two lubricating oils from a Chrysler C-24s plant motor: (a), motor oil MIL-L-2104B; ( b ) , motor oil MIL-L-46152. The repeatability obtained (Tables VII and VIII) is very good, as the differences from the mean are less than 0.1 TBN unit; such values are obtained with the ASTM standard method only when the potentiometric graphs show good inflection points2 The precision was established (Table IV) for a synthetic sample of ECA-5830, TBN value 7.4, and shows a standard deviation of 0.2 and 5.6% relative error (95% probability). The estimated accuracy of the method can be deduced from the results recorded in Table VII. In this wide range of TBN values (5-64), the differences between the obtained and expected values are less than or equal to 0.1, except for one sample.We must emphasise the need748 FERNANDEZ et al. : HIGH-FREQUENCY MICROTITRIMETRIC Analyst , VoZ. 104 - ( a ) 20 10 - - 00 90 80 70 60 50 0 20 60 100 140 180 Volume of lo-' M hydrochloric acid solution/pI / a / Fig. 7. Typical graphs obtained in the determination of TBN for synthetic, new and used lubricating oils by (a) micro- and (b) macro- titration: A, synthetic Lubrizol 56 (TBN 64.2-65.7); B, new Teseo S-3, SAE-30 (TBN 9.74-9.553 ; C, synthetic Lubrizol 56 + Lubrizol 1360 (TBN 15.9) ; D, used marine oil (TBN 13.0-14.2, vessel D, Table IX) ; and E, used marine oil (TBN 6.58-5.51, vessel A, Table IX). for a good determination of the blank correction for the solvent, particularly with solvent mixtures containing acetic acid (Table V).Samples with low TBN values can be resolved by cautious additions of Tris or, better, by adding sodium acetate. Nevertheless, this method is not totally satisfactory from the standpoint of the definition of TBN value according to ASTM method D-664, as basic constituents with a pKb of about 10 cannot be titrated. I t must be taken into account that the high-frequency conductimetric end-points are generally located at the intersection of two ascending lines. The slope of the titration line before the end-point becomes closer to the slope of the excess acid line as the basic strength of the sample decreases and the end-point becomes increasingly difficult to locate accurately.In this sense, the differences observed in titrating samples of different sizes can be related to their content in very weak species. TABLE IX COMPARISON OF TBN VALUES DETERMINED BY MACRO- AND MICROTITRATION ON SYNTHETIC AND NEW LUBRICATING OILS Sample TBN value Microtitration Macrotitration r A I Lubrizol4266* . . .. .. .. 5.66 5.60 Lubrizol 56* .. .. 64.2 65.7 Lubrizol 56 + LubAzol 1360* . . .. 15.9 15.9 Lubrizol 3882 + Lubrizol 1360* .. 9.39 9.40 ECA-5830* .. .. .. .. 7.96 7.99 Rodaje Chryslert . . .. .. .. 5.81 5.96 Teseo S-3 SAE-307 . . .. .. 9.74 9.65 * Synthetic samples: additives dissolved in base oil. t CEPSA samples, trade name. In spite of the difficulties described, high-frequency titration appears to be a very encouraging technique, as by using mixed solvents it is possible to differentiate weak bases in mixtures.6 A limitation in sample size can be avoided by use of the OscimhometerfoAzqpst, 1979 DETERMINATION OF CONSTITUENTS IN LUBRICATING OILS.PART 11 TABLE X DETERMINATION OF TBN VALUES FOR SAMPLES OF USED OILS FROM MARINE ENGINES Vessel A Vessel B Vessel C Vessel D TBN value r * - 7 - - - - t t t t Determined . . 5.61 5.50 6.99 6.51 6.55 6.77 13.0 14.4 5.48 5.63 6.88 7.01 6.66 7.82 12.8 14.0 5.64 5.42 6.94 6.89 6.98 6.47 13.1 14.1 Mean . . .. 5.58 6.54 6.94 6.80 6.73 7.02 13.0 14.2 New motor oil .. 8.00 9.74 15.13 25.64 * Microtitration. t Macrotitration. 749 according to the procedures described. In connection with this, Tables IX and X also report results obtained in tests performed by use of macrotitration using this oscillator, and one can see that the mean TBN values obtained with both techniques are virtually equal, except for one sample.Likewise, the shapes of the titration graphs are very similar [Figs. 7(a) and (b)]. With used oils, Table X suggests that the repeatability obtained by use of microtitration is better than that obtained with macrotitration. However, this suggestion cannot be substantiated firmly because there are too few results. Workers in CEPSA labora- tories are now establishing a co-operative testing programme using macrotitration in order to solve this problem. Compared with the standard potentiometric method, the use of high-frequency analysis avoids the possibility of electrode contamination and damage, as the electrodes are outside the titration cell.A sharp break is always obtained at the end-point of the titration graph, and no buffer is needed. Once the apparatus has been set up, a single TBN determination can be performed in about 10 min, instead of about 1 h, which is required for the potentio- metric determination when a manual instrument is used. The outstanding advantages of the high-frequency method of TBN determination are its good accuracy, the validity of its end-points and the smaller spread of results. As a com- parison, four synthetic and four trade oil samples were analysed by the standard potentio- metric method; results shown in Tables VII and VIII prove that TBN values obtained by use of the potentiometric method are not as consistent as those obtained with the high- frequency method. In back-titration, the high-frequency method (Table I) is an improve- ment on the equivalent potentiometric method, ASTM D-2896. The method developed for TBN determination can be suggested as an alternative to ASTM and IP methods. A similar conclusion was reached in our paper on TAN determination by high-frequency titrati0n.l This similarity can be used as an argument for the acceptance of €ugh-frequency titration in both determinations. References 1. 2. 3. 4. 6. 6. 7. 8. 9. Fernkndez, T., Rocha, J. M., Rufino, N., Garcia Luis, A., and Garcia Montelongo, F., Analyst, 1978, “1977 Book of ASTM Standards,” Part 23, American Society for Testing and Materials, Philadelphia, “Institute of Petroleum Standards 1978,” Part I, Volume 1, Institute of Petroleum, London, 1978, “1977 Book of ASTM Standards,” Part 24, American Society for Testing and Materials, Philadelphia, “Institute of Petroleum Standards 1978,” Part I, Volume 2, Institute of Petroleum, London, 1978, McCurdy, W. H., Jr., and Galt, J., Analyt. Chem., 1958, 30, 940. Caughley, B. P., and Joblin, M. V., Analyt. Chem., 1969, 41, 1211. Pungor, E., “Oscillometry and Conductometry,” Pergamon Press, London, 1965. Blaedel, W. J., and Petitjean, D. L., “High-Frequency Method of Chemical Analysis,” in Bed, W. G., Editor, “Physical Methods in Chemical Analysis,” Volume 111, Academic Press, New York, ” ‘Oscimhometer’ Type OK-105 Instrument Manual,” Research Institute for Non Ferrous Metals, Budapest, 1972. Lacroix, Y., “Analyse Chimique (Interpretation des RCsultats par le Calcul Statistique),” Masson, Paris, 1962. 103, 1249. 1977, Methods D664-58. Method 177/64. 1977, Method D2896-73. Method 276175. 1956, pp. 107-134. 10. 11. NOTE-Reference 1 is to Part I of this series. Received December 4th, 1978 Accepted March lst, 1979