Determination of ionization constants by chromatography - Analytical

Palalikit, and John H. Block. Anal. Chem. , 1980, 52 (4), ... Francesca D'Anna , Salvatore Marullo , Paola Vitale and Renato Noto. The Journal of Orga...
7 downloads 0 Views 913KB Size
Anal. Chem. 1980, 52, 624-630

624

temperature to 900 "C in furnace 4 (Figure 3) a t step 2 of the separation procedure. However, neither the yield nor the thickness of the deposited layer of 241Amwas satisfactory, so that modifications of the procedure will be necessary. Interferences of neodymium found by Knab (14) subsequent to ion-exchange separations of americium from soil might be overcome, since a separation of Am and Nd in the gas phase should be possible (15).

is then increased of course. T h e separation technique has until now been tested only on sandy soils, so it is an open question whether this technique could be used for all kinds of environmental samples, or whether the procedure would have to be modified. But the results show in any case that it is possible to separate picogram amounts of plutonium from gram amounts of soil by gas chromatography, as long as it is possible to obtain CY plates by simple vapor deposition of nearly the same quality as those made by electrodeposition. The time needed for the analysis is almost the same for the gas chromatographic analysis as for the classical analysis. The detection limit applied to a 2-g sample is somewhat better for the classical analysis, because of the better yield of 80% compared with 60% and the higher counting rate of 32% compared with 1870,but since the detection limit is still low enough for environmental control analysis, this is not a serious drawback for the gas chromatographic analysis. An advantage is the simple control for memory effects. For this control, the separation tube has simply to be heated up to 800 "C for 1 h, while a stream of C1, is passed through, until no plutonium is condensed on the glass disk. Only the HF decomposition device remains as a source of contamination, and this may easily be controlled separately, while for the classical analysis the whole separation procedure has to be carried out in order to control for memory effects. T h e gas chromatographic analysis may be run automatically with only slight modifications. The end of the separation tube has to be split into 2 outlets, one for A1C13. FeCI,, and Pl)Cl> and the second in connection with a capillary for the condensation. T h e gas chromatographic separation method is probably also suitable for the determination of americium in soil. Preliminary experiments where the soil sample wa5 spiked with 100 pCi 241Am/ghave shown. that A m CY plates could be obtained by vapor deposition only by increasing the

ACKNOWLEDGMENT The authors thank M. Pimpl, who participated in the early steps of the volatilization experiments.

LITERATURE CITED Schuttelkopf, H. "Rapid Methods for Measuring Radioactivity in the Environment", Proc. IAEA, Vienna, 1971, 183-200. Chu, N. J. Anal. Chem. 1971, 43, 449-452. Zvarova, T. S.,Zvara, I. J . Chromatogr. 1970, 4 9 , 290-292. Travnikov, S. S.; Davydov, A. V.; Myasoedov, 0 . F. Radiochern. Radioanal. Lett. 1976, 24, 281-289. Merinis, I.; Legaux, Y.; Bouissieres, G. Radiochem. Radioanal. Lett 1970, 3 , 255-261. Jouniaux. 6.Thesis, University Pierre and Marie Curie, Paris, 1979, Naumann. D. Kernenergie 1963, 6 , 116-121. Kubaschewski, 0.;Alcock, C. B. "Metallurgical Thermochemistry", 5th ed.; Pergamon Press: Oxford, 1979. Keller C. "The Chemistry of the Transuranium Elements"; Verbg Chemie: Weinheim. Germany, 1971; p 374. Binnewies, M.; Schafer, H. Z . Anorg. AUgern. Chem. 1974, 4 7 0 , 149- 155. Schafer, H. Angew. Chem. 1976, 88, 775-789. Gruen, D. M.; @a, H. A . Inorg. Nucl. Chem. Lett. 1967, 3 , 453-455. Sill. C. W.; Puphal, K. W.; Hindman. F. D. Anal. Chern. 1974, 46, 1725-1 737. Knab, D. Anal. Chem. 1979, 57, 1095-1097. Steidl, G . ; Bachmann, K. Technical University, Darmstadt, Germany, private communication, 1979.

R E C E I ~for ~ I review I September 28, 1979. Accepted December 10, 1979. The authors are grateful to the "Bundesministerium fur Forschung und Technologie'' for their financial support.

Determination of Ionization Constants by Chromatography Darawan Palalikit and John H. Block" School of Pharmacy, Oregon State University, Corvailis, Oregon 9733 7

termination of the ionization constant of a compound. Each method has its advantages and disadvantages. The spectrophotometric method requires chromophores which absorb ultraviolet or visible light. Also the relevant ionic and molecular species must have different spectra. The potentiometric titration method is very commonly used, but good water-solubility with an adequate quantity of the compound is needed. The conductometric method is quite tedious and more time consuming. The proton magnetic resonance method has proved useful for substances whose ultraviolet spectra do not change upon ionization, but the compound must be water-soluble. T h e limitations of the proton magnetic resonance method are the types of buffer that must be used, and the fact that a t least one proton must show a significant chemical shift when going from the unionized to the ionized species. All of these four methods require a compound which i. very pure. An alternative approach is to use chromatography Advantages of this method are that (1) small quantities of compounds are required; ( 2 ) poor water-solubility need not

The determination of the pK,'s of several organic acids and bases using chromatography was investigated. Normal-phase buffer-impregnated paper chromatography was found to be unsuitable. TLC plates impregnated with mineral oil were examined as a possible reversed-phase method, with poor results. The use of XAD-2 copolymer as the stationary phase In a simple HPLC method gave good results. Experimentally determined pK, values were confirmed using UV spectrophotometry with the identical solvents and buffers used in the HPLC method. Statistlcal comparisons using both pK, and K, values were performed. The advantage of using a computer method involving the second derivative to determine pK, is illustrated.

Several experimental methods sucli as ~VectrcJi)lic,tometry ( I ) , poteritiometry (2), conductornetry ( ; j ) , and p r o t n r i magnetic resonance spectrometry (2, 4 ) are available for the deC1003-2700,83/3357-0624$0 1 O O / O

L

1980 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 52, NO. 4, APRIL 1980

be a serious drawback; and (3) the samples need not be pure. There are many kinds of chromatographic methods which could be used: gas chromatography, paper chromatography, thin-layer chromatography, and high pressure liquid chromatography (HPLC). T h e gas chromatographic method is of limited usefulness because compounds are in t h e vapor phase rather than in aqueous solution. However, it has been valuable in finding t h e characteristics of nonideal solutions (5). All of the liquid chromatography methods involve the use of mobile phases of different p H values. For example, paper chromatographic methods were reported (6, 7). The resulting ionization constants, when compared with previously published results using traditional techniques, were found to be in good agreement. T h e determination of oil/water partition coefficients of substances has been studied by thin-layer chromatographic methods (8-10), but so far the determination of the ionization constants of substances has not been reported by this method. Since t h e partition coefficient relates to the ionization constant, i t should be possible to determine the ionization cons t a n t by thin-layer chromatography. Using the HPLC method, either ion-exchange chromatography or reversed phase chromatography would appear to be appropriate, The traditional ion-exchange resins have several drawbacks. They lack mechanical strength and will collapse under the pressures common to HPLC. They also swell or shrink as the counterion changes in the mobile phase (11, 12). Both of these handicaps cause changes in column volumes and inconsistent retention times. Many of the columns currently used in reversed phase HPLC suffer from two drawbacks. First, the columns cannot be operated above p H 8 because deterioration of the silica support becomes significant. Second, these columns have active silanol sites where no bonding of hydrocarbon to the silica has occurred. Some type of silation is usually required to cover these active sites ( 1 3 , 14). Recently a series of nonionic stationary phases have been reported. They are all organic and have been shown to withstand the pressures common to low flow rates ( 1 1 ) . Further, those polymers are devoid of ester linkages and would appear to be usable a t any p H which can be tolerated by the pumps and seals. Also, since they are nonionic polymers, they are not subject to shrinkage and swelling due to changes in the counterions in the mobile phase. Finally, being totally organic, and nonionic, they should lack the active adsorption sites characteristic of the silica-based internal phases. It was decided to investigate the use of one of these nonionic copolymer resins, XAD-2, in the determination of pK,'s by HPLC, and to compare the results with literature values and with those obtained by UV spectrophotometry using two methods of calculation.

ascending flow of amyl alcohol saturated with buffer. The ionization constants could be determined by plotting R, values as a function of the pH of the mobile phase (6). T h i n - L a y e r Chromatography. A nonaqueous stationary phase was obtained by impregnating the support (Kieselguhr G ) either directly or indirectly using a 2% v / v solution of Mineral Oil USP in an acetoneedioxane mixture (10). Methanolic solutions (0.3% w/v) of benzoic acid, o-chlorobenzoic acid, phenol, and ethanolamine were prepared, and l-pL volumes of solutions were spotted onto the plates. The plates were placed in chromatographic chambers that had been equilibrated for 16 h with the mobile phase. The mobile phase, saturated with mineral oil, consisted of inorganic phosphate buffer solutions of pH range 2-9. After development, the plates were air dried prior to being sprayed with a variety of detecting reagents (18). Hulshoff and Perrin derived a relationship between the R, value and the dissociation constant in thin-layer partition chromatography (10).

R, = log (,P) + log

K, Ka + [H+I

+ log r

Where R, is defined as log ( l / R f - l),,P is the partition coefficient expressed as the molar concentration in the stationary phase divided by the molar concentration in the mobile phase, and r is the phase-volume ratio expressed as the volume of the stationary phase divided by the volume of the mobile phase, and is a constant for given chromatographic system. H i g h Pressure Liquid Chromatography, A short column 9 cm long with 0.23-cm i.d. proved satisfactory. Phosphate buffers (0.01 M) were maintained at an ionic strength of 0.1 M by adding an appropriate amount of KCl (19). Sample solutions of 0.5% w/v were prepared in methanol. The column was slurry packed with XAD-2 using water as a slurry agent. Pressures were in the range of 600-800 psi, and flow-rate was 1.0 mL/min. Buffer-acetonitrile mixtures were expressed as percent by volume. The pH values of solvents containing the buffer were determined by a pH meter before and after adding acetonitrile. The pH meter was standardized each day at pH 4 and pH 7. The capacity factor was evaluated from the chromatogram (20) and is defined as: k' = ( t , - t o ) / t ,

where t , was the retention time for sample component under investigation. The tovalue was obtained from the retention time of NaN02. It has been shown by Equation 2 that k'is dependent on the dissociation of the compound and the pH of the mobile phase (19). For acids: k'=

k0

1 + K,/[H+]

-+

k-1

1 + [H+]/K,

(2)

where ko and k-l are defined as the capacity factors for the completely unionized (protonated, HA) and completely ionized (conjugate base, A-) forms of the acids, and K, is the ionization constant. Equation 2 can be written as:

EXPERIMENTAL Apparatus. A liquid chromatograph (Waters Associates Model 201), a variable wavelength UV detector (Varian Model 635LC) and a sample injection system (Waters Associates Model U6K) were used for the HPLC method. A spectrophotometer (Beckman Model DB-GT) and 1-cm matched cells were used for the spectrophotometric method. Reagents. All chemicals were used as received, Solvents were of an analytical grade and purchased in glass bottles. Amberlite XAD-2 was purchased as 20-50 mesh beads. Procedures for the cleaning and sizing of XAD-2 copolymers have been described previously (11,15-17). Particles of 4 5 4 5 pm mesh size were used. Procedure. Paper Chromatograph). Chromatographic paper (Whatman 3MM) was impregnated with inorganic phosphate buffer solution which was saturated with amyl alcohol. Methanolic solutions (0.3% W / V ) of benzoic acid, phenol and ethanolamine were prepared, and l-@Lvolumes of each solution were spotted onto the paper. The chromatography was carried out with an

625

(3) Solving for K,:

(4) Taking the negative logarithm: (5)

Equation 3 was rewritten in a linearized form (21) as: k ' = ko

-

K, ( k ' - k_,)/[H+]

(6)

By plotting k'vs. ( k ' - k_,)/[H+],the slope is -Kaand the intercept is k,.

626

ANALYTICAL CHEMISTRY, VOL. 52, NO. 4, APRIL 1980

Either Equation 5 or 6 can be used to determine pK, provided the ko and k-l values were known. It should be noted that in Equation 6, k’ occurs on both sides of the equation. For bases, a relationship between the capacity factor ( k 3 and pH of the mobile phase has been shown (19) to be:

where k, and ko were the capacity factors for the completely ionized (protonated, BH+) and completely unionized (conjugate base B)

Setting

Solving for K,: (12)

Taking the negative logarithm:

Equation 11 was rewritten in a linearized form (21) as:

k ’ = kl

+ K,(ko - k ? / [ H + ]

(14)

By plotting k’vs. ( k , - k?/[H+],the slope is K , and the intercept is k l . The pK, can be determined by either Equation 13 or 14 provided the kl and ko values were known. The same limitation applies to Equation 14 as with Equation 6. Spectrophotometric Method. Stock solutions of the same compounds as in the HPLC method were prepared ( 5 X Mj. To facilitate dissolution, an acid compound was dissolved in 0.005 N KOH and basic compounds were dissolved in 0.005 N HC1. Methanol was also added if the compound was not very soluble, but the alcohol content did not exceed 1% v / v in the stock M in solution. The stock solution was then diluted to buffer-acetonitrile solution to make the fmal solution at a specified pH. The absorbance of the final solution was then measured. Ionization constants of acids and bases were determined by Equations 15 and 16, respectively (2). Ai - A

pK, = pH

+ log A-A,

pK, = pH

+ log Ai - A

A - A,

At a specified wavelength, A is the absorbance at a certain pH, Ai is the absorbance when all of the sample is in the ionized state, and A , is the absorbance when all of the sample is in the neutral state. The wavelength used was between 220-300 nm. Determination of p K , using Computer Program. A program was written in FORTRAN which would calculate the ionization constant from the second derivative of the k’vs. pH curve for the HPLC method and the absorbance vs. pH curve for the spectrophotometric method. The program was based on a program written by Isenhour and Jurs (22). This enlarged program which is interactive is available upon request from J. H. Block and accomplishes the following: (1) Reads up to 100 data pairs. (2) Orders the data pairs that make up each point on the curve such that the r axis will start with the lowest pH and end with the highest.

(3) Decides if the curve is “uphill“ or “downhill“ and proceeds accordingly. (4) Plots k’vs. pH, the first derivative, and the second derivative. ( 5 ) Prints out data tables with appropriate labeling. (6) Prints out a limited number of error messages when the input data is incomplete or contain certain types of error. (7) Obtains the pK, by taking the cross-over value in the second derivative as the equivalence point (pK, = pH).

RESULTS AND DISCUSSION P a p e r Chromatography. With the experimental method that was described, unsatisfactory results were obtained. The reasons were: (1)the paper was destroyed by the detecting agent, and ( 2 ) it took too long for the chromatogram to develop. T h i n - L a y e r Chromatography. Hulshoff and Perrin (10) derived a relationship between t h e R, value and t h e dissociation constant in oil/water partition chromatography as shown in Equation 1. Problems were encountered with this procedure. The R , values for several of the compounds behaved independently of the p H of the mobile phase. In most cases the mobile phase had t6 be significantly nonaqueous to obtain a meaningful and reproducible R , value. While it might have been possible to extrapolate back to the R, values in pure water, the purpose of this study was to develop rapid procedures. Finally, the thin-layer chromatography procedure suffered from the same problem as with paper chromatography; i.e., finding appropriate detecting agents. Sometimes the mineral oil seemed to “protect” the compound from the detecting reagent. At other times the mineral oil would be affected by harsh visualization reagents with subsequent darkening of the entire plate. H i g h P r e s s u r e Liquid Chromatography. Horvath (21) showed that a plot of k’vs. p H produced a curve similar in appearance to that of a regular titration curve. He obtained dissociation constants by finding the midpoint of the curve where the concentrations of the acid form and conjugate base form were equal. His procedure suffered from two limitations. First, it was necessary to use a high concentration of acetonitrile in the mobile phase to obtain reasonable retention times for the unionized species. In addition, he needed to obtain k’values for the completely ionized form and unionized form of the compound (Equations 2-16). Thus solutes with pK, values lower than 3 (requiring buffers of p H 5 2) or greater than 9 (requiring buffers of pH 2 10) could be difficult t o study because of damage to the pumps, seals, and columns. Alternatively, the pK, could be calculated from a computer program (22) which obtains the first (Figure l b ) and second (Figure IC)derivatives from the plot of k’vs. pH (Figure l a ) . Using this method it was not necessary to know the k’values for the completely ionized form and unionized form. T h e program inputs a set of data representing a curve as a series of points (about six pairs). It numerically calculates the first and second derivatives, and then searches for the cross-over in the second derivative. An interpolation is made to find the exact cross-over value. Spectrophotometric Method. Because the relationship between the absorbance and p H is similar to the relationship between the capacity factor and p H (i.e., a titration curve), the pK,’s of organic acids and bases can also be calcuiated by using the second derivative. This is advantageous because the absorbance of the completely ionized and unionized species are not needed. Two different comparisons were made. In one, the results obtained by the HPLC procedure were compared with those obtained spectroscopically. In the other, pK,’s calculated from the second derivative were compared with those obtained from Equations 5 , 13, 15 and 16. These comparisons are summarized in Tables I and 11.

ANALYTICAL CHEMISTRY, VOL. 52, NO. 4, APRIL 1980

Table I. Comparison of the pK, Values of Organic Acids HPLC comequation compound name puter 5 2.76 2.86 2.90 3.22 3.12 3.49 3.94 4.09 4.54

o-chlorobenzoic acid (10% ACN) o-chlorobenzoic acid (20% ACN) salicylic acid (10% ACN)a p-nitrobenzoic acid (10% ACN) p-nitrobenzoic acid (20% ACN) acetylsalicylic acid (10% ACN) p-chlorobenzoic acid (20% ACK) benzoic acid (10% ACN) cinnamic acid (20% ACN)

spectrophotometer comequation puter 15

3.19 (10.12)b 3.32 ( t 0 . 1 6 ) 3.01 ( i O . 1 0 ) 3.50 ( i 0 . 0 7 ) 3.43 ( ~ 0 . 0 6 ) 3.64 ( * 0 . 0 8 ) 3.99 ( t 0 . 0 4 ) 4 . 1 3 (10.04) 4.45 ( t 0 . 1 6 )

2.47 2.98

__

3.61 3.23 3.59 3.96 4.02 4.85

lit. value

ref.

2.92 2.92 3.00 3.44 3.44 3.51 3.99 4.19 4.44

2, 23, 28 2, 23, 28 26, 27 2, 23, 29 2, 23, 29 29 2, 23, 29 2, 28 23, 28

3.03 ( i O . l O ) b 3.20 ( i 0 . 0 9 ) 3.01 ( t O . 1 0 ) 3.47 ( ~ 0 . 0 8 ) 3.50 ( 3 0.1 1) 3.60 ( i 0 . 0 7 ) 3.96 (i 0.06) 4.29 ( i 0 . 0 8 ) 4.39 (tO.ll)

627

(1 Some pK, values could not be determined because of excessive retention times or because of a lack of difference in the Sample standard deviation. spectra of the ionized and unionized forms.

-

Table 11. Comparison of the pK, Values of Organic Bases computer

compound name

13

4.62 4.50 4.91 6.91 7.21 7.24 9.19 4.13 8.87 5.14 6.03 8.08

aniline (10% ACN) aniline (20% ACN) pyridine (10% ACN) p-nitrophenol ( 1 0 % ACN) p-nitrophenol (20% ACN) phenobarbital sodium (20% ACN) ephedrine (10% ACN)= pyrilamine maleate (20% ACN)a quinoline (20% ACN) strychnine hydrochloride (20% ACN)a

spectrophotometer comequation 16 puter

HPLC equation 4.44 ( t 0 . 1 2 ) b 4.23 ( t 0 . 0 8 ) 5.1 2 (tO.09) 7.06 (10.10) 7.13 ( - 0 . 0 3 ) 7.55 (k0.13) 3.94 (10.11)

__

4.57 4.55 4.97 6.90 6.93 7.40 9.56 4.17 8.70

4.99 ( i 0 . 0 6 ) 6.01 ( + 0 . 0 8 ) 8.1; (10.03)

4.89 6.08 ._

lit.

4.56 ( ~ 0 . 0 3 ) ~ 4.52 (tO.09) 5.11 ( t 0 . 0 7 ) 7.17 (k0.14) 7.01 ( t 0 . 0 6 ) 7.60 ( i 0 . 1 6 ) 9.60 ( i 0 . 0 8 ) 4.01 ( t O . 0 7 ) 8.93 (10.12) 4.76 ( i 0 . 0 5 ) 5.94 (iO.10) .-

value

ref.

4.62 4.62 5.19 7.15 7.15 7.44 9.60 4.00

29 29 2, 26 2, 28 2, 28 29 27 27 27 28 26 28, 29

8.90

4.90 6.00 8.26

' Some pK, values couid not be determined because of excessive retention times or because of a lack of difference in the spectra of the ionized and unionized forms. Sample standard deviation. Table 111. Statistical Summary from Tables I and I1 for Acids and Bases HPLC computer equations 5 and 1 3

-

IF no. of compounds mean ( D and R ) significance level of mean S. D.

correlation (Lit. pK,'s vs. D or R ) significance level of correlation

21

0.092