Anal. Chem. 1991, 63,2642-2645
2042
Centrifugal Partition Chromatography of Palladium(I I) and the Influence of Chemical Kinetic Factors on Separation Efficiency Y.Surakitbanharn, S.Muralidharan,* a n d H.Freiser Strategic Metals Recovery Research Facility, Department of Chemistry, University of Arizona, Tucson, Arizona 85721
Separatlon efflclencles In tha centrlfugal padtion chromatoOraphic (=) -forthe dlsmwr#ng rpecles, a -P Ion complex and an organk solute (3-pkollne) for the heptamwater phase pair have been compared. The column effklency for the Pd( 11)-TOP0 (trloctylphosphlne oxlde) system (N = 500 f 40) has been found to be slgnlfkantly krs than forthe 3-pkolh sydem ( N = 1280 f 80) at a fkw rate of 0.50 f 0.05 mUmln and Is dependent on tts distrlbutlon ratlo. The behavbr of the Pd( I I)-TOP0 system in CPC can be attrlbuted to the slow klnetlcs of backaxtractlon of the PdCI,(TOPO), complex. The dlssoclatlon rate constant for thls complex In aqueous solutlon has been determined to be 168 f 8 M-' s-' by the stopped-flow klnetlcs procedure. This klnetlc study revealed that the parameter channel equivalent of a theoretlcal plate (CETP), whlch Is analogous to the reduced plate height, can be correlated to the half-life for the backaxtractlon and the dlstrlbutlon ratlo of Pd( 11). These flndhgs have general appkatbn to all chranatographlc separatbns lnvolvlng metal complex formatlon and dlssoclatlon because reactlon kinetics too fast to be observed by ordlnary means wlll have a dgnlflcant effect on chromatographic efflclency.
INTRODUCTION Centrifugal partition chromatography (CPC), a countercurrent liquid-liquid distribution technique, has been widely used for the separation of organic compounds, biological materials, and natural products (I,2). It has been only recently that this technique has been applied for the separations of metal ions (3, 4). We were the first to demonstrate the efficient separation by CPC of adjacent lanthanides including the separation of both light and heavy lanthanides in a single run using gradient pH elution (5). We were also the first to achieve the separation by CPC of palladium(I1) from platinum(II), rhodium(III), and iridium(II1) (6). These studies revealed that, under comparable conditions, CPC efficiencies for metal ion separations were significantly lower, by a factor of 4-5, than those regularly seen for organic compounds (7). Separations of metal ions by CPC involves formation and dissociation of extractable complexes using suitable ligands, most usually dissolved in the stationary organic phase. Further, the column efficiency is also a function of the distribution ratio of a given metal species, unlike the case of the organic compounds. Similar low efficiencies and the dependence of these efficiencies on the distribution ratios of the extracted metal species are also observed in the separation of metals by derivatized solid supports (8). Generally, in chromatography the column efficiency is constant for a given set of operating conditions, exhibiting no dependence on the distribution ratios of the species being separated. This puzzling observation in metal separations, which is yet to be addressed, prompted us to undertake a systematic investigation of the influence of chemical kinetic factors on the efficiencies of metal separations by liquid chromatography. CPC, being a liquid-liquid separation technique, has the 0003-2700/91/0363-2642$02.50/0
advantage over HPLC for elucidating the influence of chemical kinetic factors on chemical separations without the interference from the adsorption-desorption processes of the analyte encountered in HPLC. This study describes our efforts to determine whether, and in what manner, chemical factors, in contrast to simple solvation and desolvation as well as mass-transfer factors, were responsible for the differences in CPC efficiencies for metals and organic separations. CPC studies coupled with solution kinetic studies using stopped flow has allowed us to clearly establish the influence of chemical kinetics on CPC column efficiencies and correlate them to the half-lives of the chemical reaction responsible for the lowered chromatographic efficiencies. These are discussed in this paper.
EXPERIMENTAL SECTION Apparatus. A Sanki, Co., Japan, assembly consisting of a Model SPL centrifuge containing six analytical/semipreparative cartridges each having 400 channels (2400 total channels), a Model CPC FCU-V loop injector, and a Model LBP-V pump were used for CPC experiments. The current setup has been modified to obtain an internal volume of 120 mL compared to our original work on the separation of lanthanides where the internal volume was 130 mL (5). A UV-vis spectrophotometric detector (Model 770, Schoeffel Instrument Co.) with a 0.1-mL cell volume and 8 m m path length was used. It was set at 238 nm and 255 nm for Pd(I1) and 3-picoline experiments, respectively. Data were acquired every 10 s using an IBM/PC interfaced to the detector and a DASH-8 program (Metrabyte Co.). A Hi-Tech Scientific stopped flow SHU spectrophotometer was used for the kinetic study of the formation and dissociation of the Pd(I1) and TOPO complex. Data acquisition and treatment were accomplished with the associated Hi-Tech software. Pd(I1) and TOPO as well as chloride (NaC1) were dissolved in aqueous solutions of various concentrations of surfactant (0.7-4% Triton X-100). The kinetics was monitored at 420 nm. The viscosities of heptane solutions of TOPO were measured using an Ostwald viscometer. Reagents. Trioctylphosphine oxide, TOPO (kindly supplied by Dr. Richard Boyle of American Cyanamid Co.), was recrystallized from acetone (mp = 53-54 "C). All other reagents were of analyticalgrade. Metal-free heptane and water solutions were equilibrated overnight before use. Palladium stock solutions M) were prepared by dissolving a weighed quantity of palladium(II) chloride (59.9% Pd, Alfa Products) in 0.10 M HCl solution. 3-Picoline (Eastman Kodak Co.) was purified by distillation before use. The stock solution of 1W2 M 3-picolinewas prepared in water. Succinic acid buffer of pH = 6.10 (3.88 X M succinic acid and 6.94 X M NaOH) was prepared using Perrin's method (9). Deionized-distilled water was used throughout this study. Procedure. All CPC, kinetic, and viscosity measurements were performed at 25 OC. CPC experiments were conducted with 0.1-0.5 M TOPO in heptane as the stationary phase and an aqueous phase at pH 3 and appropriate chloride concentration (using NaCl) as the mobile phase, pumped in the descending mode. Equilibration of the two phases at 800 rpm resulted in 20 mL of heptane and 100 mL of water. A 1-mL aliquot of palladium stock solution M) was injected into the CPC system for each run. Flow rates between 0.5 and 4.0 ml/min were used in these experiments. The experimental configuration for the 3-picoline experiment was identical to that for palladium except that the detection 0 1991 American Chemical Soclety
4NALYTICAL CHEMISTRY, VOL. 63, NO. 22, NOVEMBER 15, 1991 2649
Table I. CETP and D Values for 3-Picoline and Pd(I1) at Different TOPO and Chloride Concentrations at a Flow Rate of 2.00 mL/min
[TOPOI, M
7, CP
0.0 0.2 0.3 0.4 0.5 0.5 0.5 0.5
0.386 0.485 0.556
0.643 0.738 0.738 0.738 0.738
3-picoline CETP D 3.2 3.6 3.9 4.2 4.4 4.4 4.4 4.4
fc1-1,
Pd(I1)
M
CETP
t i p , sb
D
0.1 0.1 0.1 0.1 0.05 0.20 0.30
10 17 25 33 68 13 11
2.2 3.3
0.4 0.9 1.4
0.9 4.4 5.5
11.0 2.7 1.1
2.8 8.3 0.6
0.4
Calculated values except in pure heptane. Calculated from eq 6. Flgurr 1. Comparison of the chromatograms at identical retention
volumes (132 mL) of (a) W(I1) ([TOPO] = 0.3 M, [Cr] = 0.1 M.)and (b) 3-picoHne (pH = 6.10). The origin of the dead volume peak (c) is
uncertain.
a 20
15
10
5
0
0
05
1
1
5
2
25
3
35
4
Fbw Rate M./d Figure 2. Van Deemter type plot for (a) the W(II)-TOPO system at 0.1 M [TOPO], lo-* M [Cl-] and pH = 3 and (b) lo3 M 3-picoline in succinic acid buffer of pH = 6.10.
wavelength was set at 255 nm. The mobile phase was adjusted to pH 6.10 using succinic acid buffer before injecting 1mL of 10-9 M 3-picoline.
RESULTS AND DISCUSSION Comparison of CPC Efficiencies of Pd(I1) and 3Picoline. We have completely characterized the extraction of Pd(I1) by TOPO by batch solvent extraction and CPC, in terms of the nature of the Pd(I1) species extracted and the dependence of the distribution ratio of this species on the concentrations of TOPO and Cl- (6). It was shown that Pd(II) was extracted as the PdC12(TOP0)2complex and that the distribution ratio increased with [TOPO] and decreased with [Cl-1. We define the term CETP, channel equivalent of a theoretical plate (2400/N), which is the equivalent of reduced plate height, to characterize the separation efficiency of CPC. The dependence of CETP on flow rates of Pd(I1) and 3-picoline were compared in the heptane:water system. The peak widths for palladium and 3-picoline at the same retention volume and flow rate of 2.0 mL/min are compared in Figure 1, which shows a significantly lower efficiency for Pd(I1) compared to 3-picoline. It is evident from Figure 2 that the CETP for Pd is higher than that for 3-picoline and that it exhibited a greater sensitivity to flow rate. The Pd(I1) and 3-picoline separations became less efficient (higher CETP) with increasing flow rate, analogous to conventional liquid chromatography. Chromatographic efficiency is typically limited by various types of diffusion processes which are operating for both
3-picoline and Pd(I1). Simple diffusion alone cannot account for the significantly lower efficiency for Pd(II), however, even when the effect of TOPO on increasing the viscosity, 7,of heptane solution is taken into account. The CETP values of 3-picoline at the viscosities of TOPO solutions in heptane were calculated from the CETP value in pure heptane by assuming a linear relationship between CETP and 7°.6,given the narrow range of viscosities, as indicated by the Knox equation (10). As seen from Table I, the viscosity of TOPO solution in heptane and the CETP values for 3-picoline and Pd(I1) increase with the concentration of TOPO. The CETP values for Pd remain distinctly higher, indicating that viscosity alone does not account for the increase in CETP for Pd with increasing TOPO concentration. It is also evident from this table that the CETP values for Pd(I1) are dependent on ita distribution ratio. The CETP values are larger a t higher distribution ratios, indicating poorer column efficiencies at higher D values. The additional peak broadening in the case of Pd(I1) compared to 3-picoline under identical conditions and at the higher distribution ratios of Pd(II) must arise from chemical kinetic factors. Approximately 50% of the 3-picoline molecules are protonated at pH = 6.1 (pK, = 6), at which ita CPC chromatograms were obtained. The deprotonation reaction will not cause CPC band broadening in the case of 3-picoline, as it is very rapid and indeed is termed diffusion controlled. In the manner of the Van Deemter and Knox equations in which the reduced plate height is seen as a s u m of contributions from various factors, we can express the observed CETP for P d (CETP,,,J as a sum of contributions from diffusion (CETP& and chemical kmetics (CETPO) (eq 1). CETP,b, = CETPdif CETPcK (1)
+
We can isolate the contribution due to the chemical kinetics to the observed CETP for Pd(I1) by subtracting the CETP values for 3-picoline at the appropriate TOPO concentrations. We shall show that CETPcK is related to the half-life for the kinetics of back-extraction of the PdC12(TOPO),complex and the D value of Pd(I1). A further examination of the involvement of chemical kinetic factors in the CPC efficiency for Pd(I1) was obtained by studying the dependence of CETP on the concentration of chloride in the aqueous phase and the concentration of TOPO in the heptane stationary phase. It is evident from eq 2 that the distribution ratio of Pd(I1) should depend on PdCl,(TOPO),
+ 2C1- + PdCld2-+ 2TOPO
(2)
the chloride and TOPO concentrations, with D being directly proportional to [TOP0I2and inversely proportional to [C1-I2. It is not obvious, however, that CETP should depend on these concentrations as well. It was found that increasing [Cl-] and decreasing [TOPO] provided better CPC efficiencies for Pd(II)
ANALYTICAL CHEMISTRY, VOL. 63,NO. 22, NOVEMBER 15, 1991
2844 O'
1
o
Table 11-Variation of kob.with Triton X-100, Chloride, and TOPO Concentrations
i
m
m
r
"
Retenton V0t.m~ [ n i l Figure 3. CPC chromatograms for Pd (lo-' M) at 0.5 M TOPO and [CI-] = 0.05-0.2 M. 01
E
8
4
Om 004
ace 0
[Cl-l, M
lO'[TOPO], M
kob, s-'
0.011 0.016 0.032 0.064 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016
0.1 0.1 0.1 0.1 0.05 0.2 0.3 0.4 0.5 0.6 0.5 0.5 0.5 0.5
1 1 1 1 1 1 1 1 1 1 5 10 20 30
15.99 17.05 17.97 16.49 7.90 32.18 46.59 61.36 80.64 96.54 75.00 66.72 56.50 47.52
that mass transfer of PdC12(TOPO), between the bulk aqueous and the bulk micellar phases was rapid. The value of kob increased with increasing [Cl-1 at constant [TOPO] and decreased with increasing [TOPO] at constant [Cl-1. These observations can be rationalized on the basis of the mechanism in eqs 3-5 for the decomposition of PdC12(TOP0)2. The rapid preequilibrium step, eq 3, and the
om
f
[Triton X-1001,M
PdC12(TOP0)2 o
i
m
Re"
m
m
v m ~~)
PdCl,(TOPO)
a
Flgure 4. CPC chromatograms for Pd [TOPO] = 0.1-0.5 M.
(lo3
M) at 0.05 M chloride and
(Figures 3 and 4). As may be seen, CETP is directly proportional to [TOPO] and inversely proportional to [Cl-] (Table I). These observations are consistent with the hypothesis that the kinetics of the back-extraction reaction, eq 2, is the major factor in affecting the CPC efficiency for Pd(I1). Study of the Back-Extraction of the PdC12(TOP0)2 Complex Using Stopped Flow. Since the CPC results indicated the involvement of slow chemical kinetics, a systematic study of the kinetics of the process of extraction and backextraction of Pd(I1) seemed in order. We learned from preliminary experiments of extraction and back-extraction of Pd(I1) by TOPO in the heptane:water system using the microporous Teflon phase separator (MTPS) (111, that these kinetics were too fast to be measured by this system which can measure half-lives 5 s and longer. In order to understand the influence of the kinetics of formation and dissociation of PdC12(TOP0)2on CETP, we resorted to stopped-flow analysis and examined the kinetics in the presence of micelles, which as we have demonstrated earlier, provide excellent models for extraction systems (12). The micellar system containing up to 4% Triton X-100, a neutral surfactant, to solubilize TOPO was chosen to compare the formation and dissociation rates of PdC12(TOP0)2. The formation of PdC12(TOP0)2under a variety of conditions was too fast even for the stopped-flow apparatus (limit of t l l z = 0.4 ms). The dissociation kinetics of PdC12(TOP0)2could be followed by the stopped-flow method. The PdCl,(TOP0)2 complex was formed in the Triton X-100 system using M Pd(I1). The dissociation of this complex under pseudo-first-order conditions in [Cl-] was monitored as a function of [Triton X-1001, [Cl-1, and [TOPO]. The observed pseudo-first-order rate constant was independent of the Triton X-100 concentration (Table 11), indicating that the dissociation of PdC12(TOP0)2occurred essentially in the bulk aqueous phase. This also indicated that no appreciable dissociation occurred in the micellar phase and
e K
+ TOPO
(3)
+ C1- 2 PdC13- + TOPO slow
(4)
PdCl,(TOPO)
+ C1-
PdC1,-
fast
PdC1,2-
(5)
rate-limiting step, eq 4, yield the following expression for kob under pseudo-first-order conditions in [Cl-1: kobs
=K
[Cl-] + k2K [TOPO]
Equation 6 can be rewritten as
+
-["-I= -
L[TOPO] (7) k2 Kk2 The plot of [C1-]/kob as a function of [TOPO] yielded a straight line. The k2 and K values derived from the intercept and slope of this plot are 168 f 8 M-ls-l and 0.004 f 0.0002 M, respectively. Equation 6 indicates two limiting cases, K >> [TOPO] and K
