Anal. Chem. 1990, 62, 2189-2193
2189
Characterization of the Oligomeric Dispersion of Poly(oxyalky1ene)diamine Polymers by Precolumn Derivatization and Capillary Zone Electrophoresis with Fluorescence Detection L a w r e n c e N. A m a n k w a , John Scholl, a n d Werner G. K u h r *
Department of Chemistry, University of California, Riverside, California 92521
The oligomer distribution of several poly(oxyaikylene)diamine (Jeffamine) polymers, ranging in molecular weight from 600 to 2000, has been characterized by precoiumn derivatiration and Capillary zone electrophoresis with fluorescence detection. The separation mechanism Is solely based on the differences between the electrophoretic mobilities of the oiigomer units. The effect of the precolumn derivatiration reaction on oligomer separation is also presented. Oligomer separation is only possible for a 1:l (sample-derivatizing agent) derivatized sample. I t Is possible by this method to separate 30 oligomers with an average molecular weight of 600 which differ in sire by only a methylene unit in less than 15 min with a separation efficiency of about 200000 theoretical plates.
INTRODUCTION Polymer products are among materials that are widely used both in industry and in households today. In consequence, there has been an increase in the number of these materials on the market. These products are most often formulated to consist of mixtures of isomers, homologues, and even oligomers having a wide distribution of molecular weight in order to meet specific physical and chemical requirements. The complexity of the chemical composition of these materials has increased the need for more reliable analytical methodologies for both process and quality control of these materials. High-performance liquid chromatography (HPLC) is one of the methods that is commonly used for separating and analyzing polymer composition (1-3). Nonionic surfactants containing different lengths of poly(ethy1ene oxide) oligomers have been separated and characterized according to the number of ethylene oxide units by normal-phase chromatography on an amino column. The separated components were monitored by both evaporative light scattering (ELS) and UV absorption detectors (1). In the same report, alkyl ether sulfates and sulfonates were separated and characterized by reversed-phase chromatography. Separations based on ion-exchange chromatography have also been widely used for characterizing oligomeric distribution of polymer products (4-6). High-resolution supercritical fluid chromatography has also been used extensively for studying the oligomeric distribution of polymer products (7-9). In one report (3,mixtures of alkyl ethoxylates were separated by using C 0 2 as mobile phase on a BP-10 column. Gel permeation chromatography (10, 11), field-flow fractionation (12),and mass spectrometry (13-16) have all been used as analytical techniques for characterizing polymer products. Among all the available techniques, mass spectrometry is the only method that can provide both structural information as well as the oligomeric distribution of a particular material. Its application in polymer charac0003-2700/90/0362-2189$02.50/0
terization is, however, less common owing to the high cost of instrumentation. In the past decade capillary zone electrophoresis (CZE) has developed into an efficient and fast analytical technique for the separation of complex mixtures such as amino acids, proteins, peptides, nucleotides, and biopolymers (17-21). Inspite of the high efficiency separation capability of CZE there has only been one reported application (22) of this technique for the separation and analysis of other polymeric materials. In that study, polystyrene nanospheres were separated by CZE. The authors showed that it was possible to resolve polystyrene particles in the size range 39-683 nm. In this case, the separation mechanism was believed to involve differences in surfactant-mediated particle-capillary wall interactions and not by differences in electrophoretic mobilities of the particles. In this report we describe the application of CZE with fluorescence detection for the separation and characterization of the oligomeric distribution of two poly(oxyalky1ene)diamine (Jeffamine-ED) polymers. These compounds have been suggested for use as antistatic and epoxy curing agents (23, 24). The separation mechanism utilized in this study is based on the differential electrophoretic mobilities (18)of the protonated NDA oligomers in the presence of an applied electric field. Eluted components (oligomers) are detected by monitoring the fluorescence emission of a fluorescent label (2,3naphthalenedialdehyde, NDA) attached by precolumn derivatization of each oligomer in the sample prior to separation. The effect of the degree of derivatization of the sample with the fluorescent label and the pH of the electrophoresis buffer on oligomer resolution is discussed. Finally, structural information on each separated oligomer is obtained by comparing the oligomer distribution obtained by CZE with the distribution independently obtained by Fourier transform mass spectrometry (FTMS) (15). EXPERIMENTAL SECTION Apparatus. Two CZE instruments were used in this study. One was built in our laboratory according to a design that has previously been described (21). The high-voltage (anodic) end of the high-voltage power supply (*50 kV; Glassman High Voltage, Inc., Model PS(EH50R02.0),Whitehouse Station, NJ) is isolated in a 20 cm X 20 cm X 20 cm Plexiglas enclosure which has an interlocked door to ensure operator safety. The inside of the enclosure is designed to accommodate a rotating (manual) sample tray which can hold up to eight 15-cm3sample vials. A 200-mR resistor (100 W) is connected to the high voltage end to provide a parallel resistance to the anode to serve as a current shunt to ground. The low voltage end of the separation capillary is connected to electrical ground. Both the cathode and the capillary are connected to the waste buffer vial through a vacuum septum. In this design the waste vial is connected via Teflon tubing (0.5 mm id.) to a low pressure (-9.0 in. of Hg) vacuum chamber maintained by a Vacuum regulator (Kontes Glass Co., Vineland, NJ). The injection end of the capillary is attached to a movable 0 1990 American Chemical Society
2190
ANALYTICAL CHEMISTRY, VOL. 62, NO. 20, OCTOBER
15, 1990
Plexiglas handle which enables switching of the capillary from a sample vial into an elution buffer vial and vice versa. Sampling can be achieved by either electromigration or vacuum injection. With injection by electromigratian, an electronic counter controls the time the injection voltage is applied. Vacuum injection is achieved by applying a constant low pressure to the low voltage end of the capillary for a specified length of time while the other end is immersed in sample solution. Timing in this case is controlled manually Detection was achieved on column by focusing the 442-nm (30 mW) output of a HeCd laser (Liconix Model 4240NB, Sunnyvale, CA) through a 1-cm focal length quartz lens, onto a spot near the low voltage end of the separation capillary that had been cleared of the polyimide coating. The capillary was mounted at Brewsters angle to the incident laser beam on an X-Y piezioelectric microdrive (Burleight Inchworm, Fisher, NY) stage. The fluorescence generated from the illumination of a fluorescent sample was collected perpendicular to the plane containing the capillary and the incident beam was imaged through a microscope objective, a spatial and two longpass glass color filters (Schott, No. G6495) onto a photomultiplier tube (Hamamatsu, R928). Data were either recorded on a strip chart recorder (Model No. EV41-885835, Kipp & Zonen, Holland) or acquired digitally after analog to digital (A/D) conversion and stored in a personal computer (Zenith Data Systems, St. Joseph, MI). The second instrument used in this work is a commerical electrophoresis instrument (Model 270A, Applied Biosystems, Inc., San Jose. CA) It is microprocessor controlled and consists of an autosampler, a vacuum injection system, an electromigration system, a thermostated (30-60 "C) capillary compartment and a UV-vis absorbance detector (25, 26). Additionally, this instrument is equipped with a prototype fluorescence detector, including a 75-W xenon arc lamp, powered by an auxiliary power supply (Model No. C2177-01. Hamamatsu, Japan), mounted in standard configuration Excitation wavelengths between 190 and 700 nm were selected via the inbuilt monochromator. The detector cell consists of 0 5-1.0 cm window on the capillary where the polyimide coating has been removed. This optical window is mounted in an assembly where the excitation light is focused radially by a spherical saphire lens into the center of the capillary so that the intensity of the light passing through the solution is maximized. The fluorescence emission is collected at right angles to the plane of the source beam and the capillary by two fiber optics, one on each side of the detector cell. Light from these fiber optics is filtered through a manually interchangeable 475-nm longpass glass color filter and detected by a photomultiplier tube. The instriiment can hr operated in both manual and automatic modes. Reagents. Water was distilled and deionized (Millipore, Bedford, MA). Methanol (Fisher Scientific, Fair Lawn, NJ) was of analytical grade and was used without further purification. Stock buffer was composed of 0.1 M trisodium phosphate (Fisher), 0.2 M boric acid (Sigma Chemical Co.. St. Louis, MO), and 0.05 M citric acid (Fisher). The electrophoresis buffer was made bv diluting the stock buffer with appropriate volumes of methanol and water such that the final concentration of the buffer components are half that of the stock buffer a t the appropriate pH (27) and also contain 20% methanol (v/v). The final buffer was filtered through a 0.45-pm (felman No. 60173 filter (Fisher) prior to use. Stock sample solutions (10.0 mM) were prepared by weighing appropriate amounts of Jeffamine (donated by Texaco Chemical Co., Bellaire, TX) and dissolving in 10.00 mL of water. These solutions were stable and hence were prepared weekly. NDA derivatizing reagent (10.0 mM) was prepared fresh daily by dissolving appropriate weights of NDA (Molecular Probes, Inc., Eugene, OR) in methanol. A stock solution (0.10 M) of NaCN (Mallinckrodt, Inc. Paris, KY) was prepared in water. Separation Capillary. Two sizes of fused silica capillaries were used in this study, one for each instrument. That used with the homebuilt instrument was 75 cm long (60 cm to the detector), 25 pm i.d. and 143 wm 0.d. That used with the Applied Biosystems Instrument was of the same length but was 50 pm i.d. and 370 pm o.d. Prior to use, each capillary was cleaned with 1.0 M NaOH for 10 min followed a IO-min rinse with water before filling with the electronhor buffer Cleaning was done hv connecting
one end of the capillary to a vacuum line (=m‘ d V ,
Figure 5. Electropherogram showing oligomer separation of Jeffamine ED 2001 by CZEILIFD. CZE parameters: sample concentration = 180 pM. Buffer pH = 4.2 and is composed of 77 mM boric acid, 19.4 mM citric acid, 11.3 mM trisodium phosphate in 20 % (v/v) methanol solution. Capillary, 100 cm (85 cm to detector), 25 pm i.d.; injection, 3 s at 30 kV; run, 25 kV; current = 1.8 pA.
the most intense peak in the mass spectra with the most intense peak in the electropherogram. The compositions of the rest of the peaks in the electropherogram will therefore correspond to that of corresponding peaks in the mass spectra. This assignment is purely based on conjecture; however, it is justified since both the electropherogram and the mass spectra are composed of peaks that are separated either directly or indirectly according to charge density. By this assignment oligomers that are separated in mass by only 2 amu were not resolved by CZE. Although baseline resolution was
Anal. Chem. 1990, 62, 2193-2198
not achieved for all the peaks, the separation efficiency for the most intense peak in Figure 4B corresponds to approximately 200 000 theoretical plates. Shown in Figure 5 is the electropherogram for Jeffamine ED-2001. The mass spectrum for the polymer (not shown) indicates a broad distribution of oligomers with molecular weights ranging from 1400 to 2400. Although the mass spectra peaks are also separated by 14 amu, it is apparent from Figure 5 that most of the oligomers were eluted together in the CZE experiment. Better resolution of these higher molecular weight components might be possible through the use of gels, which can more easily discriminate between higher molecular weight species (32). ACKNOWLEDGMENT The loan of a capillary electrophoresis system (Model 270A) by Applied Biosystems, Inc., and helpful conversations with Michael Albin and Steve Moring are gratefully acknowledged. LITERATURE CITED (1) Bear, G. R. J. Chromatogr. 1988,459, 91-107. (2) Desbene, P. L.; Desmazieres, B.; Basselier, J. J.; DesbensMonvernay, A. J. Chromatogr. 1989,461, 305-313. (3) Aserin, A.; Gartl. N.; Frenkei, M. J. Li9. Chromatogr. 1984, 7 (E), 1545-1557. (4) Ernst, J. M. Anal. Chem. 1984,56, 834-835. (5) Tsygankov, A. Y.: Motorin, Y. A.; Wolfson, A. D.; Kirpotin. B. D.; Oviovsky, A. F. J. Chromatogr. 1989,465,325-329. (6) Okada, T. Anal. Chem. 1990,62, 327-331. (7) Chester, T. L. J. Chromatogr. 1984,299, 424-434. (8)Richter, B. E. HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun. 1985,8 , 297-300. (9) Ashrafkhorssani, M.: Taylor, L. T. LC-GC 1990,8 (4), 314-320. (10) Paveiich, W. A.; Livigni, R. A. J. Po/ym. Sci., Part C 1968, 21, 215-223. (11) Garcia-Rubio, L. H.: McCregor, J. F.; Hamielec, A. E. Po/ymer Characterization; Craver, C. D., Ed.; American Chemical Society: Washington, DC, 1983;pp 31 1-344.
2193
(12) Giddings, J. C.; Graff, K. A.; Caldwell, K. D.; Myers, M. N. Po/ymer Characterization; Craver, C. D., Ed.; American Chemical Society: Washington, DC, 1983;pp 257-269. (13) Otsuki, A.; Shiraishi, H. Anal. Chem. 1979,57, 2329-2332. (14) Shiraishi, H.; Otsuki, A.; Fuwa, K. Bull. Chem. SOC.Jpn. 1982, 55, 1410- 1415. (15) Nuwaysir, L. M.; Wilkins, C. L.; Simonsick. W. J., Jr. J. Am. SOC. Mass Spectrom. 1990, 1, 66-71. (16) Siegel, M. M.; Tsao, R.; Oppenheimer, S.; Chang, T. T. Anal. Chem. 1990,62,322-327. (17) Jorgenson, J. W.; Lukacs. K. D. Science 1983,222, 260-272. (18) Jorgenson, J. W.; Lukacs, K. D. J. Chromatogr. 1981,278,209-216. (19) Kuhr, W. G. Anal. Chem. 1990,62, 404R-414R. (20) Lukacs, K. D.; Jorgenson, J. W. HRC CC, J. High Res. Chromatogr. Chromatogr. Commun. 1985,8 , 407-411. (21) Kuhr, W. G.; Yeung, E. S. Anal. Chem. l98& 6 0 , 1832-1834. (22) Vanorman, B. B.; McIntire, G. L. J. Microcolumn Sep. 1989, 1 (6), 289-293. (23) Pretka, J. E. (E. I. du Pont de Nemours & Co.) U S . Patent 3,021,232, Feb. 13, 1962. (24) Technical Report, Texaco Chemical Co.,Texas USA. (25) Grossman, P. D.; Lauer, H. H.; Moring, E. S.; Mead, D. E.;Oldham, F. M.; Nickel, J. H.; Gouberg, J. R. P.; Krever, A.; Ranson, D. H.; Coiburn, J. C. Am. Biotechnol. Lab. 1990,8(2), 35-43. (26) Moring, E. S.; Colburn, J. C.; Grossman, P. D.; Lauer, H. H. LC-GC 1990. 8 .111. 34-46. (27) Perrin, D. b.: Dempsey, B. Buffers for pH and Metal Ion Control; ChaDman and Hall: London. 1974. (28) Roach, M. C.; Harmony, M. 'D. Anal. Chem. 1987, 59, 411-415. (29) Carlson, R. G.; Srinivasachar, K.; Givens, R. S.; and B. K. Matuszewski, B. K. J. Org Chem. 1986,51,3978-3987. (30) Dissociation Constants of Organic Bases In Aqueous Solutions ; Perrin, D. D.. Dalzeli, D., Eds.; Butterworths: London, 1965. (31) Rose, J. D., Jr.; Jorgenson, J. W. J. Chromatogr. 1988, 447, 117-131. (32) Cohen, A. S.; Najarian, D. R.; Pauius, A.; A. Guttman, A,; Smith, J. A.; Karger, B. L. Proc. Natl. Acad. Sci. U . S . A . 1988,85,9660-9663.
....
RECEIVED for review May 10, 1990. Accepted July 16, 1990. This work was supported, in part, by the Society for Analytical Chemists of Pittsburgh, the National Science Foundation (Grant No. CHE-8957394), and the Office of Naval Research (Grant No. N00014-89-5-3071).
Absorption Detection in Capillary Electrophoresis by Fluorescence Energy Transfer Tommy W. G a r n e r a n d E d w a r d S.Yeung*
Ames Laboratory-USDOE and Department of Chemistry, Iowa State University, Ames, Iowa 50011
Normally, only highly fluorescent materials can be detected at low concentratlons In the small detectlon volumes typical of caplllary electrophoresls In a laser-based fluorometer. We report here the detectlon of absorbing but nonfluoresclng analytes by laser-excited fluorescence. This Is possible If the excited analytes transfer thelr energy to a fluorescent addltlve In the running buffer to Increase the backgroundfluorescence level. Two different fluorophores and four different absorbing analytes were tested In thls detectlon scheme. Concentrations as low as 6 X lo-' M and amounts as small as 4 amol at lnjectlon are detectable. The data support a long-range energy-transfer scheme, but the transfer efflclency Is much larger than those reported for other donor-acceptor palrs.
Recent research efforts in microcolumn separations, particularly in capillary electrophoresis (CE) (1,2) and in open tubular capillary liquid chromatography (OTCLC) (3-8),have
* To w h o m correspondence should b e addressed.
shown that very high separation efficiencies can be obtained. I t appears that 2-10 pm i.d. open tubes provide the best performance. In CE, these small diameters restrict the current flow and thus reduce the amount of band broadening induced by joule heating. In OTCLC, small diameters are needed to maximize the interaction with the stationary phase. On the other hand, the small elution volumes and the short optical path lengths available across these columns put severe restrictions on the selection of detection schemes for the analytes. The easiest way to detect analytes in capillary columns is by UV absorption (9-12). At 210 nm or so, most organic analytes have reasonable molar absorptivities. Indeed, the many commercial capillary electrophoresis systems on the market all rely on this detection mode. However, if one assumes an absorbance detection limit of 5 X for the detector, for an analyte with a fairly high molar absorptivity of lo4 L mol-l cm-', the concentration limit of detection (LOD) at the detector can be estimated to range only from 5 X to 1 X M for path lengths of 10-50 pm, respectively. In fact, as one goes to small columns, the decreasing amount of light available through the column degrades the absorbance
0003-2700/90/0362-2193$02.50/00 1990 American Chemical Society
