Anal. Chem. 2004, 76, 1857-1864
Surface Chemical Analysis of Carbohydrate Materials Used for Chromatography Media by Time-of-Flight Secondary Ion Mass Spectrometry Peter Sjo 1 vall,*,† Jukka Lausmaa,† Bo-Lennart Johansson,‡ and Mikael Andersson‡
Department of Chemistry and Materials Technology, SP Swedish National Testing and Research Institute, P.O. Box 857, SE-50115 Borås, Sweden, and Research and Development, Amersham Biosciences AB, SE-751 84 Uppsala, Sweden
The surface chemical structure of two raw materials (agarose and dextran) and four base matrixes used in the manufacture of chromatography media were analyzed using time-of-flight secondary ion mass spectrometry (TOF-SIMS). The results show that the small differences in molecular structure between these materials result in significant differences in the TOF-SIMS spectra and that these differences can be identified and quantified using either of two different approaches. In a novel approach, fragment ion distributions were extracted from the TOFSIMS spectra for each material, providing an immediate and systematic overview of the spectral features. Difference fragment distributions were used to highlight spectral differences between the materials. The results of the fragment ion distribution analysis, in terms of identification and quantification of spectral variations between different materials, were found to be in agreement with the results from a principal component analysis using the same set of data. Both methods were found capable of (i) distinguishing between agarose and dextran and (ii) detecting and quantifying the degree of cross-linking present in the four base matrix materials. In addition, using a deuterated chemical cross-linker, it was possible to identify spectral features specifically connected to the cross-link molecular structure. Liquid chromatography is one of the most important methods for the separation of chemical substances, for example, in biotechnology.1 The increasing demands on the pharmaceutical industry regarding consistency and purity of products continuously stimulate the development of new, well-characterized chromatography media. In this process, new and better characterization methods are needed in order to obtain more detailed chemical and structural information about these materials. The chemical properties of chromatography media have been characterized with a variety of spectroscopic methods, including total internal reflection fluorescence,2 UV-visible diffuse reflectance spectroscopy,3 Fourier transform infrared techniques,4 * Corresponding author: (e-mail)
[email protected]; (fax) +46 (33) 103388. † SP Swedish National Testing and Research Institute. ‡ Amersham Biosciences AB. (1) Sofer, G. J. Chromatogr., A 1995, 707, 23-28. (2) Rangnekar, V. M.; Oldham, P. B. Anal. Chem. 1990, 62, 1144-1147. (3) Rutan, S. C.; Harris, J. M. J. Chromatogr., A 1993, 656, 197. 10.1021/ac035457g CCC: $27.50 Published on Web 03/02/2004
© 2004 American Chemical Society
Raman spectroscopy,5 confocal scanning laser microscopy,6 electron spin resonance spectroscopy,7 nuclear magnetic resonance, and magnetic resonance imaging spectroscopy.8,9 Since the function of liquid chromatography media depends critically on the chemical composition on the surfaces, analysis methods that specifically probe the surface of these materials are particularly interesting. However, only a few such studies have been reported.10 Time-of-flight secondary ion mass spectrometry (TOF-SIMS) has during the last 10 years emerged as an important method for surface characterization of polymers.11 The high sensitivity and chemical specificity of this technique, in combination with its imaging capabilities and surface sensitivity, makes it a powerful tool for providing exclusive information about the chemical structure of polymer surfaces. A large variety of polymer systems and related issues have been studied with TOF-SIMS, including surface characterization of bulk polymers, copolymers, and polymer blends,12-14 polymer additives,15 polymer surface modification and functionalization,12,16 polymer aging,17 and molecular weight distribution.18,19 (4) Suffolk, B. R.; Gilpin, R. K. Anal. Chem. 1985, 57, 596-601. (5) Ho, M.; Cai, M.; Pemberton, J. E. Anal. Chem. 1997, 69, 2613-2616. (6) Ljunglo ¨f, A.; Larsson, M.; Knuuttila, K.-G.; Lindgren, J J. Chromatogr., A 2000, 893, 235-244. (7) Gilpin, R. K.; Kasturi, A.; Gelerinter, E. Anal. Chem. 1987, 59, 1177-1179. (8) Chen, S.; Qin, F.; Watson, A. T. AIChE J. 1994, 40, 1238. (9) Gladden, L. F.; Hollwand, M. P.; Alexander, P. AIChE J. 1995, 41, 894. (10) Barrett, D. A.; Brown, V. A.; Davies, M. C.; Shaw, P. N. Anal. Chem. 1996, 68, 2170-2178. (11) Delcorte, A. In ToF-SIMS: Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; IM Publications and SurfaceSpectra Limited: Charlton, Chichester, West Sussex, U.K., 2001; pp 161-194. (12) Briggs, D. In ToF-SIMS: Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; IM Publications and SurfaceSpectra Limited: Charlton, Chichester, West Sussex, U.K., 2001, pp 497-524. (13) Galuska, A. A. In ToF-SIMS: Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; IM Publications and SurfaceSpectra Limited: Charlton, Chichester, West Sussex, U.K., 2001, pp 525-542. (14) Eynde, X. V. In ToF-SIMS: Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; IM Publications and SurfaceSpectra Limited: Charlton, Chichester, West Sussex, U.K., 2001, pp 543-571. (15) Me´dard, N.; Poleunis, C.; Eynde, X. V.; Bertrand, P. Surf. Interface Anal. 2002, 34, 565-569. (16) McArthur, S. L.; Wagner, M. S.; Hartley, P. G.; McLean, K. M.; Griesser, H. J.; Castner, D. G. Surf. Interface Anal. 2002, 33, 924-931. (17) Mo ¨ller, K.; Jansson, A.; Sjo¨vall, P. Polym. Degrad. Stab. 2003, 80, 345352. (18) Eynde, X. V.; Bertrand, P. Surf. Interface Anal. 1997, 25, 878-888. (19) Coullerez, G.; Lundmark, S.; Malmstro ¨m, E.; Hult, A.; Mathieu, H. J. Surf. Interface Anal. 2003, 35, 693-708.
Analytical Chemistry, Vol. 76, No. 7, April 1, 2004 1857
base matrixes made from agarose, Sepharose 6B, Sepharose CL-6B, Sepharose 6 Fast Flow, and Sepharose High Performance, were studied in order to investigate the possibility to detect and quantify the degree of cross-linking in these materials using TOFSIMS. The results are analyzed using normalized fragment distributions extracted from the TOF-SIMS spectra and by PCA.
Figure 1. Principal molecular structures of (a) dextran and (b) agarose.
Although TOF-SIMS spectra intrinsically contain chemical information about the analyzed polymer surface, the interpretation of the spectra is often not straightforward. The TOF-SIMS spectrum from a polymer surface typically contains a large number of nonspecific fragment peaks in the low-mass region of the spectrum (< ∼150 u) and a few “characteristic” peaks, which are more or less specific for a particular functional group of the polymer, in the higher mass region.11 In the cases where no characteristic peak is observed, the distribution of peaks in the low-mass region may be used for identification in a “fingerprint” manner. To make use of as much information as possible from the TOF-SIMS spectra, multivariate statistical methods, such as principal component analysis (PCA), have been successfully used for the purpose of quantification and detection of small but systematic variations in TOF-SIMS spectra from different polymer samples.14-16,18,19 The principal chemical structures of the carbohydrate materials investigated in the present work are shown in Figure 1. The agarose structure consists of alternating D-galactose and 3,6anhydro-L-galactose structures, the latter of which include an anhydro bridge instead of the methoxy group normally present in the glucose structure. In the dextran structure, the glucose methoxy group is integrated in the bond between the ring structures. Despite these distinct differences in chemical structure between agarose and dextran, it is not trivial to find analysis methods, especially surface-sensitive methods, that can detect these differences. In terms of stoichiometry, the chemical state of the included elements, and the active group content, the two materials are very similar. In a recent work,20 we showed that the ligand structure of a number of different functionalized chromatography media could be identified and characterized in detail using TOF-SIMS. In the present work, TOF-SIMS is used to analyze different carbohydrate polymer materials used as raw materials and base matrixes in the manufacture of liquid chromatography media. Two commonly used raw materials with similar chemical structure, agarose and dextran (see Figure 1), were analyzed in order to investigate whether the chemical specificity of TOF-SIMS is sufficient to distinguish between these materials. Furthermore, four different (20) Johansson, B.-L.; Andersson, M.; Lausmaa, J.; Sjo¨vall, P. J. Chromatogr., A 2004, 1023, 49-56.
1858
Analytical Chemistry, Vol. 76, No. 7, April 1, 2004
EXPERIMENTAL SECTION Agarose and dextran (irregular-shaped particles with sizes of 50-300 µm) were obtained from Cambrex Corp. and Amersham Biosciences (Uppsala, Sweden), respectively. To achieve wellcharacterized molecular weight distributions, the materials were controlled and chemically adjusted by Amersham Biosciences prior to being supplied for analysis. The base matrixes, Sepharose 6B, Sepharose CL-6B, Sepharose 6 Fast Flow, and Sepharose High Performance, were obtained from Amersham Biosciences in the form of dried gel beads (∼20-150-µm diameter). Before analysis, the base matrixes were washed individually on a glass filter funnel: first in distilled water, then in 0.5 M NaCl solution, and finally in distilled water. The media were then dried by suction to remove excess water and finally completely dried at 43 °C in a Speedvac evaporation system under vacuum conditions (down to 1 mbar). The sample material was attached to a metal block using double-sided tape (3M, No. 665). A generous layer of sample powder was pressed onto the tape-covered metal block by using aluminum foil. The metal block was immediately mounted on the sample holder and introduced into the UHV chamber for analysis. TOF-SIMS spectra from clean tape were recorded for reference, to check for possible contamination of the samples from the tape. TOF-SIMS spectra were recorded on a TOF-SIMS IV instrument (ION-TOF GmbH) using a Ga+ primary ion beam (0.5 pA) and a low-energy electron flood gun for charge compensation. The analysis area was typically 200 × 200 µm2, and the analysis time was ∼100 s, always keeping the accumulated ion dose below the so-called static limit,21 1013 ions/cm2. The roughness and insulating properties of the sample particles caused a severe loss in mass resolution in the recorded TOFSIMS spectra. By obtaining an ion image of the analysis area and extracting new spectra from limited parts of the analysis area (one single or a few neighboring beads), it was possible to obtain a mass resolution of typically m/∆m ) 3000-4000. For each recorded spectrum, three to four new spectra were extracted. The spectrum analysis included 159 peaks corresponding to all detected CxHyOz+ fragment ions with masses up to 125 u. The signal intensities of these peaks were normalized to the total signal intensity of all selected peaks in each spectrum (the added normalized signal intensity is thus 1 for each spectrum). For each recorded spectrum, the peak intensities of the three to four extracted spectra were first added and then normalized. For agarose and dextran, a minimum of four spectra were recorded, two each from two separately prepared samples, while for the Sepharose materials, the presented results are from two different analysis areas from a single sample. The error bars in the presented diagrams represent the maximum and minimum values of the different recorded spectra from each material. In addition, (21) Vickerman, J. C. In ToF-SIMS: Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; IM Publications and SurfaceSpectra Limited: Charlton, Chichester, West Sussex, U.K., 2001; pp 1-40.
high impurity levels of foreign substances, e.g., phthalate. However, the results from the contaminated samples, analyzed using the methods described below, indicate that organic impurities can be observed, quantified, and identified using the technique presented in this work. PCA was carried out using a commercial software (The Unscrambler 6, Camo AS). The input data consisted of the normalized signal intensities of the 159 selected CxHyOz+ peaks as described above. The pretreatment further included mean-centering of the data, while autoscaling was not applied.
Figure 2. Positive TOF-SIMS spectra of dextran and agarose. The inset shows an expanded portion of the agarose spectrum at nominal mass 58 u.
the reproducibility of the results was controlled by analysis of additional samples, which, unfortunately, displayed unacceptably
RESULTS AND DISCUSSION Positive TOF-SIMS spectra for dextran and agarose are shown in Figure 2. The spectra are dominated by peaks corresponding to different CxHy+ and CxHyOz+ fragment ions, as expected from the carbohydrate structure of these materials (see Figure 1). The inset in the agarose spectrum shows a magnification of the C2H2O2+, C3H6O+, and C4H10+ peaks around 58 u, demonstrating that the spectra contain several, well-resolved peaks at each nominal mass. At higher masses, the intensities of the
Figure 3. CxHyOz+ fragment distributions obtained from TOF-SIMS spectra for dextran and agarose. See text for description.
Analytical Chemistry, Vol. 76, No. 7, April 1, 2004
1859
Figure 4. CxHyO+ fragment distributions obtained from TOF-SIMS spectra for dextran and agarose. See text for description.
CxHyOz+ fragment ion peaks gradually decreased and no “characteristic” peaks were observed. Observed peaks corresponding to impurities include Na+, Mg+, Al+, Si+, Ca+, and a phthalate fragment (149 u). As for the positive spectra, the negative spectra of agarose and dextran are dominated by peaks corresponding to CxHyOzfragments, however, with a more rapidly decreasing signal intensity with increasing ion mass than in the positive spectra. Due to experimental difficulties, it was not possible to obtain sufficient mass resolution, resulting in overlapping peaks and less reliable peak assignments. The negative spectra therefore contained less information than the positive ones and will not be further discussed in this work. Although no characteristic peaks were observed, the positive TOF-SIMS spectra show significant differences between dextran and agarose in the fingerprint region (
