536 Maguy Abi Jaoude´ Yannick Lassalle ´ ome ˆ Jer Randon Institut des Sciences Analytiques, Universite´ Claude Bernard Lyon1, Universite´ de Lyon, Villeurbanne, France Received September 24, 2013 Revised November 24, 2013 Accepted December 6, 2013
J. Sep. Sci. 2014, 37, 536–542
Research Article
Separation of xanthines in hydro-organic and polar-organic elution modes on a titania stationary phase Hydrophilic interaction LC was investigated in hydro-organic and nonaqueous elution modes on a titania column by using a set of N-methyl xanthines as neutral polar probes. To get information regarding the mechanisms that are behind the discrimination of these analytes in hydrophilic interaction, we focused our study on the type and amount of organic modifier as a critical yet rarely explored mobile phase parameter. Several alcohols such as methanol, ethanol, and isopropanol were studied as substitutes to acetonitrile in hydroorganic elution mode. Compared to silica, the investigation of the eluotropic series of these alcohols on titania highlighted a different implication in the retention mechanism of the xanthine derivatives. At low amounts of protic solvents, the adsorption mainly characterized the retention of analytes on bare silica; whereas mixed interactions including adsorption and ligand exchange were identified on native titania. To investigate the peculiar behavior of alcohols on the metal oxide, methanol, ethanol, and ethylene glycol were tested in replacement of water in polar-organic elution mode. Distinctive effects on the chromatographic retention and selectivity of xanthines were noticed for the dihydric alcohol, which was found to be a stronger eluting component than water on titania. Keywords: Eluotropic series / Hydrophilic interaction liquid chromatography / Polar organic elution / Titania / Xanthines DOI 10.1002/jssc.201301054
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction In 1990, Alpert [1] introduced hydrophilic interaction liquid chromatography (HILIC) as a viable substitute to classic normal-phase (NP) chromatography for the analysis of hydrophilic substances such as carbohydrates, peptides, and nucleic acids [2, 3]. According to the basic foundations of HILIC, the chromatographic retention is driven by the “partition” of analytes between a water-enriched layer, which is adsorbed onto a polar support, and a mobile phase that is usually composed of 5 to 40% v/v of water in acetonitrile (MeCN) [2–4]. From a practical perspective, the advantages of HILIC over classic NP chromatography are numerous. For instance, water traces originating from impurities in the organic solvent phase are no longer an issue since water is involved in the partition process and used as the strongest eluting solvent in the mobile phase. Moreover, better solubi´ ome ˆ Correspondence: Professor Jer Randon, Institut des Sciences Analytiques, Universite´ Claude Bernard Lyon1, Universite´ de Lyon, Villeurbanne, France E-mail: [email protected] Fax: 33-4-37-42-37-00
Abbreviations: HILIC, hydrophilic interaction liquid chromatography; NA, nonaqueous; NP, normal phase C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
lization of hydrophilic compounds can be achieved with the use of hydro-organic mobile phases instead of aprotic eluents. Finally, a high acetonitrile content in the mobile phase is highly desirable considering the great compatibility and sensitivity enhancement in ESI-MS [5, 6]. In the method development of HILIC separations, classic mobile phase parameters include the volume ratio of water, the pH, the salt concentration, the column temperature, and to a smaller extent the nature of the organic phase, which is studied as the weaker eluent [2–4]. Critical evaluations of these HILIC parameters have recently shown that the basic concept of “hydrophilic partition” does not fully explain the retention behavior of polar analytes in the case of specific interactions of the solutes with the surface of the chromatographic support (i.e. adsorption, ion, or ligand exchange) [7, 8]. To investigate the mechanisms underlying HILIC separations, Bicker et al. [9] have recently suggested the use of nonaqueous “polar-organic” eluents, as an extension of the classic HILIC mobile phase parameters. In their study, nonaqueous (NA) HILIC was performed by replacing H2 O in H2 O/MeCN mobile phases by organic protic modifiers such as alcohols. The investigation of HILIC and NA-HILIC mobile phase conditions demonstrated the dependence of the eluotropic series of the protic modifiers on the surface www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
chemistry of the chromatographic support. Therefore, the elution strength order of solvents in aqueous and nonaqueous HILIC modes can be a relevant parameter to provide complementary information and comprehensive understanding of mixed retention mechanisms on novel chromatographic materials. Recently, we investigated the use of native titania as a new polar material for HILIC or aqueous normal phase separations, using a set of neutral N-methyl xanthine derivatives, ionizable betablockers, and phosphorylated adenosines as polar probes [10–13]. The distinctive surface chemistry of titania is particularly characterized by mixed amphoteric and ligand exchange properties, which are both pH dependent. Anion or cation exchanges occur after protonation or deprotonation of hydroxyl groups (TiOH) on the surface, at pH values respectively lower or higher than the isoelectric point of the material (pHiep ∼ 5) [14, 15]. Ligand exchange interactions also take place on Ti4+ Lewis acid sites that exhibit high affinity toward electron donor molecules such as carboxylates and phosphates [16, 17]. Compared to silica, the surface chemistry of titania is still barely explored in separation science. Poor chromatographic efficiency is frequently observed with this material using conventional mobile phase composition, which currently limits its application to specific analytical separations. Despite a relatively poor knowledge of the interaction mechanisms involved, the original surface chemistry of titania explains its widespread use for sample preparation [18–23]. Further understanding of the chromatographic behavior of this original material is a prerequisite for future method development in separation sciences. In this context, the present work provides additional insights on HILIC performed on titania, with emphasis on the role of organic modifiers such as alcohols as weak or strong protic solvents in aqueous and nonaqueous elution conditions. Previously, we demonstrated in standard HILIC mode the ability of titania to discriminate N-methyl xanthines by the number and position of the methyl groups [11]. In this study, we focus our investigation on closely retained compounds such as 1,7-dimethylxanthine, 3-methylxanthine, and 1-methylxanthine, in relation to the key implication of the nitrogen atom at position 3 in their ligand exchange process on the surface of titania. An intercolumn comparison is also presented between titania and silica phases to bring out additional information on the involved interactions, based on the analysis of the eluotropic series of the organic modifiers and the retention profiles of the test solutes.
2 Materials and methods 2.1 Chemicals and materials Naphthalene, 1-methylxanthine (1MX), 3-methylxanthine (3MX), 1,7-dimethylxanthine (1,7DMX), 1,3dimethylxanthine (1,3DMX), 3,7-dimethylxanthine (3,7DMX), 1,3,7-trimethylxanthine (1,3,7TMX), and am C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
537
monium acetate were purchased from Sigma–Aldrich (Saint-Quentin Fallavier, France). HPLC-grade methanol and ethylene glycol (Saint-Quentin Fallavier, France), acetonitrile and isopropanol (VWR Prolabo, Fontenay sous Bois, France) in addition to ethanol (SDS, Vitry sur Seine, France) were used as organic solvents in the mobile phase. Ultra-purified water was collected from Elga Pure Lab UHQ system (Veolia Water STI, Le Plessis Robinson, France). R -NP titania columns (100 mm × Two different Sachtopore 2.1 mm id, dp 5 m [100 Å, 55 m2 /g] or [300 Å, 30 m2 /g]) from Zirchrom Separations (Anoka, MN, USA) and a R SILICA column (50 mm × 4.6 mm id, dp HYPERSIL 3 m [120 Å, 170 m2 /g]) from Thermo Electron Corporation (Runcorn, UK) were used in this work.
2.2 Instrumentation LC was performed using a Waters 2690 HPLC system with a Waters 996 photodiode array UV detector (Milford, MA, USA). The chromatographic data were processed using the R Pro software. Temperature regulation Waters Empower of chromatographic columns was carried with a Selerity PolarathermTM series 9000 oven (Salt Lake City, UT, USA).
2.3 Chromatographic method The test solutions were prepared by dissolving each compound individually with naphthalene (the column hold-up volume marker) in 66:34 v/v MeCN/H2 O. The solutions were sonicated for 30 min prior to injection (injection volume, 5 L). The eluents were prepared by dissolving ammonium acetate at a total concentration of 5 × 10−3 mol/L in MeCN-based mobile phases (no further pH adjustment) and were then sonicated during 30 min before use. Since water molecules cannot be totally removed from the surface of native titania, to perform polar-organic NA-HILIC experiments under anhydrous conditions, a residual amount of 1% v/v H2 O was introduced in both solvents of the mobile phase, which also enhanced the solubility of the ammonium acetate additive. For both aqueous and polar organic elution modes, the chromatographic columns were first equilibrated for an equivalent time of 30 void volumes.
3 Results and discussion In LC, investigating the role of the polar organic solvent is central to get a better understanding of the chemical interactions occurring on the surface of a separation material. To study the influence of the type of organic solvent, such as alcohols in hydrophilic interaction chromatography on native titania, two elution modes were considered: the hydroorganic mode as in classic HILIC (organic solvent/H2 O), and the polar organic mode for nonaqueous elution conditions (MeCN/organic solvent). www.jss-journal.com
538
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
Figure 1. Effect of water content on the retention of N-methyl xanthine derivatives ˚ 170 on (A) a silica column SiO2 (120 A, ˚ m2 /g) and (B) a titania column TiO2 (300 A, 30 m2 /g). Mobile phase, ammonium acetate (5 × 10−3 mol/L) in MeCN/H2 O (100−x):x, v/v; column temperature, 30⬚C; detection, UV 272 nm; solutes: 1MX, 3MX, 1,7DMX (300 mg/L) in MeCN/H2 O 66:34, v/v; injection volume, 5 L; column hold-up volume marker, naphthalene.
3.1 Hydro-organic elution mode—classic HILIC 3.1.1 MeCN/H2 O, effect of H2 O content in MeCN Since water is by reference a standard component in HILIC mobile phases [2–4], the retention profiles of 1,7DMX, 3MX, and 1MX were initially investigated as a function of the water content in the presence of ammonium acetate (5 ×10−3 mol/L) as salt additive on native silica (120 Å, 170 m2 /g), and titania (300 Å, 30 m2 /g). A typical chromatogram of xanthines separation on titania is shown in the Supporting Information. According to Fig. 1A and B, the retention of the set of xanthines on both silica and titania columns typically decreased with water amounts ranging from 2 to 40% v/v. The regularity of the elution patterns of the xanthines was further analyzed according to the well-established models of partition log10 k = f(%H2 O) and adsorption log10 k = f(log10 (% H2 O)) [11]. Assuming each of these two mechanisms on silica and titania, the calculated values of the coefficients of determination (Supporting Information Table S1) were closest to 1 for the adsorption model on both materials. On native silica, the adsorption of neutral xanthines would typically engage hydrogen bonding and dipolar interactions with the silanols groups (Si–OH) and siloxane bridges (Si–O–Si). On native titania, a multimodal retention regime is suspected since a greater retention was found on this column (Fig. 1A and B) despite the lower value of the specific surface area of the material (i.e. 30 m2 /g for titania compared to 170 m2 /g for the silica column). Unlike silica, the titania column was able to discriminate the two mono-N-methyl xanthine isomers, 3MX and 1MX, despite their similar molecular polarity. The position of the methyl group seems to be fundamental in the molecular interaction of the xanthine derivatives with the surface of titania, and responsible for such high selectivity. In addition to nonspecific adsorption on Ti–OH and Ti–O–Ti groups, the retention of the tested xanthine derivatives on titania (300 Å) is apparently governed by ligand exchange. Such a mechanism would occur between the water molecules occupying the Ti4+ Lewis acid sites and the free nitrogen atoms (i.e. N1, N3, or N7) of the xanthines, as we previously observed for titania (100 Å) [11].
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.1.2 X/H2 O, effect of the nature of the organic modifier Four different organic modifiers, namely MeOH, EtOH, i-PrOH, and MeCN were used as weak solvents added to water for hydrophilic chromatography of xanthines on silica and titania columns. Figure 2A and B depict the evolution of the retention factor of 1MX, 3MX, and 1,7DMX versus the type of organic modifier for both silica and titania columns. With 2% water, alcohols on silica were found to be stronger eluting components than acetonitrile. The eluotropic series of the alcohols typically followed their adsorption energy (g o ) on silica: H2 O MeOH (g o = 0.73) > EtOH (g o = 0.68) > i-PrOH (g o = 0.60) > MeCN (g o = 0.50) [24]. The elution strength order of the alcohols highlights the main implication of their protic character as competitor to the hydrogen bonding adsorption of xanthines on silica. For hydro-alcoholic eluents, the retention of 1,7DMX is always higher than the retention of the two mono-N-methyl xanthines (1MX and 3MX); whereas this order is reversed with MeCN. For all organic modifiers, the selectivity between the two monomethyl isomers remains very low on silica. With 10% H2 O, typical solvent effects were observed on titania compared to silica (Fig. 2B). The elution strength order of the different solvents on titania decreased in the sequence: i-PrOH > EtOH MeCN > MeOH. The eluotropic series of the different alcohols is completely modified on titania and methanol mostly acted as weaker eluent than acetonitrile. This means that the adsorption forms and pathways of alcohols on titania are totally different from that observed on silica. Several infrared/thermodesorption studies have previously demonstrated that reversible surface hydroxylation/dehydroxylation may occur upon adsorption of alcohols on titanium dioxide (anatase) at room temperature [25–28]. This sorption pathway is established through coordination bonding with Ti4+ Lewis acid sites, and involves reversible deprotonation/protonation of the adsorbed alcohols, which typically result in a surface populated by chemisorbed alkoxy groups (RO) and intact alcohol molecules (ROH) [25, 27–29]. In contrast to acetonitrile, alcohols possess hydroxyl groups that can act as competitors to the coordination process of
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
539
Figure 2. Effect of the nature of the organic modifier on the retention of N-methyl xanthine derivatives on (A) silica column ˚ 170 m2 /g), ammonium acetate SiO2 (120 A, (5 × 10−3 mol/L) in organic modifier/H2 O 98:2, v/v and (B) titania column TiO2 ˚ 30 m2 /g), ammonium acetate (300 A, (5 × 10−3 mol/L) in organic modifier/H2 O 90:10, v/v. Other experimental conditions are identical to Fig. 1, N = 3 repetitions, error bars: ±1 SD.
xanthines on titania. Based on this assumption, the ligand exchange between analyte (L), alcohol (ROH), and water on titania can be described using the following equilibria: Ti(OH)(H2 O) + ROH ↔ Ti(OH)(ROH) + H2 O
(1)
Ti(OH)(H2 O) + ROH ↔ Ti(OR)(H2 O) + H2 O
(2)
Ti(OH)(H2 O) + L ↔ Ti(OH)(L) + H2 O
(3)
Ti(OR)(H2 O) + L ↔ Ti(OR)(L) + H2 O
(4)
Ti(OH)(ROH) + L ↔ Ti(OH)(L) + ROH
(5)
Ti(OR)(H2 O) + L ↔ Ti(L)(OH) + ROH
(6)
Where Equations (1) and (2) refer respectively to associative and dissociative adsorption of alcohol on titania. Equations (3) and (4) account for exchange of coordinated water with a ligand analyte; whereas Equations (5) and (6) describe, respectively, the displacement of undissociated alcohol and alkoxide group from coordination by a ligand analyte. Considering the coordination hypothesis of alcohols on titania, the stability of titanium alcoholic complexes should increase with the length of the alkyl chain (R) as follows: MeOH < EtOH < i-PrOH. This predicated order is consistent with the eluotropic series obtained in hydro-alcoholic conditions on titania (Fig. 2B). Since methanol is the weakest eluent among the other alcohols, we assume on account of Equations (1), (2), (5) and (6), that this solvent is more easily displaced by water and analyte molecules during ligand exchange.
3.2 Polar organic elution mode—NA-HILIC 3.2.1 MeCN/MeOH, effect of MeOH content in MeCN To confirm whether methanol or acetonitrile is the weakest eluent on titania, the retention of xanthine derivatives was C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
investigated as a function of the MeOH content in MeCN in NA-HILIC. Figure 3A and B compares the results obtained for caffeine (1,3,7TMX), theophylline (1,3DMX), theobromine (3,7DMX), and 3-methylxanthine (3MX) in NA-HILIC on TiO2 (100 Å) with the data previously obtained in standard HILIC using water as the protic modifier in MeCN [11]. As a general observation, the retention factors of all the xanthines are greater in MeCN/MeOH (k > 20; for 1,7DMX and 1MX), despite a higher concentration of acetate competitor in the NA-HILIC mobile phase. Thus substituting water with methanol reduces the overall polarity and elution strength of the mobile phase. As we have already observed for MeCN/H2 O, two opposite regimes of retention were also observed for MeCN/MeOH when the amount of MeOH varied from 5 to 99% v/v in MeCN (Fig. 3A). Methanol acts as the strongest solvent for MeOH contents lower than 70% in MeCN, and above this value, the elution strengths of solvents were inverted, leading to a progressive increase of the retention factor of 3MX with MeOH content. The inversion of the elution strengths of solvents in NA-HILIC at high concentrations in MeCN still remains unclear on titania as it is for silica. For methanol contents lower than 50% in MeCN/MeOH, the retention data of the xanthine derivatives were plotted according to the models of partition log10 k = f(%MeOH) and adsorption log10 k = f(log10 (%MeOH)) (Fig. 4) to gain insight on the retention mechanisms involved in this range of composition. According to the correlation coefficients values (Supporting Information Table S2), the linearity of the different elution patterns fitted the adsorption mechanism for both polar-organic and hydro-organic modes. 3.2.2 MeCN/X, effect of the nature of the polar modifier The evolution of the retention factor of 3MX on titania (100 Å) was evaluated according to the type of alcohol (Fig. 5) for MeCN/X mobile phases in order to establish a scale of elution strength for methanol, ethanol, ethylene glycol, and water in acetonitrile. The eluotropic series at 10% v/v in MeCN follows the sequence: (CH2 OH)2 H2 O > EtOH > MeOH. The elution strength order between ethanol and methanol in acetonitrile is identical to that observed in water using hydroalcoholic mixtures (Section 3.1.2). It is interesting to notice www.jss-journal.com
540
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
Figure 3. Retention factors of Nmethyl xanthine derivatives on titania R ˚ Sachtopore -NP (100 A) as a function of the protic modifier content. (A) MeCN/MeOH [(99–x ):x , v/v] + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L) and (B) MeCN/H2 O [(100– x):x, v/v], ammonium acetate (1 × 10−3 mol/L). Other experimental conditions are identical to Fig. 1, N = 3 repetitions, error bars: ±1 SD.
Figure 4. Elution patterns of N-methyl xanthine derivatives on titania for MeOH contents less than 50% in MeCN/MeOH. (A) Partition model, log10 k = f(%MeOH) and (B) adsorption model, log10 k = f(log10 (%MeOH). Experimental conditions are listed in Fig. 3A.
hydrogen-bonded molecular networks and macro-clusters on the surface of titanium dioxide. 3.2.3 MeCN/X, ethylene glycol versus water
Figure 5. Retention factors of 3MX on titania using methanol, ethanol, ethylene glycol, and water as protic modifiers in MeCN. R ˚ mobile Chromatographic column, TiO2 Sachtopore -NP (100 A); phase, MeCN/protic modifier (90:10, v/v) + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L). Other experimental conditions are similar to Fig.1., N = 3 repetitions; error bars: ±1 SD.
that ethylene glycol is considerably stronger than water on titania, since the retention factor of 3MX was divided by a factor of 10 upon use of this alcohol as protic modifier in NA-HILIC. Bicker et al. [9] earlier reported a similar but less significant behavior of ethylene glycol on bare silica. As a dihydric solvent (i.e. diol), ethylene glycol has a large dipole moment (2.36 D versus 1.85 D for water) [30] and two OH groups that are responsible for the substantial adsorption of this alcohol on oxide surfaces. Sekuli´c et al. [31] reported that free and adsorbed ethylene glycol molecules can even establish C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The retention of 1,7DMX, 3MX, and 1MX on titania (100 Å) was evaluated as a function of the amount of ethylene glycol to further characterize the effects of the diol as strong protic modifier in NA-HILIC. In comparison with water, the retention data obtained for less than 50% v/v ethylene glycol contents in acetonitrile were plotted in the logarithmic scale, assuming the adsorption model. According to Fig. 6A and B, the retention factors of the xanthine derivatives decrease with increasing amounts of ethylene glycol or water in the mobile phase. The elution patterns typically followed the adsorption model for both solvents. The substitution of water with ethylene glycol led to lower retentions on titania and the selectivity between the xanthines remained important in all of these polar organic conditions. The low retention of the analytes in the presence of ethylene glycol highlights substantial bonding between the diol and the surface of the metal oxide. The desorption of ethylene glycol from the surface of titania appeared as a slow process and a special washing procedure had to be implemented in order to recover the native titania surface. The retention factor of 1,7DMX using ammonium acetate 5 × 10−3 mol/L in MeOH/H2 O (90:10 v/v) as mobile phase was originally close to 25, but after using ethylene glycol as mobile phase, the retention factor of 1,7DMX with the same ammonium acetate mobile phase was three www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
541
Figure 6. Elution profiles of N-methyl xanthines on titania for contents less than 50% of ethylene glycol (A) and water (B). ChroR matographic column, titania Sachtopore ˚ mobile phase, MeCN/protic NP (100 A); modifier ((99–x):x, v/v) + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L). Other experimental conditions are identical to Fig. 1. Plots are represented in decimal logarithmic scale, assuming the adsorption model: R2 (EG) > 0.95; R2 (water) ∼ 0.99.
times lower than previously observed. A 2 h column wash with (50:50, v/v) MeOH/NH3 ,H2 O (0.3% v) at 1 mL/min was not efficient enough and only a 12 h treatment at 60⬚C with (50:50, v/v) MeOH/H2 O at 1 mL/min was able to restore the retention of 1,7DMX. According to Sekuli´c et al. [31], the presence of high amounts of water in contact with titania may disrupt the physisorbed ethylene glycol network and macroclusters through hydrogen bonding, so free water molecules will have access to coordination sites and will progressively replace the chemisorbed ethylene glycol. We assume that a high washing temperature can improve the diffusion of free water molecules into the pores of titania, whereas the presence of a hard Lewis base such as ammonia in the washing solution can accelerate the exchange of ethylene glycol for hydroxide anions. As great ligand competitor on titania, biocompatible ethylene glycol, used to replace water in acetonitrile-based mobile phases, may potentially offer novel selectivity for extended applications of titania in sample treatment of complex biological mixtures.
classic HILIC and led also to interesting selectivities between the xanthines. The authors have declared no conflict of interest.
5 References [1] Alpert, A. J., J. Chromatogr. 1990, 499, 177–196. ¨ [2] Hemstrom, P., Irgum, K., J. Sep. Sci. 2006, 29, 1784–1821. [3] Jandera, P., J. Sep. Sci. 2008, 31, 1421–1437. [4] Gama, M. R., Silva, R. G. C., Collins, C. H., Bottoli, C. B. G., TrAC Trends Anal. Chem. 2012, 37, 48–60. [5] Kind, T., Tolstikov, V., Fiehn, O., Weiss, R. H., Anal. Biochem. 2007, 363, 185–195. [6] Garbis, S. D., Melse-Boonstra, A., West, C. E., Van Breemen, R. B., Anal. Chem. 2001, 73, 5358–5364. [7] Buszewski, B., Noga, S., Anal. Bioanal. Chem. 2012, 402, 231–247. [8] Pesek, J. J., Matyska, M. T., Boysen, R. I., Yang, Y., Hearn, M. T. W., TrAC Trends Anal. Chem. 2013, 42, 64–73.
4 Conclusions The type and amount of protic modifier is an interesting mobile phase parameter that can be carefully exploited to tune the chromatographic retention and selectivity of polar compounds in hydrophilic interaction LC. The evaluation of the eluotropic series of alcohols as protic modifiers in both HILIC and NA-HILIC elution modes provided, in combination with intercolumn comparison, valuable comprehension of the retention mechanisms of N-methyl xanthines on silica and titania. The retention of neutral xanthine at low contents of protic modifiers exhibited simple molecular physisorption on silica and a mixed-mode mechanism composed of adsorption and ligand exchange on titania. Adsorptive interactions of analytes on silica are strengthened when the polarity of the protic solvent is reduced. However, the elution strength of protic modifiers on titania is highly influenced by their Lewis base properties, since these solvents are directly involved in a ligand-exchange competition process with the analytes. Particularly, the use of ethylene glycol as organic protic modifier in NA-HILIC provided faster analysis time than water did in C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[9] Bicker, W., Wu, J. Y., Lammerhofer, M., Lindner, W., J. Sep. Sci. 2008, 31, 2971–2987. [10] Randon, J., Guerrin, J.-F., Rocca, J.-L., J. Chromatogr. A. 2008, 1214, 183–186. ´ M., Randon, J., J. Chromatogr. A 2011, 1218, [11] Abi Jaoude, 721–725. ´ M., Randon, J., Anal. Bioanal. Chem. 2011, [12] Abi Jaoude, 400, 1241–1249. ´ M., Morin, N., Miege, ` [13] El Debs, R., Abi Jaoude, C., Randon, J., J. Sep. Sci. 2011, 34, 1805–1810. [14] Winkler, J., Marme, S., J. Chromatogr. A 2000, 888, 51–62. ¨ M., Kurganov, A. A., Schacht, S., Schuth, ¨ [15] Grun, F., Unger, K. K., J. Chromatogr. A 1996, 740, 1–9. [16] Zhou, T., Lucy, C. A., J. Chromatogr. A 2008, 1187, 87–93. [17] Zhou, T., Lucy, C. A., J. Chromatogr. A 2010, 1217, 82–88. [18] Ikeguchi, Y., Nakamura, H., Anal. Sci. 1997, 13, 479–483. [19] Kuroda, I., Shintani, Y., Motokawa, M., Abe, S., Furuno, M. Anal. Sci. 2004, 20, 1313–1319. [20] Pinkse, M. W. H., Uitto, P. M., Hilhorst, M. J., Ooms, B., Heck, A. J. R., Anal. Chem. 2004, 76, 3935–3943.
www.jss-journal.com
542
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
[21] Sekiguchi, Y., Mitsuhashi, N., Inoue, Y., Yagisawa, H., Mimura, T., J. Chromatogr. A 2004, 1039, 71–76.
[26] Kim, K. S., Barteau, M. A., Farneth, W. E., Langmuir 1988, 4, 533–543.
[22] Hata, K., Morisaka, H., Hara, K., Mima, J., Yumoto, N., Tatsu, Y., Furuno, M., Ishizuka, N., Ueda, M., Anal. Biochem. 2006, 350, 292–297.
[27] Shepot’ko, M. L., Davydov, A. A., Theor. Exp. Chem. 1991, 27, 210–214.
[23] Richardson, B. M., Soderblom, E. J., Thompson, J. W., Moseley, M. A., J. Biomol. Tech. 2013, 24, 8–16. [24] Meyer, V. R., Practical High-Performance Liquid Chromatography, 3rd Edition, John Wiley & Sons, Hoboken, 1998. [25] Ramis, G., Busca, G., Lorenzelli, V., J. Chem. Soc. Faraday Trans. I 1987, 83, 1591–1599.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[28] Wang, C., Groenzin, H., Shultz, M. J., J. Am. Chem. Soc. 2005, 127, 9736–9745. [29] Lusvardi, V. S., Barteau, M. A., Farneth, W. E., J. Catal. 1995, 153, 41–53. [30] Lide, D. R., CRC Handbook of Chemistry and Physics, 82th Edition, CRC Press, Boca Raton 2000. ´ J., Elshof, J. E. T., Blank, D. H. A., Langmuir 2005, [31] Sekulic, 21, 508–510.
www.jss-journal.com
J. Sep. Sci. 2014, 37, 536–542
Research Article
Separation of xanthines in hydro-organic and polar-organic elution modes on a titania stationary phase Hydrophilic interaction LC was investigated in hydro-organic and nonaqueous elution modes on a titania column by using a set of N-methyl xanthines as neutral polar probes. To get information regarding the mechanisms that are behind the discrimination of these analytes in hydrophilic interaction, we focused our study on the type and amount of organic modifier as a critical yet rarely explored mobile phase parameter. Several alcohols such as methanol, ethanol, and isopropanol were studied as substitutes to acetonitrile in hydroorganic elution mode. Compared to silica, the investigation of the eluotropic series of these alcohols on titania highlighted a different implication in the retention mechanism of the xanthine derivatives. At low amounts of protic solvents, the adsorption mainly characterized the retention of analytes on bare silica; whereas mixed interactions including adsorption and ligand exchange were identified on native titania. To investigate the peculiar behavior of alcohols on the metal oxide, methanol, ethanol, and ethylene glycol were tested in replacement of water in polar-organic elution mode. Distinctive effects on the chromatographic retention and selectivity of xanthines were noticed for the dihydric alcohol, which was found to be a stronger eluting component than water on titania. Keywords: Eluotropic series / Hydrophilic interaction liquid chromatography / Polar organic elution / Titania / Xanthines DOI 10.1002/jssc.201301054
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction In 1990, Alpert [1] introduced hydrophilic interaction liquid chromatography (HILIC) as a viable substitute to classic normal-phase (NP) chromatography for the analysis of hydrophilic substances such as carbohydrates, peptides, and nucleic acids [2, 3]. According to the basic foundations of HILIC, the chromatographic retention is driven by the “partition” of analytes between a water-enriched layer, which is adsorbed onto a polar support, and a mobile phase that is usually composed of 5 to 40% v/v of water in acetonitrile (MeCN) [2–4]. From a practical perspective, the advantages of HILIC over classic NP chromatography are numerous. For instance, water traces originating from impurities in the organic solvent phase are no longer an issue since water is involved in the partition process and used as the strongest eluting solvent in the mobile phase. Moreover, better solubi´ ome ˆ Correspondence: Professor Jer Randon, Institut des Sciences Analytiques, Universite´ Claude Bernard Lyon1, Universite´ de Lyon, Villeurbanne, France E-mail: [email protected] Fax: 33-4-37-42-37-00
Abbreviations: HILIC, hydrophilic interaction liquid chromatography; NA, nonaqueous; NP, normal phase C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
lization of hydrophilic compounds can be achieved with the use of hydro-organic mobile phases instead of aprotic eluents. Finally, a high acetonitrile content in the mobile phase is highly desirable considering the great compatibility and sensitivity enhancement in ESI-MS [5, 6]. In the method development of HILIC separations, classic mobile phase parameters include the volume ratio of water, the pH, the salt concentration, the column temperature, and to a smaller extent the nature of the organic phase, which is studied as the weaker eluent [2–4]. Critical evaluations of these HILIC parameters have recently shown that the basic concept of “hydrophilic partition” does not fully explain the retention behavior of polar analytes in the case of specific interactions of the solutes with the surface of the chromatographic support (i.e. adsorption, ion, or ligand exchange) [7, 8]. To investigate the mechanisms underlying HILIC separations, Bicker et al. [9] have recently suggested the use of nonaqueous “polar-organic” eluents, as an extension of the classic HILIC mobile phase parameters. In their study, nonaqueous (NA) HILIC was performed by replacing H2 O in H2 O/MeCN mobile phases by organic protic modifiers such as alcohols. The investigation of HILIC and NA-HILIC mobile phase conditions demonstrated the dependence of the eluotropic series of the protic modifiers on the surface www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
chemistry of the chromatographic support. Therefore, the elution strength order of solvents in aqueous and nonaqueous HILIC modes can be a relevant parameter to provide complementary information and comprehensive understanding of mixed retention mechanisms on novel chromatographic materials. Recently, we investigated the use of native titania as a new polar material for HILIC or aqueous normal phase separations, using a set of neutral N-methyl xanthine derivatives, ionizable betablockers, and phosphorylated adenosines as polar probes [10–13]. The distinctive surface chemistry of titania is particularly characterized by mixed amphoteric and ligand exchange properties, which are both pH dependent. Anion or cation exchanges occur after protonation or deprotonation of hydroxyl groups (TiOH) on the surface, at pH values respectively lower or higher than the isoelectric point of the material (pHiep ∼ 5) [14, 15]. Ligand exchange interactions also take place on Ti4+ Lewis acid sites that exhibit high affinity toward electron donor molecules such as carboxylates and phosphates [16, 17]. Compared to silica, the surface chemistry of titania is still barely explored in separation science. Poor chromatographic efficiency is frequently observed with this material using conventional mobile phase composition, which currently limits its application to specific analytical separations. Despite a relatively poor knowledge of the interaction mechanisms involved, the original surface chemistry of titania explains its widespread use for sample preparation [18–23]. Further understanding of the chromatographic behavior of this original material is a prerequisite for future method development in separation sciences. In this context, the present work provides additional insights on HILIC performed on titania, with emphasis on the role of organic modifiers such as alcohols as weak or strong protic solvents in aqueous and nonaqueous elution conditions. Previously, we demonstrated in standard HILIC mode the ability of titania to discriminate N-methyl xanthines by the number and position of the methyl groups [11]. In this study, we focus our investigation on closely retained compounds such as 1,7-dimethylxanthine, 3-methylxanthine, and 1-methylxanthine, in relation to the key implication of the nitrogen atom at position 3 in their ligand exchange process on the surface of titania. An intercolumn comparison is also presented between titania and silica phases to bring out additional information on the involved interactions, based on the analysis of the eluotropic series of the organic modifiers and the retention profiles of the test solutes.
2 Materials and methods 2.1 Chemicals and materials Naphthalene, 1-methylxanthine (1MX), 3-methylxanthine (3MX), 1,7-dimethylxanthine (1,7DMX), 1,3dimethylxanthine (1,3DMX), 3,7-dimethylxanthine (3,7DMX), 1,3,7-trimethylxanthine (1,3,7TMX), and am C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
537
monium acetate were purchased from Sigma–Aldrich (Saint-Quentin Fallavier, France). HPLC-grade methanol and ethylene glycol (Saint-Quentin Fallavier, France), acetonitrile and isopropanol (VWR Prolabo, Fontenay sous Bois, France) in addition to ethanol (SDS, Vitry sur Seine, France) were used as organic solvents in the mobile phase. Ultra-purified water was collected from Elga Pure Lab UHQ system (Veolia Water STI, Le Plessis Robinson, France). R -NP titania columns (100 mm × Two different Sachtopore 2.1 mm id, dp 5 m [100 Å, 55 m2 /g] or [300 Å, 30 m2 /g]) from Zirchrom Separations (Anoka, MN, USA) and a R SILICA column (50 mm × 4.6 mm id, dp HYPERSIL 3 m [120 Å, 170 m2 /g]) from Thermo Electron Corporation (Runcorn, UK) were used in this work.
2.2 Instrumentation LC was performed using a Waters 2690 HPLC system with a Waters 996 photodiode array UV detector (Milford, MA, USA). The chromatographic data were processed using the R Pro software. Temperature regulation Waters Empower of chromatographic columns was carried with a Selerity PolarathermTM series 9000 oven (Salt Lake City, UT, USA).
2.3 Chromatographic method The test solutions were prepared by dissolving each compound individually with naphthalene (the column hold-up volume marker) in 66:34 v/v MeCN/H2 O. The solutions were sonicated for 30 min prior to injection (injection volume, 5 L). The eluents were prepared by dissolving ammonium acetate at a total concentration of 5 × 10−3 mol/L in MeCN-based mobile phases (no further pH adjustment) and were then sonicated during 30 min before use. Since water molecules cannot be totally removed from the surface of native titania, to perform polar-organic NA-HILIC experiments under anhydrous conditions, a residual amount of 1% v/v H2 O was introduced in both solvents of the mobile phase, which also enhanced the solubility of the ammonium acetate additive. For both aqueous and polar organic elution modes, the chromatographic columns were first equilibrated for an equivalent time of 30 void volumes.
3 Results and discussion In LC, investigating the role of the polar organic solvent is central to get a better understanding of the chemical interactions occurring on the surface of a separation material. To study the influence of the type of organic solvent, such as alcohols in hydrophilic interaction chromatography on native titania, two elution modes were considered: the hydroorganic mode as in classic HILIC (organic solvent/H2 O), and the polar organic mode for nonaqueous elution conditions (MeCN/organic solvent). www.jss-journal.com
538
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
Figure 1. Effect of water content on the retention of N-methyl xanthine derivatives ˚ 170 on (A) a silica column SiO2 (120 A, ˚ m2 /g) and (B) a titania column TiO2 (300 A, 30 m2 /g). Mobile phase, ammonium acetate (5 × 10−3 mol/L) in MeCN/H2 O (100−x):x, v/v; column temperature, 30⬚C; detection, UV 272 nm; solutes: 1MX, 3MX, 1,7DMX (300 mg/L) in MeCN/H2 O 66:34, v/v; injection volume, 5 L; column hold-up volume marker, naphthalene.
3.1 Hydro-organic elution mode—classic HILIC 3.1.1 MeCN/H2 O, effect of H2 O content in MeCN Since water is by reference a standard component in HILIC mobile phases [2–4], the retention profiles of 1,7DMX, 3MX, and 1MX were initially investigated as a function of the water content in the presence of ammonium acetate (5 ×10−3 mol/L) as salt additive on native silica (120 Å, 170 m2 /g), and titania (300 Å, 30 m2 /g). A typical chromatogram of xanthines separation on titania is shown in the Supporting Information. According to Fig. 1A and B, the retention of the set of xanthines on both silica and titania columns typically decreased with water amounts ranging from 2 to 40% v/v. The regularity of the elution patterns of the xanthines was further analyzed according to the well-established models of partition log10 k = f(%H2 O) and adsorption log10 k = f(log10 (% H2 O)) [11]. Assuming each of these two mechanisms on silica and titania, the calculated values of the coefficients of determination (Supporting Information Table S1) were closest to 1 for the adsorption model on both materials. On native silica, the adsorption of neutral xanthines would typically engage hydrogen bonding and dipolar interactions with the silanols groups (Si–OH) and siloxane bridges (Si–O–Si). On native titania, a multimodal retention regime is suspected since a greater retention was found on this column (Fig. 1A and B) despite the lower value of the specific surface area of the material (i.e. 30 m2 /g for titania compared to 170 m2 /g for the silica column). Unlike silica, the titania column was able to discriminate the two mono-N-methyl xanthine isomers, 3MX and 1MX, despite their similar molecular polarity. The position of the methyl group seems to be fundamental in the molecular interaction of the xanthine derivatives with the surface of titania, and responsible for such high selectivity. In addition to nonspecific adsorption on Ti–OH and Ti–O–Ti groups, the retention of the tested xanthine derivatives on titania (300 Å) is apparently governed by ligand exchange. Such a mechanism would occur between the water molecules occupying the Ti4+ Lewis acid sites and the free nitrogen atoms (i.e. N1, N3, or N7) of the xanthines, as we previously observed for titania (100 Å) [11].
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.1.2 X/H2 O, effect of the nature of the organic modifier Four different organic modifiers, namely MeOH, EtOH, i-PrOH, and MeCN were used as weak solvents added to water for hydrophilic chromatography of xanthines on silica and titania columns. Figure 2A and B depict the evolution of the retention factor of 1MX, 3MX, and 1,7DMX versus the type of organic modifier for both silica and titania columns. With 2% water, alcohols on silica were found to be stronger eluting components than acetonitrile. The eluotropic series of the alcohols typically followed their adsorption energy (g o ) on silica: H2 O MeOH (g o = 0.73) > EtOH (g o = 0.68) > i-PrOH (g o = 0.60) > MeCN (g o = 0.50) [24]. The elution strength order of the alcohols highlights the main implication of their protic character as competitor to the hydrogen bonding adsorption of xanthines on silica. For hydro-alcoholic eluents, the retention of 1,7DMX is always higher than the retention of the two mono-N-methyl xanthines (1MX and 3MX); whereas this order is reversed with MeCN. For all organic modifiers, the selectivity between the two monomethyl isomers remains very low on silica. With 10% H2 O, typical solvent effects were observed on titania compared to silica (Fig. 2B). The elution strength order of the different solvents on titania decreased in the sequence: i-PrOH > EtOH MeCN > MeOH. The eluotropic series of the different alcohols is completely modified on titania and methanol mostly acted as weaker eluent than acetonitrile. This means that the adsorption forms and pathways of alcohols on titania are totally different from that observed on silica. Several infrared/thermodesorption studies have previously demonstrated that reversible surface hydroxylation/dehydroxylation may occur upon adsorption of alcohols on titanium dioxide (anatase) at room temperature [25–28]. This sorption pathway is established through coordination bonding with Ti4+ Lewis acid sites, and involves reversible deprotonation/protonation of the adsorbed alcohols, which typically result in a surface populated by chemisorbed alkoxy groups (RO) and intact alcohol molecules (ROH) [25, 27–29]. In contrast to acetonitrile, alcohols possess hydroxyl groups that can act as competitors to the coordination process of
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
539
Figure 2. Effect of the nature of the organic modifier on the retention of N-methyl xanthine derivatives on (A) silica column ˚ 170 m2 /g), ammonium acetate SiO2 (120 A, (5 × 10−3 mol/L) in organic modifier/H2 O 98:2, v/v and (B) titania column TiO2 ˚ 30 m2 /g), ammonium acetate (300 A, (5 × 10−3 mol/L) in organic modifier/H2 O 90:10, v/v. Other experimental conditions are identical to Fig. 1, N = 3 repetitions, error bars: ±1 SD.
xanthines on titania. Based on this assumption, the ligand exchange between analyte (L), alcohol (ROH), and water on titania can be described using the following equilibria: Ti(OH)(H2 O) + ROH ↔ Ti(OH)(ROH) + H2 O
(1)
Ti(OH)(H2 O) + ROH ↔ Ti(OR)(H2 O) + H2 O
(2)
Ti(OH)(H2 O) + L ↔ Ti(OH)(L) + H2 O
(3)
Ti(OR)(H2 O) + L ↔ Ti(OR)(L) + H2 O
(4)
Ti(OH)(ROH) + L ↔ Ti(OH)(L) + ROH
(5)
Ti(OR)(H2 O) + L ↔ Ti(L)(OH) + ROH
(6)
Where Equations (1) and (2) refer respectively to associative and dissociative adsorption of alcohol on titania. Equations (3) and (4) account for exchange of coordinated water with a ligand analyte; whereas Equations (5) and (6) describe, respectively, the displacement of undissociated alcohol and alkoxide group from coordination by a ligand analyte. Considering the coordination hypothesis of alcohols on titania, the stability of titanium alcoholic complexes should increase with the length of the alkyl chain (R) as follows: MeOH < EtOH < i-PrOH. This predicated order is consistent with the eluotropic series obtained in hydro-alcoholic conditions on titania (Fig. 2B). Since methanol is the weakest eluent among the other alcohols, we assume on account of Equations (1), (2), (5) and (6), that this solvent is more easily displaced by water and analyte molecules during ligand exchange.
3.2 Polar organic elution mode—NA-HILIC 3.2.1 MeCN/MeOH, effect of MeOH content in MeCN To confirm whether methanol or acetonitrile is the weakest eluent on titania, the retention of xanthine derivatives was C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
investigated as a function of the MeOH content in MeCN in NA-HILIC. Figure 3A and B compares the results obtained for caffeine (1,3,7TMX), theophylline (1,3DMX), theobromine (3,7DMX), and 3-methylxanthine (3MX) in NA-HILIC on TiO2 (100 Å) with the data previously obtained in standard HILIC using water as the protic modifier in MeCN [11]. As a general observation, the retention factors of all the xanthines are greater in MeCN/MeOH (k > 20; for 1,7DMX and 1MX), despite a higher concentration of acetate competitor in the NA-HILIC mobile phase. Thus substituting water with methanol reduces the overall polarity and elution strength of the mobile phase. As we have already observed for MeCN/H2 O, two opposite regimes of retention were also observed for MeCN/MeOH when the amount of MeOH varied from 5 to 99% v/v in MeCN (Fig. 3A). Methanol acts as the strongest solvent for MeOH contents lower than 70% in MeCN, and above this value, the elution strengths of solvents were inverted, leading to a progressive increase of the retention factor of 3MX with MeOH content. The inversion of the elution strengths of solvents in NA-HILIC at high concentrations in MeCN still remains unclear on titania as it is for silica. For methanol contents lower than 50% in MeCN/MeOH, the retention data of the xanthine derivatives were plotted according to the models of partition log10 k = f(%MeOH) and adsorption log10 k = f(log10 (%MeOH)) (Fig. 4) to gain insight on the retention mechanisms involved in this range of composition. According to the correlation coefficients values (Supporting Information Table S2), the linearity of the different elution patterns fitted the adsorption mechanism for both polar-organic and hydro-organic modes. 3.2.2 MeCN/X, effect of the nature of the polar modifier The evolution of the retention factor of 3MX on titania (100 Å) was evaluated according to the type of alcohol (Fig. 5) for MeCN/X mobile phases in order to establish a scale of elution strength for methanol, ethanol, ethylene glycol, and water in acetonitrile. The eluotropic series at 10% v/v in MeCN follows the sequence: (CH2 OH)2 H2 O > EtOH > MeOH. The elution strength order between ethanol and methanol in acetonitrile is identical to that observed in water using hydroalcoholic mixtures (Section 3.1.2). It is interesting to notice www.jss-journal.com
540
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
Figure 3. Retention factors of Nmethyl xanthine derivatives on titania R ˚ Sachtopore -NP (100 A) as a function of the protic modifier content. (A) MeCN/MeOH [(99–x ):x , v/v] + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L) and (B) MeCN/H2 O [(100– x):x, v/v], ammonium acetate (1 × 10−3 mol/L). Other experimental conditions are identical to Fig. 1, N = 3 repetitions, error bars: ±1 SD.
Figure 4. Elution patterns of N-methyl xanthine derivatives on titania for MeOH contents less than 50% in MeCN/MeOH. (A) Partition model, log10 k = f(%MeOH) and (B) adsorption model, log10 k = f(log10 (%MeOH). Experimental conditions are listed in Fig. 3A.
hydrogen-bonded molecular networks and macro-clusters on the surface of titanium dioxide. 3.2.3 MeCN/X, ethylene glycol versus water
Figure 5. Retention factors of 3MX on titania using methanol, ethanol, ethylene glycol, and water as protic modifiers in MeCN. R ˚ mobile Chromatographic column, TiO2 Sachtopore -NP (100 A); phase, MeCN/protic modifier (90:10, v/v) + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L). Other experimental conditions are similar to Fig.1., N = 3 repetitions; error bars: ±1 SD.
that ethylene glycol is considerably stronger than water on titania, since the retention factor of 3MX was divided by a factor of 10 upon use of this alcohol as protic modifier in NA-HILIC. Bicker et al. [9] earlier reported a similar but less significant behavior of ethylene glycol on bare silica. As a dihydric solvent (i.e. diol), ethylene glycol has a large dipole moment (2.36 D versus 1.85 D for water) [30] and two OH groups that are responsible for the substantial adsorption of this alcohol on oxide surfaces. Sekuli´c et al. [31] reported that free and adsorbed ethylene glycol molecules can even establish C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The retention of 1,7DMX, 3MX, and 1MX on titania (100 Å) was evaluated as a function of the amount of ethylene glycol to further characterize the effects of the diol as strong protic modifier in NA-HILIC. In comparison with water, the retention data obtained for less than 50% v/v ethylene glycol contents in acetonitrile were plotted in the logarithmic scale, assuming the adsorption model. According to Fig. 6A and B, the retention factors of the xanthine derivatives decrease with increasing amounts of ethylene glycol or water in the mobile phase. The elution patterns typically followed the adsorption model for both solvents. The substitution of water with ethylene glycol led to lower retentions on titania and the selectivity between the xanthines remained important in all of these polar organic conditions. The low retention of the analytes in the presence of ethylene glycol highlights substantial bonding between the diol and the surface of the metal oxide. The desorption of ethylene glycol from the surface of titania appeared as a slow process and a special washing procedure had to be implemented in order to recover the native titania surface. The retention factor of 1,7DMX using ammonium acetate 5 × 10−3 mol/L in MeOH/H2 O (90:10 v/v) as mobile phase was originally close to 25, but after using ethylene glycol as mobile phase, the retention factor of 1,7DMX with the same ammonium acetate mobile phase was three www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 536–542
541
Figure 6. Elution profiles of N-methyl xanthines on titania for contents less than 50% of ethylene glycol (A) and water (B). ChroR matographic column, titania Sachtopore ˚ mobile phase, MeCN/protic NP (100 A); modifier ((99–x):x, v/v) + 1% v/v ammonium acetate (total concentration, 5 × 10−3 mol/L). Other experimental conditions are identical to Fig. 1. Plots are represented in decimal logarithmic scale, assuming the adsorption model: R2 (EG) > 0.95; R2 (water) ∼ 0.99.
times lower than previously observed. A 2 h column wash with (50:50, v/v) MeOH/NH3 ,H2 O (0.3% v) at 1 mL/min was not efficient enough and only a 12 h treatment at 60⬚C with (50:50, v/v) MeOH/H2 O at 1 mL/min was able to restore the retention of 1,7DMX. According to Sekuli´c et al. [31], the presence of high amounts of water in contact with titania may disrupt the physisorbed ethylene glycol network and macroclusters through hydrogen bonding, so free water molecules will have access to coordination sites and will progressively replace the chemisorbed ethylene glycol. We assume that a high washing temperature can improve the diffusion of free water molecules into the pores of titania, whereas the presence of a hard Lewis base such as ammonia in the washing solution can accelerate the exchange of ethylene glycol for hydroxide anions. As great ligand competitor on titania, biocompatible ethylene glycol, used to replace water in acetonitrile-based mobile phases, may potentially offer novel selectivity for extended applications of titania in sample treatment of complex biological mixtures.
classic HILIC and led also to interesting selectivities between the xanthines. The authors have declared no conflict of interest.
5 References [1] Alpert, A. J., J. Chromatogr. 1990, 499, 177–196. ¨ [2] Hemstrom, P., Irgum, K., J. Sep. Sci. 2006, 29, 1784–1821. [3] Jandera, P., J. Sep. Sci. 2008, 31, 1421–1437. [4] Gama, M. R., Silva, R. G. C., Collins, C. H., Bottoli, C. B. G., TrAC Trends Anal. Chem. 2012, 37, 48–60. [5] Kind, T., Tolstikov, V., Fiehn, O., Weiss, R. H., Anal. Biochem. 2007, 363, 185–195. [6] Garbis, S. D., Melse-Boonstra, A., West, C. E., Van Breemen, R. B., Anal. Chem. 2001, 73, 5358–5364. [7] Buszewski, B., Noga, S., Anal. Bioanal. Chem. 2012, 402, 231–247. [8] Pesek, J. J., Matyska, M. T., Boysen, R. I., Yang, Y., Hearn, M. T. W., TrAC Trends Anal. Chem. 2013, 42, 64–73.
4 Conclusions The type and amount of protic modifier is an interesting mobile phase parameter that can be carefully exploited to tune the chromatographic retention and selectivity of polar compounds in hydrophilic interaction LC. The evaluation of the eluotropic series of alcohols as protic modifiers in both HILIC and NA-HILIC elution modes provided, in combination with intercolumn comparison, valuable comprehension of the retention mechanisms of N-methyl xanthines on silica and titania. The retention of neutral xanthine at low contents of protic modifiers exhibited simple molecular physisorption on silica and a mixed-mode mechanism composed of adsorption and ligand exchange on titania. Adsorptive interactions of analytes on silica are strengthened when the polarity of the protic solvent is reduced. However, the elution strength of protic modifiers on titania is highly influenced by their Lewis base properties, since these solvents are directly involved in a ligand-exchange competition process with the analytes. Particularly, the use of ethylene glycol as organic protic modifier in NA-HILIC provided faster analysis time than water did in C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[9] Bicker, W., Wu, J. Y., Lammerhofer, M., Lindner, W., J. Sep. Sci. 2008, 31, 2971–2987. [10] Randon, J., Guerrin, J.-F., Rocca, J.-L., J. Chromatogr. A. 2008, 1214, 183–186. ´ M., Randon, J., J. Chromatogr. A 2011, 1218, [11] Abi Jaoude, 721–725. ´ M., Randon, J., Anal. Bioanal. Chem. 2011, [12] Abi Jaoude, 400, 1241–1249. ´ M., Morin, N., Miege, ` [13] El Debs, R., Abi Jaoude, C., Randon, J., J. Sep. Sci. 2011, 34, 1805–1810. [14] Winkler, J., Marme, S., J. Chromatogr. A 2000, 888, 51–62. ¨ M., Kurganov, A. A., Schacht, S., Schuth, ¨ [15] Grun, F., Unger, K. K., J. Chromatogr. A 1996, 740, 1–9. [16] Zhou, T., Lucy, C. A., J. Chromatogr. A 2008, 1187, 87–93. [17] Zhou, T., Lucy, C. A., J. Chromatogr. A 2010, 1217, 82–88. [18] Ikeguchi, Y., Nakamura, H., Anal. Sci. 1997, 13, 479–483. [19] Kuroda, I., Shintani, Y., Motokawa, M., Abe, S., Furuno, M. Anal. Sci. 2004, 20, 1313–1319. [20] Pinkse, M. W. H., Uitto, P. M., Hilhorst, M. J., Ooms, B., Heck, A. J. R., Anal. Chem. 2004, 76, 3935–3943.
www.jss-journal.com
542
M. A. Jaoude´ et al.
J. Sep. Sci. 2014, 37, 536–542
[21] Sekiguchi, Y., Mitsuhashi, N., Inoue, Y., Yagisawa, H., Mimura, T., J. Chromatogr. A 2004, 1039, 71–76.
[26] Kim, K. S., Barteau, M. A., Farneth, W. E., Langmuir 1988, 4, 533–543.
[22] Hata, K., Morisaka, H., Hara, K., Mima, J., Yumoto, N., Tatsu, Y., Furuno, M., Ishizuka, N., Ueda, M., Anal. Biochem. 2006, 350, 292–297.
[27] Shepot’ko, M. L., Davydov, A. A., Theor. Exp. Chem. 1991, 27, 210–214.
[23] Richardson, B. M., Soderblom, E. J., Thompson, J. W., Moseley, M. A., J. Biomol. Tech. 2013, 24, 8–16. [24] Meyer, V. R., Practical High-Performance Liquid Chromatography, 3rd Edition, John Wiley & Sons, Hoboken, 1998. [25] Ramis, G., Busca, G., Lorenzelli, V., J. Chem. Soc. Faraday Trans. I 1987, 83, 1591–1599.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[28] Wang, C., Groenzin, H., Shultz, M. J., J. Am. Chem. Soc. 2005, 127, 9736–9745. [29] Lusvardi, V. S., Barteau, M. A., Farneth, W. E., J. Catal. 1995, 153, 41–53. [30] Lide, D. R., CRC Handbook of Chemistry and Physics, 82th Edition, CRC Press, Boca Raton 2000. ´ J., Elshof, J. E. T., Blank, D. H. A., Langmuir 2005, [31] Sekulic, 21, 508–510.
www.jss-journal.com
