Instrumental Analysis Chemistry

Determination of Suitable Column Geometries by Means of van Deemter and Kinetic Plots for Isothermal and Isocratic Method Development in High-Temperature Liquid Chromatography Isotope Ratio Mass Spectrometry Paul Ermisch,† Steffen Wiese,† Harald Weber,‡ and Thorsten Teutenberg*,†

†Institut fuer Energie- und Umwelttechnik e. V., Bliersheimer Strasse 58-60, 47229 Duisburg, Germany ‡Niederrhein University of Applied Science, 47798 Krefeld, Germany

*S Supporting Information

ABSTRACT: The method of high-temperature liquid chro- matography isotope ratio mass spectrometry (HTLC-IRMS) is used to determine the origin or authenticity of compounds. Currently, the drawback of this hyphenation is the interface which causes pronounced band broadening due to a large extra-column volume. Therefore, the aim of this study is to determine suitable column geometries and particle sizes at different temperature and to study the effect of extra-column band broadening. The tools to assess the efficiency of columns are van Deemter and kinetic plots. By comparison of different column geometries and particle sizes, it could be shown that 3.0 mm ID columns achieve a higher performance than 2.1 mm ID columns and a particle size of 1.7 μm is advantageous over 3.5 and 5.0 μm particles when the injection volume is adjusted to 2 μL and the temperature is higher than 60 °C. Because water was the mobile phase, the retention factor could not be kept constant at different column temperatures. The lower retention factor at elevated temperatures leads to a decrease of the plate number, because of the relatively larger contribution to extra-column band broadening at lower retention factors. This is the reason why 3.0 mm ID columns should be preferred for the HTLC-IRMS hyphenation when the separation is carried out under isothermal and isocratic conditions.

The use of elevated temperatures in liquid chromatographyhas some advantages compared to liquid chromatography at ambient conditions.1−6 By increasing the temperature, the viscosity of the mobile phase and hence the pressure drop over the column decreases.7 Therefore, a longer column can be used at the same flow rate or shorter analysis times can be achieved using higher flow rates without sacrificing efficiency. Furthermore, the diffusion coefficient DM, described by the Wilke-Chang equation, increases with increasing temperature.8

= * Ψ

η −D

M

V T7.4 10M

8 2 2

1 0.6

(1)

Here, Ψ2 is the association factor of the solvent (1 for nonpolar solvents and 2.6 for water), M2 [g mol

−1] is the molecular weight of the solvent, T [K] is the absolute temperature, V1 [mL] is the molar volume of the solute, and η [cP] is the viscosity of the solvent which strongly depends on temperature. Another effect is that the static permittivity and hence polarity of the mobile phase will be reduced when the temperature is increased.9 At high temperatures, water has a similar polarity like an organic solvent at ambient temperature. Consequently, a solvent gradient can be replaced by a temperature gra-

dient.10−12 This allows the use of special hyphenation techniques, where only water can be used as mobile phase.13−18

One of these techniques is isotope ratio mass spectrometry (IRMS), which is used to determine the origin or authenticity of samples.14,15,19,20 The variation in stable isotope ratio of different samples is due to fractionation effects during chemical or biological reactions. Before determining the isotope ratio of components, complex samples must be separated chromato- graphically. For this reason, the isotope ratio mass spectrometer is coupled to an HPLC system. In the interface, all carbon which is in the mobile phase is oxidized to CO2. Afterward, the CO2 is separated from the mobile phase for the isotope analysis. To prevent a high blank value, only water can be used as eluent in the HPLC unit. The drawback of this hyphenation is the interface which

causes pronounced band broadening due to a long residence time of the analytes in the oxidation reactor. The flow rate of the mobile phase is adapted to the oxidation time and cannot be increased further. Currently, a maximal flow rate of 0.7 mL min−1 can be adjusted on commercially available instrumenta-

Received: October 25, 2011 Accepted: January 4, 2012 Published: January 20, 2012

Article

pubs.acs.org/ac

© 2012 American Chemical Society 1565 dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−1571

tion. However, it would be advantageous to increase the flow rate in order to prevent work in the B-term dominated region of the van Deemter curve at elevated temperatures. A loss in separation efficiency would also negatively effect the resolution between adjacent peaks. One of the basic requirements for IRMS analysis is that a baseline separation of all compounds in the mixture has to be achieved. Therefore, the best possible separation in the HPLC unit is necessary to prevent a coelution of the CO2 fractions at increased flow rates. The aim of this study was to determine suitable column geometries and particle sizes for a temperature range of 60 to 180 °C using water as the mobile phase and to study the effect of extra-column band broadening on the overall efficiency.

■ THEORY When assessing the efficiency of columns, van Deemter plots can be used. They show the lowest achievable plate height Hmin and the corresponding optimal linear velocity uopt. Van Deemter plots are based on the van Deemter equation. Here, H [μm] is the plate height, A [μm] is the Eddy diffusion, B [μm2 s−1] is the longitudinal diffusion, C [μm s−1] is the mass transfer, and u [mm s−1] the linear mobile phase velocity.

= + +H u A B u

Cu( ) (2)

Another tool to compare the efficiency of columns is the kinetic plot method. Kinetic plots are derived from experimental van Deemter plots and show the highest plate number a column can achieve in the shortest time when working at the maximum pressure of the system.21 A transformation of experimentally generated van Deemter plots to kinetic plots can be performed with some simple equations. For the basic plate number (N) versus dead time (t0) kinetic plot, the following equations are applied

= Δ

η ⎜ ⎟ ⎛ ⎝

⎞ ⎠N

p K uH

max v

exp (3)

= Δ

η ⎛ ⎝⎜

⎞ ⎠⎟t

p K u

0 max v

2 exp (4)

Here, for every experimentally determined (u, H)-couple, the column is prolonged to the length which yields a pressure drop equal to the system pressure limit Δpmax at the given linear velocity u. Hence, every data point of a kinetic plot curve is achieved with another column length at different dead times. The system pressure limit can be freely chosen. For a theoretical consideration, the pressure limit can be set to values beyond the real system limit. Thus, one obtains information about the performance a column can achieve on a UHPLC system even though the van Deemter data are generated on a conventional HPLC system. However, extrapolating HPLC data to UHPLC conditions can over- estimate the performance of the system at ultrahigh pressures because of effects like viscous heating.

■ EXPERIMENTAL SECTION Chemicals and Columns. The model compounds

sulfadiazine, sulfamerazine, and sulfamethazine, with a purity grade of >99%, were purchased from Sigma-Aldrich (Seelze, Germany) and dissolved in water/acetonitrile (75/25, v/v) at a concentration of 100 μg mL−1. High-purity deionized water was

produced by an Elix 10-Milli-Q Plus water purification system (Millipore, Eschborn, Germany) and acidified with 0.1% formic acid, obtained from Fluka (Seelze, Germany). Acetonitrile (Optigrade) was purchased from LGC Standards (Wesel, Germany). The columns, which were used in this study and filled with ethylene-bridged hybrid (BEH) C18 particles, were provided by Waters (Eschborn, Germany). Columns containing different particle sizes from 1.7, 3.5, and 5.0 μm and with internal diameters of 2.1 and 3.0 mm ID were used in this study. The column length was always 5 cm in order to obtain a better signal-to-noise ratio on the 2.1 mm i.d. column, because HPLC-IRMS is not a trace analytical technique. For monitoring the degradation of the stationary phases, a mixture of uracil, naphthalene, and acenaphthene, purchased from Fluka, was measured with a water/acetonitrile (50/50, v/v) mobile phase after finishing measurement series at 60, 120, and 180 °C, respectively.

Instrumentation. The experiments were always performed in triplicate on two HPLC systems at temperatures of 60, 120, and 180 °C. Columns with 3.5 and 5.0 μm particles were used on an Agilent 1100 Series HPLC system (Agilent Technologies, Waldbronn, Germany) in combination with a commercially available SIM HT-HPLC 200 high temperature column oven (SIM, Scientific Instruments Manufacturer, Oberhausen, Germany). For the column packed with 1.7 μm particles, an UltiMate 3000 UHPLC System (Dionex GmbH Deutschland, Idstein, Germany) in combination with a Metalox Model 200-C high-temperature column oven (Systec, New Grighton, USA) was used. The SIM HT-HPLC oven included an eluent preheater and an eluent cooler, which decreased the mobile phase temperature to 50 °C before it was introduced into the detector. The investigated flow rate range was between 0.05 and 1.5 mL min−1. The injection volume was always 2 μL which was regarded to be the lower injection limit. A lower injection volume would not make sense for HPLC-IRMS hyphenation due to the relatively high amount of carbon which has to be injected on column. Detection was carried out at a wavelength of 270 nm. To simulate the above-described HPLC-IRMS-Interface, a

PEEK capillary with an internal diameter of 0.17 mm and a length of 1 m was connected between the column and the detector and heated at 80 °C. The volume of this additional capillary was about 23 μL. All measurements were carried out first without and subsequently with additional capillary.

Methodology. Data fitting to the van Deemter curves was performed by minimizing the sum of error squares. From the experimentally determined van Deemter data, kinetic plots were transformed using the Kinetic Plot Analyzer software (version 6.7), provided by the Vrije Universiteit Brussel.22 The permeability K [mm2 s−1 bar−1], specific permeability Kv [mm

2], and porosity ε needed to create kinetic plots were calculated as specified in the literature.23 The mobile phase dynamic viscosity η [mPa s] was calculated according to ref 24. Lc [mm] is the column length, Δpcol [bar] is the pressure drop across the column, t0 [s] is the dead time, F [mL min

−1] is the flow rate, and dc [mm] is the internal column diameter. The dead times were calculated with the simulation software DryLab 2000 Plus, Ver. 3.8.

= ηε−K K10v 8

(5)

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711566

= *

Δ =

Δ * K

u L p

L p t

c

col

c 2

col 0 (6)

ε = * *

F t

d L 21.2 0

c 2

c (7)

The column pressure drop (Δpcol) was determined by subtracting the system pressure drop without column (Δpcap) from the entire system pressure drop including the column (Δpsys) at every flow rate:

Δ = Δ − Δp p pcol sys cap (8)

To investigate the influence of the temperature on the separation efficiency, the retention factor should be kept constant. This condition can be fulfilled by adapting the composition of the mobile phase to the temperature.25,26 In this work, however, only water was used as mobile phase and hence the retention factor could not be kept constant.

■ RESULTS AND DISCUSSION All experiments were performed on silica ethylene-bridged hybrid (BEH) stationary phases because of their excellent stability at high temperatures.27 Figure 1 shows the influence of an additional extra-column volume on the separation power of two columns with an ID of 3.0 and 2.1 mm at 60 and 120 °C. When the 3.0 mm ID column is used at 60 °C which is depicted in Figure 1(a), an additional extra-column volume has almost no influence on the plate height minimum. This can also be observed when the temperature is increased to 120 °C as is shown in Figure 1(b). There is just a slight increase of the plate height at higher flow rates. In contrast, the 2.1 mm ID column is subjected to a decreased efficiency at temperatures of 120 °C and higher when an extra-column volume is added (Figure 1(d)). Here, the minimal plate height is increased from 28 to 31 μm. At 60 °C, no loss in efficiency is observed (Figure 1(c)). At a temperature of 180 °C, however, a loss of the separation efficiency for the 3.0 mm ID column is also observed. The minimal plate height of sulfamerazine increases from 43 to 47 μm when an additional extra-column volume is added (data not shown). For the 2.1 mm ID column, the data could not be

Figure 1. van Deemter plots of sulfamerazine on a 3.0 mm ID (a, b) and 2.1 mm ID (c, d) column with 5 μm particles at 60 °C (a, c) and 120 °C (b, d). The black solid curves ( × ) were obtained without and the red dashed curves (○) were obtained with additional extra-column volume.

Table 1. Minimum Plate Height [μm] for Sulfamerazine at Optimal Linear Velocity uopt and at Flow Rates of 0.7 and 1.5 mL min−1 Including the Influence of an Additional Extra-Column Volume (aecv)

2.1 mm ID 3.0 mm ID

T, °C uopt 0.7 mL min −1 1.5 mL min−1 uopt 0.7 mL min

−1 1.5 mL min−1

60 without aecv 15 20 33 13 13 17 with aecv 15 19 33 14 14 18

120 without aecv 28 51 96 18 20 31 with aecv 31 56 103 18 20 33

180 without aecv 43 83 162 with aecv 47 81 148

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711567

fitted to a valid van Deemter curve when the temperature was increased to 180 °C, because the data points showed no minimum and a large scatter. In Table 1, the plate heights for the above-discussed curves at

the optimal linear velocity as well as flow rates of 0.7 and 1.5 mL min−1 are summarized. These flow rates have been chosen because 0.7 mL min−1 is the highest flow rate which can be adjusted on commercially available instrumentation for HPLC- IRMS hyphenation. In order to increase the speed of analysis, an increase of the flow rate by a factor of 2 is desirable. When intending to increase the flow rate from 0.7 to 1.5 mL min−1, the loss of efficiency has to be as low as possible. From Table 1, it can be derived that a two times higher flow rate leads to an increase in the plate height. However, when a 3.0 mm ID column is used, the loss of efficiency is lower than with a 2.1 mm ID column. Especially when increasing the temperature from 60 to 180 °C, the plate height increases from 13 to 43 μm for the 3.0 mm ID column. Interestingly, the plate height minimum without the extra column capillary is slightly higher than with additional extra column volume. We assume that this is because of the extremely low retention factors (k < 1) at these temperatures so that the error in calculating H is rather high. As already mentioned above, an increase of the temperature

in theory leads to a shift of the van Deemter minimum to higher linear velocities, but the plate height minimum remains constant.25 This is the result of a higher B-term- and lower C- term-contribution at elevated temperatures. However, this behavior can only be observed when the retention factor is kept constant at all temperatures by adjusting the mobile phase composition. Figure 2(a) shows the influence of temperature

when the retention factor is not constant due to the use of water as sole mobile phase. By increasing the temperature, the curve minimum increases from 15 μm at 60 °C to 82 μm at 180 °C. Furthermore, the slope of the C-term increases, and the curve fit is worse than at lower temperatures. This reduced separation power is caused by a lower retention factor at higher temperatures, because of the relatively larger contribution to extra-column band broadening at lower retention factors. The model compound in Figure 2(a), sulfadiazine, has a retention factor of 10 at 60 °C, 2.3 at 120 °C, and 0.7 at 180 °C. Neue already showed that the plate number will decrease when the retention factor decreases.28 Here, the loss of efficiency is dependent on the internal diameter of the column. The lower the internal column diameter, the higher is the efficiency loss at a given retention factor. This is also underlined by the comparison of the separation

power of the two column IDs (2.1 and 3.0 mm) at different temperatures (60 and 120 °C) in Figure 2(c,d). At 60 °C, both columns generate a minimal plate height of 15 μm (Figure 2(c)). When increasing the temperature to 120 °C (Figure 2(d)), the 3.0 mm ID column still achieves the same efficiency at the optimal linear velocity in the plate height minimum. In contrast, the separation power of the 2.1 mm ID column decreases, and the plate height minimum has a value of 22 μm. When the temperature is increased further to 180 °C, the performance of both columns is significantly reduced. The 3.0 mm ID column has a doubled minimal plate height (15 to 28 μm). Because of an invalid van Deemter curve at 180 °C, no minimal plate height for sulfadiazine on the 2.1 mm ID column could be calculated. The curve could not be fitted because it showed no minimum and a large scatter of data points. These

Figure 2. (a) van Deemter plots of sulfadiazine at 60 °C (black solid curve, × ), 120 °C (red dashed curve, ○), and 180 °C (blue dotted curve, □), obtained with a 3.0 mm ID column with 3.5 μm particles. (b) van Deemter plots of sulfamethazine on columns with an ID of 2.1 mm and 1.7 μm (black solid curve, × ), 3.5 μm (red dashed curve, ○), and 5.0 μm (blue dotted curve, □) at a temperature of 120 °C. (c and d) van Deemter plots of sulfamethazine on columns containing 5 μm particles with an ID of 2.1 mm (black solid curves, ×) and 3.0 mm (red dashed curves, ○) at a temperature of 60 and 120 °C.

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711568

results are related to the above-mentioned fact of a much lower retention factor at higher temperature when the solvent strength of the mobile phase cannot be adjusted. A small internal column diameter and a low retention factor leads to very short interaction times of the analytes with the stationary phase. In this case, the influence of the column volume on efficiency is extremely high. To determine the influence of the particle size on the column

efficiency, particles with dimensions of 1.7, 3.5, and 5.0 μm were investigated. The comparison at 120 °C in Figure 2(b) shows that the 1.7 μm particles achieve the lowest plate height. In addition, the slope of the C-term at higher linear velocities decreases with smaller particles. In consequence, the minimum is shifted to higher linear velocities. These results are in complete agreement with theoretical considerations and could be shown in other studies.29,30

The following results refer to kinetic plots, transformed from experimentally obtained van Deemter plots. Figure 3(a,b) shows kinetic plots of plate number N versus dead time t0 to compare the three different particle sizes. The plots differ in the chosen working pressure of 400 (Figure 3(a)) and 1200 bar (Figure 3(b)), respectively. At low dead times, both plots show that small particles generate a higher plate number than larger particles. At high dead times, larger particles are superior over small particles. The intersection is dependent on the operating pressure. Working at 400 bar, 1.7 μm particles are advantageous up to a dead time of 24 s. In the dead time range of 24 to 107 s, 3.5 and 5.0 μm particles generate almost the same plate number but more than 1.7 μm particles. For dead times higher than 107 s, 5.0 μm particles are suited best. This means that for fast analysis times lower than 10 min small particles and for analysis times higher than 10 min large particles should be used. When the working pressure is increased to 1200 bar (Figure 3(b)), the intersection points shift to three times higher dead time values and the curves reach higher values. The advantage of small particles over larger particles increases proportional to the increase in the working pressure. Furthermore, for maximizing the absolute plate number, only large particles should be used. However, a very high number of plates is only achievable at excessively long retention times. For example, the marked data point (*) in Figure 3(a) refers to a plate number of approximately 130.000 at a dead time of 170 min. With a retention factor for sulfamerazine of 25 at 60 °C, a retention time of almost 74 h results! For usual chromatography a plate number of 10.000 is sufficient and can be generated much faster with 1.7 μm particles than with larger particles. In Figure 3(c), a kinetic plot of column length L versus plate

number N is shown. From this plot, it can be derived which column length is needed to reach a certain plate number. Short columns are filled ideally with small, long columns with large particles. Here, the performance of the three particle sizes to each other also changes. The intersection points are obtained at a column length of 24 and 69 cm, respectively. Because of the fact that conventional columns have lengths of 5, 10, and 15 cm, only the small particle packed columns should be used. Of course, longer columns can be obtained by coupling columns together, but this approach leads to excessively long retention times31 and possible analyte degradation when working at elevated temperatures. Figure 4 also shows N versus t0-kinetic plots as a function of

the maximal column pressure. It illustrates the separation efficiency of columns with internal diameters of 2.1 and 3.0 mm. As can be seen in Figure 4(a), at a temperature of 60 °C,

the efficiency of the 3.0 mm ID column is slightly higher than for the 2.1 mm ID column. This holds true for the whole dead time/retention time range. When increasing the temperature to 120 °C as is shown in Figure 4(b), the difference in efficiency becomes more pronounced. The 3.0 mm ID column achieves a much higher plate number compared to the 2.1 mm ID column. The vertical, dotted lines mark the retention times of 10, 30, and 60 min, respectively. For a given retention time, both columns generate a higher number of plates when the temperature is increased from 60 to 120 °C.

Figure 3. Kinetic plots of sulfamerazine showing a comparison of different particle sizes (1.7 (black ×), 3.5 (red ○), and 5.0 μm (blue □)) at 60 °C. (a and b) N versus t0-plot at 400 and 1200 bar, respectively. (c) Column length L versus N-plot at a pressure of 400 bar.

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711569

The dashed curves represent the efficiency which can be obtained when the columns are operated at a working pressure of 1200 bar. It is seen that the columns can generate a higher number of plates compared to a maximal pressure of 400 bar (solid lines). For a retention time of 60 min and a working pressure of 1200 bar, a 3.0 mm ID column is able to generate 100 000 plates. However, for this performance a column length of about 2.5 m is necessary! The stationary phase degradation was monitored with a

mixture of uracil, naphthalene, and acenaphthene. As expected, the columns showed a loss of retention time up to 12% after the experiments at elevated temperatures. The loss of retention is independent of the column dimension or particle size and occurs for all columns at the same degree (Figure S.1, Table S.1, Supporting Information). Finally, the effect of frictional heating has to be addressed.

The basis data for the calculated kinetic plots were almost exclusively determined at pressures below 400 bar. When calculating kinetic plots for any higher pressures, e.g., 1200 bar, frictional heating was not taken into account. When frictional heating occurs, the peak may be subjected to a band broadening due to axial and radial temperature gradients.6 In this case, the real column efficiency will be much lower than the calculated efficiency. Therefore, it is useful to extrapolate kinetic plots just for operation pressures where frictional heating can be excluded. By this means, a realistic assessment of the possible column efficiency can be made. In order to derive the conclusions, the obtained results now

have to be discussed in the context of the hyphenation of high- temperature HPLC with isotope ratio mass spectrometry. As was already explained in the introduction, this hyphenation technique completely relies on the requirement that only a water mobile phase can be used. This means that any addition of an organic solvent in order to facilitate the elution of compounds is prohibited. The conventional solvent gradient elution cannot be applied. Instead, temperature is the most important parameter to influence retention, selectivity, and efficiency. When the temperature is increased, the retention factor will decrease for most compounds in reversed phase HPLC. This in turn leads to a much higher influence of the extra-column volume to band broadening and hence efficiency. Operating the column under isothermal and isocratic conditions means that the analytes cannot be focused at the column head, especially when the temperature is increased. In order to obtain high efficiencies at isothermal conditions, a low

temperature would always be preferred because of a much higher retention. In this respect, the influence of the column diameter will only exert a negligible influence on the column efficiency. Working at low temperature has the inherent disadvantage, however, that extremely long retention times will be observed. Applying higher temperatures will increase the sample

throughput because of a much lower retention of the target analytes. However, the extra-column volume exerts a strong influence on the efficiency. Therefore, a compromise has to be found between the column inner diameter and the temperature. In our experiments, the injection volume was kept constant. The reason is that IRMS is not an analytical technique for trace analysis. The injection volume or the injected sample amount has to be high enough to get a signal. Therefore, an injection volume below 2 μL will not be considered for HPLC-IRMS hyphenation. Applying higher injection volumes up to 20 μL would lead to an even higher discrepancy between the performance of the 2.1 and 3.0 mm ID column. Increasing the column length to 10 cm for the 2.1 mm ID column would not lead to a better peak focusing at elevated temperatures when the injection volume is kept constant. Furthermore, a higher signal-to-noise ratio can also not be expected because of a longer elution of the analyte bands through the column. From the discussion above, it becomes clear that temperature

programming instead of the isothermal operation mode would offer the benefit of a sample focusing on the head of the column as well as a higher retention of all compounds. In this respect, low and highly retained solutes could be eluted in a much narrower time window, which we have shown else- where.10 Currently, the maximal flow rate which can be adjusted in high-temperature liquid chromatography (HTLC)- IRMS is around 0.7 mL min−1, although it would be desirable to increase the flow rate in order to compensate the loss in efficiency when working in the B-term dominated region of the van Deemter curve. However, our findings also demonstrated that increasing the flow rate to 1.5 mL min−1 will cause a slightly negative influence on band broadening but a doubled sample throughput when working at 60 °C and using a column diameter of 3.0 mm.

■ CONCLUSION Our results clearly show that columns with an inner diameter of 3.0 mm should be preferred over columns with an inner diameter of 2.1 mm for HPLC-IRMS hyphenation at

Figure 4. N versus t0-kinetic plots of sulfamerazine showing a comparison of different column internal diameters (2.1 mm (black ●, green ○) and 3.0 mm (red ■, blue □)) at 60 °C (a) and 120 °C (b). Solid lines at 400 bar (full symbols) and dashed lines at 1200 bar (open symbols), respectively.

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711570

temperatures above 60 °C if the following restrictions concerning method development exist: (a) The mobile phase consists of water, and the retention factor cannot be adjusted at higher temperatures by changing the mobile phase composi- tion. (b) The injection volume cannot be decreased because of the detection limit given by the detector. In this study, the injection volume was always 2 μL which can be regarded as the lower limit of injection for HPLC-IRMS hyphenation. (c) The sample is eluted under isothermal conditions. These requirements are specific for the hyphenation of high-

temperature HPLC with isotope ratio mass spectrometry, but they can also be transferred to the LC-FID hyphenation which is well described in the literature. Although a higher inner column diameter will always be better, the particle size of the stationary phase should be as low as possible if the column is shorter than 24 cm. Columns containing 5 or 3.5 μm particles will always show a lower efficiency. A small extra-column volume of about 23 μL does not exert a strong influence on column efficiency for a 3.0 mm ID column if the temperature is below 120 °C. In order to overcome these inherent drawbacks, temperature programming should be used instead of isothermal elution, which will be addressed in a forthcoming publication.

■ ASSOCIATED CONTENT *S Supporting Information Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATION Corresponding Author *Phone: +49 2065 418 179. Fax: +49 2065 418 211. E-mail: [email protected].

■ ACKNOWLEDGMENTS The authors are thankful for the financial aid supported by the German Federal Ministry of Economics and Technology on the basis of a decision by the German Bundestag (Project number KF 2025405MK9). We would also like to thank Dionex for the loan of the UltiMate 3000 UHPLC system.

■ REFERENCES (1) Greibrokk, T.; Andersen, T. J. Chromatogr., A 2003, 1000, 743. (2) Vanhoenacker, G.; Sandra, P. J. Chromatogr., A 2005, 1082, 193. (3) Lestremau, F.; Cooper, A.; Szucs, R.; David, F.; Sandra, P. J. Chromatogr., A 2006, 1109, 191. (4) Yang, Y. J. Sep. Sci. 2007, 30, 1131. (5) Teutenberg, T. Anal. Chim. Acta 2009, 643, 1. (6) Teutenberg, T. High-Temperature Liquid Chromatography-A User’s Guide for Method Development; Royal Society of Chemistry: Cambridge, U.K., 2010; Vol. 13. (7) Teutenberg, T.; Wiese, S.; Wagner, P.; Gmehling, J. J. Chromatogr., A 2009, 1216, 8470. (8) Wilke, C. R.; Chang, P. AIChE J. 1955, 1, 264. (9) Teutenberg, T.; Wiese, S.; Wagner, P.; Gmehling, J. J. Chromatogr., A 2009, 1216, 8480. (10) Wiese, S.; Teutenberg, T.; Schmidt, T. C. Anal. Chem. 2011, 83, 2227. (11) Wiese, S.; Teutenberg, T.; Schmidt, T. J. Chromatogr., A 2011, 1218, 6898. (12) Wiese, S.; Teutenberg, T.; Schmidt, T. C. J. Chromatogr., A 2012, 1222, 71. (13) de Boer, A. R.; Alcaide-Hidalgo, J. M.; Krabbe, J. G.; Kolkman, J.; van Emde Boas, C. N.; Niessen, W. M.; Lingeman, H.; Irth, H. Anal. Chem. 2005, 77, 7894.

(14) Godin, J. P.; Fay, L. B.; Hopfgartner, G. Mass Spectrom. Rev. 2007, 26, 751. (15) Krummen, M.; Hilkert, A. W.; Juchelka, D.; Duhr, A.; Schluter, H. J.; Pesch, R. Rapid Commun. Mass Spectrom. 2004, 18, 2260. (16) Reichelt, K. V.; Peter, R.; Paetz, S.; Roloff, M.; Ley, J. P.; Krammer, G. E.; Engel, K. H. J. Agric. Food Chem. 2010, 58, 458. (17) Roloff, M.; Erfurt, H.; Kindel, G.; Schmidt, C.-O.; Krammer, G. Patent Int.-Publ.-No.: WO 2006/111476 A1, Organization, W. I. P., Ed., 2006. (18) Zhang, L. J.; Kujawinski, D. M.; Jochmann, M. A.; Schmidt, T. C. Rapid Commun. Mass Spectrom. 2011, 25, 2971. (19) Benson, S.; Lennard, C.; Maynard, P.; Roux, C. Forensic Sci. Int. 2006, 157, 1. (20) Jochmann, M. A.; Kujawinski, D. M.; Zhang, L.; Schmidt, T. C.; Teutenberg, T.; Wiese, S.; Fischer, B.; Bettermann, H. GIT 2010, 54, 182. (21) Desmet, G.; Clicq, D.; Gzil, P. Anal. Chem. 2005, 77, 4058. (22) Kühne, M.; Ihnen, D.; Moller, G.; Agthe, O. J. Vet. Med. A Physiol. Pathol. Clin. Med. 2000, 47, 379. (23) Meyer, V. R. In Praxis der Hochleistungs-Flu ̈ssigchromatographie, 8 ed.; Salle + Sauerlan̈der: Frankfurt am Main, 1999; pp 106−107, 109. (24) Bundesanstalt, P.-T. PTB-Stoffdatenbla ̈tter SDB 11 Wasser; Physikalisch-Technische Bundesanstalt (PTB): Braunschweig und Berlin, 1995; pp 11. (25) Cabooter, D.; Heinisch, S.; Rocca, J. L.; Clicq, D.; Desmet, G. J. Chromatogr., A 2007, 1143, 121. (26) Heinisch, S.; Desmet, G.; Clicq, D.; Rocca, J. L. J. Chromatogr., A 2008, 1203, 124. (27) Teutenberg, T.; Hollebekkers, K.; Wiese, S.; Boergers, A. J. Sep. Sci. 2009, 32, 1262. (28) Neue, U. D. In HPLC Columns; Wiley-VCH: Weinheim, 1997; pp 55−61. (29) Cabooter, D.; Billen, J.; Terryn, H.; Lynen, F.; Sandra, P.; Desmet, G. J. Chromatogr., A 2008, 1204, 1. (30) de Villiers, A.; Cabooter, D.; Lynen, F.; Desmet, G.; Sandra, P. J. Chromatogr., A 2009, 1216, 3270. (31) Sandra, P.; Vanhoenacker, G. J. Sep. Sci. 2007, 30, 241.

Analytical Chemistry Article

dx.doi.org/10.1021/ac202819v | Anal. Chem. 2012, 84, 1565−15711571