Electrospray ionization mass spectrometry ion suppression/enhancement caused by column bleed for three mixed‐mode reversed‐phase/anion‐exchange high‐performance liquid chromatography columns

Rationale Mixed‐mode reversed‐phase/anion exchange liquid chromatography is useful for separations of mixtures containing anions (e.g. ionized acids). However, when using this form of liquid chromatography with mass spectrometry detection, the bleed of amine‐containing hydrolysis products from the columns may cause ion suppression or enhancement. Methods Using electrospray ionization tandem quadrupole mass spectrometry detection, we determined the ion suppression or enhancement caused by column bleed for three mixed‐mode reversed‐phase/weak anion‐exchange columns containing stationary phases that differ in chemical structure. Two of the stationary phases are based on silica particles, while the third uses ethylene‐bridged hybrid organic/inorganic particles, which have improved hydrolytic stability. Mixtures of acidic and basic analytes were combined with the chromatography flow postcolumn, both with and without a column, and their mass spectrometry ion signal responses (peak areas) were determined. The ratio of signal response with and without a column is the matrix factor. Positive ion electrospray measurements were carried out using 0.1% formic acid (pH ~ 2.7) as a mobile phase additive, and 10mM ammonium formate (pH ~ 6.4) was used for negative ion electrospray detection. Results The matrix factors under both positive and negative ionization modes were closest to 1 (0.74–1.16) for the hybrid particle‐based columns, showing minimal ion suppression or enhancement. In contrast, the silica‐based columns gave matrix factors ranging from 0.04 to 1.86, indicating high levels of ion suppression or enhancement. These results may be explained by the differences in the structures of the stationary phases, which affect the relative amounts of hydrolysis products that elute from the columns. Conclusions The low levels of mass spectrometry ion suppression or enhancement caused by column bleed from the hybrid particle‐based columns should allow for accurate quantitative mass spectrometric detection combined with mixed‐mode reversed‐phase/weak anion‐exchange chromatography.


| INTRODUCTION
Mixed-mode reversed-phase (RP)/anion-exchange (AX) columns, which contain stationary phases having both hydrophobic and AX functionalities, provide a means to retain and separate anions (e.g. ionized acids) under RP conditions. [1][2][3][4][5][6] These columns have been most commonly used with optical detection, for example UV absorbance. When considering their use with electrospray ionization mass spectrometry (ESI-MS), the hydrolytic stability of the stationary phase is an area of concern. Many mixed-mode RP/AX stationary phases, particularly those based on silane-bonded silica, suffer from relatively poor hydrolytic stability. This is reflected in the limited recommended pH ranges for these columns. For example, one brand of RP/AX columns is recommended for use in the pH range 1. 7 while a second has a range of 2.5-7.5. 8 Organic polymer-based mixed-mode columns have been reported to be stable over a wide pH range (e.g. [1][2][3][4][5][6][7][8][9][10][11][12][13][14], but suffer from low efficiencies. 9 In recent work to improve the hydrolytic stability of silica-based mixed-mode RP/AX materials, a polymer-coating approach was used. 10 The resulting stationary phase was shown to have good stability when used with a pH 5 mobile phase at 60 C. To improve both acid and base stability without sacrificing column efficiency, a mixed-mode RP/AX stationary phase based on ethylene-bridged hybrid organic/inorganic particles 11 was recently developed. This stationary phase has been shown to be stable from pH 2 to 10. 12 Ion suppression or enhancement is an important concern in quantitative liquid chromatography (LC) ESI-MS. [13][14][15][16] The primary cause is linked to components that co-elute with the analytes and affect the ionization process. These components may originate from the sample, from mobile phase additives or from contaminants introduced during sample preparation. 13,15,16 Hydrolysis products eluted from the column stationary phase, referred to as column bleed, are another potential source. Under acidic conditions, silane bonded phases hydrolyze to form organosilanols, which may condense to form organosiloxanes. 17 Under basic conditions, not only is the bonded phase susceptible to hydrolysis, but the underlying silica particles may begin to be attacked, forming soluble silicates. 18,19 The stability of different bonded phases varies over a wide range, depending on the attachment chemistry, the structure and surface concentration of the bonded organic groups, and the chemical and physical properties of the underlying particles. 20 Elution of bondedphase hydrolysis products containing amine groups from mixed-mode RP/AX columns has been shown to cause problems due to their strong MS response. These problems include a high background signal in total ion current mode 21 and low ion transmission in highresolution MS. 22 Here, we compare the magnitude of ion suppression or enhancement due to column bleed for hybrid particle-based mixedmode columns relative to two silica-based mixed-mode RP/AX columns. These three stationary phases employ different bonding chemistries as well as different particles. We describe these differences. The column bleed profiles of the three materials were determined using both an acidic mobile phase with positive ionization ESI (ESI+) and a neutral mobile phase with negative ionization ESI (ESIÀ). To quantify ion suppression or enhancement caused by   hydrolysis products eluted from these columns, we monitored seven   small-molecule analytes under ESI+ and five small-molecule analytes   under ESIÀ using analyte-specific selected reaction monitoring (SRM) transitions. The analyte mixtures were combined with the LC mobile phase flow postcolumn, both in the presence and in the absence of a column (using a union instead of a column). The ratio of signal response with and without a column is the matrix factor.

| Experimental plan
To evaluate ESI-MS signal suppression or enhancement due to column stationary phase bleed, seven different analytes under ESI+ and five different analytes under ESIÀ were monitored using analyte-   The stationary phases for columns P and A are based on silica particles. The structures of the bonded phases for these columns have previously been reported 1 and are shown in Figure 2. The surface concentrations of these bonded phases were calculated based on their carbon contents and surface areas. 24 The nature of the attachment to the silica surface for these bonded phases was established using 29 Si cross-polarization magic angle spinning NMR spectroscopy. 25,26 The results indicate that the stationary phase of column A is monofunctionally bonded, while that of column P is trifunctionally bonded. Unlike column W, the stationary phases for columns A and P contain bonded groups that incorporate both hydrophobic and AX moieties together in the same structure. In contrast, the stationary phase of column W is bonded with separate hydrophobic and AX groups, with the AX group surface concentration being approximately five times lower than that of the C 18 groups.

| Effect of column bleed on ESI+ MS signal response
Elution of bonded phase hydrolysis products from columns can affect the signal intensity for co-eluting analytes when using ESI-MS.

| Effect of column bleed on ESIÀ MS signal response
In the ESIÀ experiments performed using a 10mM ammonium formate (pH $ 6.4) mobile phase, the mass spectra of the column   Figure 7). Column W showed only a few peaks with intensities more than 10 times lower than those of the other two columns.
Shown in Figure 8 is a comparison of the matrix factors for the three columns for the representative analytes measured under ESIÀ with 10mM ammonium formate mobile phase additive at 10% acetonitrile ( Figure 8A) and at 90% acetonitrile ( Figure 8B). Column P (black bars) showed matrix factors significantly less than 1 for all the analytes, indicating ion suppression. In contrast, column A (red bars) showed matrix factors slightly greater than 1 for many of the analytes, indicating ion enhancement. While similar trends were observed at 10% and 90% acetonitrile, more severe suppression for column P and enhancement for column A was observed at 90% acetonitrile. In contrast, at both 10% and 90% acetonitrile, column W (blue bars) showed matrix factors ranging from 0.9 to 1.3, indicating little to no suppression or enhancement from column bleed. We attribute this to the greater hydrolytic stability of the hybrid particle-based stationary phase for column W.

| CONCLUSIONS
These results show that the silica-based RP/AX columns (columns A and P) exhibited significant MS signal suppression or enhancement, which appears to be caused by hydrolysis products from the bonded phases and/or the silica particles. In contrast, the hybrid particlebased column W exhibited minimal MS signal suppression or enhancement caused by column bleed. This is due to the improved hydrolytic stability of the stationary phase for column W, which is a consequence of its ethylene-bridged hybrid particles and bondedphase structure, containing separate C 18

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/rcm.9098.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.