Mixed-Mode Chromatography to Mitigate Diluent-Eluent Mismatch

(1)

Mixed-Mode Chromatography to Mitigate Diluent-Eluent Mismatch

Mixed-Mode kromatografi för att mildra oförenligheter mellan diluent och eluent

Olivia Björkman

Department of Engineering and Chemical sciences

Bachelor Thesis in Chemistry (KEGCX5), Drug Analysis (NGKEA) 2021 Bachelor Programme in Chemistry 180 ECTS-credits

Supervisors: Anna Granfors, Morgan Stefansson, Torgny Fornstedt Examiner: Maria Rova

2021-06-04

(2)

Abstract

The Acclaim Mixed-Mode WCX-1, 3 µm 120Å (3.0 x 50 mm), analytical column was used to study the retention of a series of charged amines and test whether neat methanol is a suitable diluent and investigate the analyte's ability to enrich inside the stationary phase. The columns silica particles have covalently bonded alkyl chains (C18), all with terminal carboxylic groups.

The functional groups make it possible to regulate the separation mechanism of the stationary phase by changing the pH and salt content of the eluent. Hydrophobic and weak ionic

exchange interactions coexist and work independently of each other in the Acclaim Mixed- Mode WCX-1 to retain and separate the analytes in the sample.

The column showed its reversed-phase nature at pH = 2.20, well below the pKa of the carboxylic groups bound to the silica sphere packing material (pKa = 4). At pH =6, the column revealed its cationic exchange feature. Injection volumes ranging from 2 – 100 µL were investigated, there were no prominent diluent-eluent incompatibilities at either pH 2.20 or pH 6. The analytes did indeed enrich, and the detection was improved for dilute samples.

Peak deformations occurred at the end of the project, which was thought to be a substance- based dissolution issue, and there was not enough time for further investigation.

Keywords: High-Performance Liquid Chromatography (HPLC), Mixed-Mode, Ion Exchange, Diluent-Eluent Mismatch, Large Injection Volumes, Peak deformation

(3)

Sammanfattning

Den analytiska kolonnen Acclaim Mixed-Mode WCX-1, 3 µm 120Å (3.0 x 50 mm) användes för studium av laddade aminers retention. Provupparbetning med ren metanol som enda diluent, samt analysobjektens anrikning på kolonnen via stora injektionsvolymer studerades.

Den stationära fasens silikapartiklar har kovalent bundna kolvätekedjor (C18), alla med terminala karboxylgrupper. De funktionella grupperna gör det möjligt för användaren att med hjälp av pH och salthalt styra fasens retentionsmekanism. I denna stationära fas samverkar hydrofobicitet och svagt katjonbyte, oberoende varandra, för att retardera och separatera analysobjekten i provet.

Kolonnen demonstrerade sin hydrofoba natur när eluentens pH-värde var lägre än de svagt katjonbytande karboxylgruppernas pKa-värde (pH = 2.00, pKa = 4). Då eluentens pH-värde var större (pH = 6) än karboxylgruppernas pKa-värde befann sig den stationära fasen i sin laddade saltform och uppvisade sina katjonbytande egenskaper. Inga egentligen negativa bieffekter med provupparbetning i ren metanol kunde bekräftas då injektionsvolymer mellan 2 och 100 µL studerades. Låga koncentrationer av analysobjekten anrikades och större

injektionsvolymer gav en ökad detektion för svaga prover. Ett substansbaserat

löslighetsproblem kan ligga bakom de deformerade topparna som uppstod vid vissa av experimenten, det fanns inte tid att undersöka detta vidare.

Nyckelord: Vätskekromatografi, Mixed-Modekromatografi, Jonbyteskromatografi, Diluent- Eluent-Missanpassning, Stora Injektionsvolymer, Toppdeformering

(4)

1. Introduction

The Italian-Russian botanist Ph.D. Mikhail Semenovich Tsvett (1872 -1919) invented adsorption chromatography in the early 1900s. Tsvett separated pigments in plant extracts through columns packed with adsorptive material like alumina or sucrose. The samples were prepared by extracting the pigments in powdered plant material with organic solvents. The pigments are separated in the glass column into colored bands according to the strength of the interaction between the pigments and the packing material. Tsvett named the procedure chromatography, which is the Greek word for "written in color." The technique was not practiced until the 1930s, years after Tsvetts' departure [1, 2].

Today, high precision liquid chromatography (HPLC) is one of the most widely used separation techniques in analytical chemistry and a workhorse in most laboratories. The instrumentation is quite simple, an injector, a pump, a column, and a detector. The analytes adsorb and desorb thousands of times inside the stationary phase (column) and are eluted, one after the other, and detected as Gaussian-shaped peaks in a chromatogram [3, 4]. The signals are detected in the UV-VIS spectrum expressed as absorbance units (mAU). For added sensitivity, a mass spectrometer can be connected to the instrument.

Commonly in HPLC, the stationary phase is packed with porous silica particles covalently bonded with octa or octadecyl hydrocarbon chains. The sample is transported through the column by the mobile phase flow, and separation based on molecular structure begins. The separation is caused by a multitude of equilibria reactions between the solutes, ligands, free ions of the mobile phase to the stationary phase [3, 5].

1.1 Reversed-Phase Chromatography

Reversed-phase liquid chromatography (RPLC) is a popular mode for separating a large variety of species. It is called the reversed-phase because it is dissimilar to the first mode of HPLC, the normal phase, which utilizes nonpolar eluent and polar columns. RPLC is characterized by a nonpolar stationary phase (e.g., C18) and a mixture of water and organic solvents (e.g., methanol or acetonitrile) as a polar mobile phase [3].

The retention can be caused by other than hydrophobic interactions, charged species in the stationary phase such as underivatized silanols (-SiO^-) or additives can interact with

protonated bases via cationic-exchange, and neutral silanols (-SiOH) or additives can donate a proton to an accepting solute through a hydrogen bond. The shape of the solute can influence retention [5].

The retention mechanism and selectivity in RPLC is constituted by five solute-column interactions: interactions between the phase's and the solute's hydrophobicity, hydrogen bonding, cation exchange with ionized silanol groups, and steric resistance (not the same as shape selectivity). The expression in equation 1 describes the five column-solute interactive powers of reversed-phase chromatography [5, 6, 7].

log 𝛼 = log 𝑘 𝑘⁄ _𝑟𝑒𝑓= 𝜂´𝐻 − 𝜎^′𝑆^∗+ 𝛽^′𝐴 + 𝛼^′𝐵 + 𝜅′𝐶 (1) The terms H, S, A, B, and C are eluent- and temperature-independent properties of the

stationary phase, while the terms η', σ', β', α', κ' are either empirical, eluent- and temperature-

(6)

dependent properties of the solute. The hydrophobic interaction occurs between the columns and the solute's "fatty" parts, such as between long hydrocarbon chains and cyclic structures.

The hydrogen bonding happens between the silanol groups and the solute. An example of cationic exchange is when a charged amine is retained by its interaction with a deprotonated silanol. The steric selectivity can dictate whether the solute is small enough to enter the space between the ligands of the silica particles [7].

𝑘 =

^𝑘⁰^+𝑘¹^∙10^{(𝑝𝐾𝑎−𝑝𝐻)}

1+10^{(𝑝𝐾𝑎−𝑝𝐻)} (2)

The retention mechanism of a basic and a neutral compound during reversed-phase chromatography can be described with equation 2, where k0 and k1 are the neutral and charged solute retention factors, respectively [8].

1.2 Ion-Exchange Chromatography

Ion-exchange chromatography (IEX) is a mode of chromatographic separation of polar species such as inorganic ions, biomolecules, carbohydrates, carboxylic acids, sample preparation, and two-dimensional separations. The IEX column has polar groups attached to the otherwise nonpolar stationary phase, e.g., sulfonate groups for anionic exchange and ammonium groups for cationic exchange, interacting with the charged solutes [4, 9].

A combination of ion-exchange and reversed-phase interactions with the stationary phase causes the separation of analytes; the eluent for IEX separations is typically comprised of a counter-ion, a buffer, water, and an organic modulator. Both the organic modifier and the counter-ion regulates the strength of the eluent and the retention of the sample. The retention is typically regulated by changing the counter-ion salt concentration. The increase of the salt concentration correlates to a decrease in the retention of ionic solutes due to repulsion [9, 10].

An eluent with high ionic strength suppresses ion exchange mechanisms in favor of hydrophobic interactions. There is a competition between the sample ions and the counter- ions, and when the salt concentration increases, there is a shift in the equilibrium so that the counter-ion is bound to the stationary phase rather than the charged solutes.

The retention factor k is directly proportional to the ionization of monovalent solutes; in other words – the dissociation, association, and retention of acids and bases are pH-dependent [9].

The retention in IEX can be explained with equations 3-5, where the interaction with the charged groups of the column, in this case, a carboxylic group that is bound to the carbon chain, is described by equilibrium constant K1 and depends on the concentration of dissolved species [X⁺] and counter-ion, in this case, sodium. The second and third reactions are

described by equilibrium constants K2 and K3, respectively, and are both acid-base reactions [9].

𝑅𝐶𝑂𝑂𝑁𝑎 + 𝑋⁺

𝐾1,[𝑋⁺],[𝑁𝑎⁺]

↔ 𝑅𝐶𝑂𝑂𝑋 + 𝑁𝑎⁺ (3)

𝑋⁺+ 𝑂𝐻⁻

𝐾₂,𝑝𝐻

↔ 𝑋𝑂𝐻 (4)

(7)

𝑅𝐶𝑂𝑂𝑁𝑎 + 𝐻⁺

𝐾3,𝑝𝐻

↔ 𝑅𝐶𝑂𝑂𝐻 + 𝑁𝑎⁺ (5)

𝐾₁ = ^[𝑋⁺^{][𝑅𝐶𝑂𝑂𝑁𝑎]}

[𝑁𝑎⁺][𝑅𝐶𝑂𝑂𝑋] (6)

The equilibrium constant for the cation exchange reaction in (3) is the displacement of the counter-ion by the solute ion seen in equation 6. For the reverse reaction, the equilibrium constant is the inverse of equation 6 [9].

log (𝑘) = 𝑎 − 𝑚𝑙𝑜𝑔(𝐶) (7)

The retention in ion-exchange chromatography can be described by equation 7. The logarithm of the retention factor (k) equals the constant a (a = log(k) for C = 1M) subtracted by m (the charge of the solute) times the logarithm of the molar concentration of the counter ion C. The constants a and m are dependent on the analyte, stationary phase, salt content, eluent pH, and temperature [4].

1.3 Mixed-Mode Chromatography

The combination of RPLC or HILIC (hydrophilic interaction liquid chromatography) systems with ion-exchange systems is often referred to as mixed-mode chromatography (MMC).

Mixed-mode chromatography is similar to IEX, though the mixed-mode columns are far more hydrophobic than IEX columns. In MMC, two or more mechanisms for retention are

combined in a single chromatographic system, which makes MMC suited for sample preparation and analysis of complex samples that contain both ionizable and neutral

molecules. Mixed-mode columns are an alternative to ion-exchange columns, its applications for biological samples are unique as they can be directly injected for purification [10].

Typically, biological matrices require extensive purification before analysis [11].

The retention in MMC can be controlled by changing the buffer, organic modifier content, pH, and salt concentration. Mixed-mode columns are used for bases, heavily retained by the charged silanol groups in an RPLC column. Mixed-mode columns are capable of separating anions, cations, zwitterions, and uncharged species all at once [4]. MMC retains polar pharmaceutical substances and other small, polar compounds as opposed to RPLC [10].

1.4 Sample-solvent incompatibility

Suppose the diluent has a stronger eluting power than the eluent. In that case, the sample may not be entirely diluted by the mobile phase after the injection, resulting in premature eluted, deformed peaks. More dilute samples or smaller injection volumes can mitigate the effects of sample solvent mismatch [4] [12].

(8)

1.5 Aim of the study

This study aims to find a way to mitigate the effects of diluent-eluent incompatibility in high- performance liquid chromatography for combined pharmaceutical samples with automatic sample preparation. The samples are dissolved in neat methanol, which has proven to bring deformed peaks in liquid chromatography or precipitation inside the column, see figure 1. The Acclaim mixed-mode weak cation-exchange stationary phase is considered an effective competitor to regular reversed-phase columns for samples containing ionizable hydrophobic species.

1) Demonstrate the dual nature of the Acclaim Mixed-Mode WCX-1 and show that the polar substances can be retained.

2) Demonstrate that samples prepared with neat methanol works.

3) Demonstrate that the species can be enriched inside the column.

A series of common beta-blockers were chosen for the project, which all would be positively charged in the operating range of the Acclaim Mixed-Mode WCX-1 (pH = 2.5-6.5).

Simultaneous organic solvent and the salt gradient were tested for the separation of the analytes. This project is aimed towards the type of methods used in the pharmaceutical industry; an automated sample preparation routine performed by a robot. The advantages of introducing more intelligent solutions, like robotics, are eliminating the human factor, better reproducibility, using harmful chemicals with minimal exposure to the laboratory staff, etc.

The sample preparation in this project should be a simple procedure to apply to an automated routine.

Figure 1. This figure demonstrates how difficult it is to attain a good peak of salbutamol sulfate in pure methanol. In this experiment, an ion-pair reagent was added to the eluent, and the salbutamol peak has a shoulder. The arrow points at an enlargement of the salbutamol peak. Source AstraZeneca.

(9)

2. Experimental

2.1 Apparatus

The experiments were performed on an Agilent Technologies 1200 series chromatographic system from Agilent Technologies (Palo Alto/ Santa Clara, CA, USA), consisting of an autosampler (G1329A), binary pump (G1312A), and diode array UV detector (G1315D). The software was ChemStation. After a few weeks, there was a move to another lab, and the rest of the project was performed with an Agilent technologies 1260 BIO Infinity series

chromatographic system.

2.2 Column Properties

The Acclaim Mixed-Mode WCX-1, 3 µm 120Å (3.0 x 50 mm), lot nr 071910, was used for the experiment. The column is packed with silica particles with covalently bonded octadecyl alkyl chains with terminal carboxyl groups, see figure 2.

Figure 2. The column material, every alkyl chain is ended with a functional group.

A few days into the project, the Agilent 1200 system or the MMC column had some issues with the performance; therefore, it was decided that control of the system performance had to be done. The performance tests were executed isocratic (25 % MeCN) on an Xbridge C18 BEH Technology Waters column known to perform correctly from a previous project. Four substances were chosen as test subjects, thiourea which would elute in the void, methyl mandelate, phenol, and phenol-3-propanol. Neither the MMC column nor the instrument failed the test. The test was also performed on the Agilent 1260; for chromatograms, see appendix A.

2.3 Chemicals

Acetonitrile (≥ 99.9 %, VWR, HiPerSolv Chromanorm quality), methanol (99.8 %, Fisher chemicals, HPLC grade), ortho-phosphoric acid (≥ 85 %, Fluka, analytical quality), acetic acid (99.8 – 100.5 %, Merck Chemicals, analytical quality), sodium hydroxide (99.2 %, VWR chemicals, AnalaR NORMAPUR), Titrisol, Sodium hydroxide solution (1.000 M Merck

(10)

Chemicals, analytical quality), alprenolol hydrochloride (CRS), atenolol (Sigma-Aldrich), formoterol fumarate dihydrate (AstraZeneca), metoprolol (used as received, from synthesis at Karlstad University), metoprolol succinate (AstraZeneca), (±)-propranolol hydrochloride (Sigma), salbutamol sulfate (AstraZeneca). The samples of budesonide, formoterol fumarate dihydrate, glycopyrronium bromide, metoprolol succinate, and salbutamol sulfate were gifted from AstraZeneca and used as received.

2.4 Buffers and Eluents

To avoid false peaks, the pH was measured in a small amount of the buffers in a separate container, and the buffer was then discarded. This made the pH harder to keep constant throughout the project. pH was in the range of 5.8-6.3 was accepted since most of the carboxyls' were then deprotonated.

2.4.1 Preparation of Buffers Phosphoric acid buffer pH = 2.20:

250 ml of the buffer was prepared by diluting 169 µL of 85 % phosphoric acid in a volumetric flask with MilliQ-water to the mark. The solution was transferred to a bottle with a screw top.

Acetic acid buffer pH = 6:

Since a salt gradient was desired, a fixed amount of the counter ion was titrated with acid to the desired pH. In this case, the salt concentration was set to 50 mM, and the required amount of acid needed was determined by a reiteration of the Henderson-Hasselbalch equation, see equation 8.

[𝐻𝐴] = ^[𝐴⁻^]

10^{𝑝𝐻−𝑝𝐾𝑎} (8)

= ^{166.67 𝑚𝑀}

10^6.00−4.75 = 9.37 𝑚𝑀 + 166.67 𝑚𝑀 = 176.04 𝑚𝑀

A 1.000 M sodium hydroxide solution was prepared from an ampoule. 166.67 mM of the sodium hydroxide was titrated with 176 mM of acetic acid, the pH was noted in a separate vessel. The solution was transferred to a bottle with a screw top.

2.4.2 Preparation of Eluents

Table I contain an overview of the eluent composition. The low pH eluent contains

phosphoric acid buffer with pH = 2.20, and the high pH eluent contains sodium acetate buffer with pH 6.

Low pH eluent

The phase for the A-line was prepared by diluting 169 µL of 85 % phosphoric acid with water to 250 mL in a volumetric flask, then adding it to a bottle with a screw top. 250 mL for the B- line was prepared by diluting 169 µL of 85 % phosphoric acid with water to 25 mL, adding it and 225 mL of acetonitrile to a bottle with a screw top. The phase for the B-line was prepared one day in advance to let the mixture come to room temperature or sonicated for ten minutes in warm water if prepared right before use.

(11)

High pH eluent

A screw-top flask was filled with water for the A-line. 250 mL of the phase for the B-line was prepared by adding 75 mL of 166.67 mM sodium acetate buffer to a 250 mL screw top flask and adding 175 mL acetonitrile to it. The phase for the B-line was prepared one day in advance to let the mixture come to room temperature or sonicated for ten minutes in warm water if prepared right before use.

Table I. An overview of the eluents.

Buffer system

pH A-line B-line

Phosphoric acid

2.20 10 mM of

phosphoric acid in water

10 mM of phosphoric acid in water and acetonitrile (30:70 v/v)

Sodium acetate

6 Water 50 mM of NaOH titrated with acetic acid to the desired pH in water and acetonitrile (30:70 v/v)

2.5 Analytes

In tables II and III, the molecular structure, charge, weight, pKa, and the logP of the analytes are stated.

Table II. The charge of the analytes in the operating pH of this study and the molecular structure of the analytes, source ChemAxon.

Substance Charge Molecular structure Alprenolol hydrochloride +1

Atenolol +1

Budesonide 0

(12)

Formoterol fumarate dihydrate

+1

Glycopyrronium bromide +1

Metoprolol base +1

Metoprolol succinate +1

Propanolol hydrochloride +1

Salbutamol sulfate +1

Table III. The analytes used for the study, their molecular weight and pKa (strongest basic), source ChemAxon.

Substance Molecular weight (gmol^-1) pKa logPoct

Alprenolol hydrochloride 285.81 9.67 2.69

Atenolol 266.33 9.67 0.57

Budesonide 430.53 -2.90 2.73

Formoterol fumarate dihydrate 420.46 9.81 1.06

Glycopyrronium bromide 398.34 -4.30 -1.40

Metoprolol base 267.36 9.67 1.76

Metoprolol succinate 652.82 9.67 1.76

Propanolol hydrochloride 295.81 9.67 3.03

Salbutamol sulfate 288.35 9.40 0.44

(13)

2.6 Gradient Design and General Instrumentation

The gradient elution is described in table IV and figure 3; this gradient is used for both pH 2.20 and 6. The acetonitrile content, the steepness of the gradient, and the time for equilibrium were tuned until acceptable retention and reproducibility were attained.

The samples were eluted by a linear gradient over ten minutes, plateauing for two minutes.

The final acetonitrile concentration was 56 %, back to equilibrium with the aqueous phase for another 13 minutes; the total time for one run was 25 minutes. Phase A was an aqueous buffer or plain water, and phase B was a 70:30 (v/v) mixture of acetonitrile and an aqueous buffer/

water. The solvent front eluted after 0.8 minutes.

Table IV. Timetable for the gradient, A is the aqueous phase, and B is the organic phase.

Time (min) A (%) B (%)

0.00 100 0

10.00 20 80

12.00 20 80

12.01 100 0

25.00 100 0

Figure 3. Programmed progression of binary solvent, the linear gradient plateaus after ten minutes with a total MeCN concentration of 56 % for two minutes, then follows 13 minutes of equilibrium with 0% MeCN.

Most parameters were kept constant, e.g., the flow rate, see table V. If nothing else is stated, the experimental conditions are in table V.

Table V. The general instrumentation

Instrument Agilent technologies 1260 Infinity BIO series Stationary phase Acclaim mixed-mode WCX-1 3 x 5 mm, 3 µm

Flow (mL/min) 0.400

Injection volume (µL) 5.00 Sample concentration (mM) 1.00

Phase A Phosphoric acid 10 mM/ Water

Phase B 70 % Acetonitrile + 30 % aqueous phase

Gradient (%B) Linear 0-80

Detection (nm) DAD, 280.4

Pressure (bar) 67-70

Temperature (C) Room temperature

(14)

2.7 Procedures

2.7.1 Preparation of the drug substance stock solutions

Stock solutions of 1 mg/mL were prepared by weighing out 10 mg of each substance and dissolving them in 10 mL of methanol, respectively. The solutions were filtrated through a 0.2 μm polypropylene syringe filter (VWR International).

2.7.2 Sample Preparation

0.10 mM samples were prepared from the stock solutions by diluting them with water or methanol.

2.7.3 Preparation of a Dilution Series for Larger Injections

To increase the system's sensitivity, it was suggested to inject more significant volumes of dilute samples, trying to achieve enrichment of the analytes on the stationary phase. The analytes chosen for the study were the least hydrophobic substance salbutamol sulfate, metoprolol in its base form and salt form, and the most hydrophobic substance, propranolol hydrochloride. Four samples of each analyte were made in the dilutions 1mM, 0.5 mM, 0.1 mM, 0.05 mM, with the intention of injecting 2, 5, 10, 20, 40, 60, 80, and 100 µL, see table VI.

Table VI. The setup for the investigation of injection volumes.

Concentration (mM) Volume (µL) Amount of substance (mol)

1.00 2 2·10^-12

5 5·10^-12

0.50 10 5·10^-12

20 1·10^-11

0.10 40 4·10^-12

60 6·10^-12

0.05 80 4·10^-12

100 5·10^-12

(15)

3. Results

3.1 pH 2.20

A series of injections were performed with a phosphoric acid buffer and an acetonitrile gradient to confirm the stationary phase's hydrophobic properties. In this mode, the columns carboxylic surface was completely protonated and uncharged. The retention of the least hydrophobic specimens, atenolol, and salbutamol sulfate, was 1.288 and 1.134 minutes, respectively. In figure 4, there is a chromatogram with all the charged amines showing the selectivity from the least to the most hydrophobic in the order left to right. The absorption of alprenolol hydrochloride was lacking in this experiment and could not be integrated,

alprenolol HCl retention time is approximately seven minutes.

Figure 4. Agilent 1200 chromatographic system, pH=2.20, 5 µL injections, 0.1 mM samples in water, 220 nm. From left to right salbutamol sulfate, atenolol, metoprolol succinate, metoprolol base, formoterol fumarate dihydrate, alprenolol hydrochloride, and propranolol hydrochloride. All the analytes eluted with the gradient.

The switch from ̴ 100 % protonation to ̴ 95% deprotonation of the carboxylic groups is approximated to require 4.5 hours at the given flow rate and salt content (0.4 mL/min and 40 mM Na⁺ ions gradient). The protonation of the column did not require a long time; less than one hour of equilibrium time produced a stable system. See the result of the repeated

injections of metoprolol succinate during the eluent shift from pH = 2.20 to pH = 6 in figure 5. The first injection of metoprolol succinate was retained for 5.707 minutes. After 4.5 hours, the analyte was retained for 11.4 minutes.

(16)

Figure 5. Metoprolol succinate in methanol was first retained by hydrophobic interactions and is gradually interacting more and more with the weak cation exchange feature of the stationary phase.

3.2 pH 6

The most polar substances, atenolol, and salbutamol sulfate, were successfully retained in the mixed mode. In this mode, the stationary phase is in its salt form. Figures 6 and 7 are

chromatograms from one month apart, and their eluents are of different batches. Atenolol and salbutamol sulfate have higher retention in figure 7. The samples in figure 6 are diluted in water, while the samples in figure 7 are diluted in methanol. In figure 6, the five rightmost analytes, metoprolol salt and base, formoterol, alprenolol hydrochloride, and propranolol hydrochloride, eluted isocratically. The former two eluted while the organic solvent and salt content were at their peak. In figure 7, the uncharged corticosteroid budesonide is included as a part of application testing. Glycopyrronium bromide was undetectable at 280.4 nm and was only detectable at 254,4 nm just after the UV cutoff of the acetate ion in the buffer.

Glycopyrronium bromide was retained for 13 minutes, see figure 8.

In table VII, the analytes retention times in the two modes can be compared to each other.

(17)

Figure 6. 1.0 mM samples in water. From left to right, atenolol, salbutamol sulfate, metoprolol base, and succinate co-eluting as one peak, formoterol fumarate dihydrate, alprenolol hydrochloride, and propranolol hydrochloride.

Figure 7. 1 mM samples in methanol. From left to right atenolol, salbutamol sulfate, budesonide (the split peak), metoprolol base, and succinate co-eluting as one peak (bright green), formoterol fumarate dihydrate, alprenolol hydrochloride, and propranolol

hydrochloride.

(18)

Figure 8. pH=6, 5.0 µL injection, 0.1 mM glycopyrronium bromide, 254.4 nm.

Table VII. The retention times of the analytes, in the left column, the reversed-phase mode, and for the mixed mode in the right column.

Substance Retention time (min)

RPLC

Retention time (min) Mixed-Mode

Alprenolol hydrochloride 7.100 12.786

Atenolol 1.288 9.101

Budesonide - 10.739; 10.889 *

Formoterol fumarate dihydrate 6.461 12.106

Glycopyrronium bromide - 13.071

Metoprolol base 4.460 11.503

Metoprolol succinate 4.442 11.560

Propranolol hydrochloride 7.823 14.238

Salbutamol sulfate 1.134 9.324

* Budesonide is an epimer appearing as two peaks in the chromatogram

(19)

3.3 Methanol as the diluent, pH 6

Figure 9 shows the result from injected samples with water and methanol as the diluent, in low pH, and no apparent difference in peak shape, response, or retention time was observed.

Figure 10 shows the same experiment but in the mixed mode, pH 6.

Figure 9. 10 mM phosphoric acid + 0-56 % MeCN gradient elution, pH=2.20, 5 µL injections, 0.1 mM samples, 280.4 nm. Metoprolol succinate samples with methanol and water as the diluent.

Figure 10. 0-40 mM sodium acetate + 0-56 % MeCN gradient elution, pH=6, 5 µL injections, 1.0 mM samples, 280.4 nm. Metoprolol succinate samples with methanol and water as the diluent (six injections of each), this time in the mixed mode.

3.4 Large Injection Volumes

The analytes chosen for the study were the least hydrophobic substance salbutamol sulfate, metoprolol in its base form and salt form, and the most hydrophobic substance, propranolol hydrochloride. The volumes and concentration are found in Table VI.

(20)

Figure 11 a and b shows the 2 µL and 100 µL injections of salbutamol sulfate, respectively.

The results for the three other analytes in the testing were similar; see figure 12 a and b for metoprolol base, figure 13 a and b for metoprolol salt, and figure 14 a and b for propranolol hydrochloride. The 2 µL injects roughly half the moles as the 100 µL injections do; this is noticeable in the y-axis as the response is twice as high in the latter case.

Figure 15 shows repeated injections of propranolol HCl, set up according to Table VI, the peak height decreases while the peak width and the retention increase with increased injection volume. The area response is linear to the injected amount of substance; following the

Lambert-Beers law, see the linear relationship in figure 16. The data used for the graphs in figure 16 are the peak area and the molar amount; the data are found in table I, appendix B.

Lambert-Beers law (𝐴 = 𝜀 ∙ 𝑏 ∙ 𝑐) states that the relative absorption of a sample relates to the molar absorptivity coefficient, the optical path length, and the sample concentration.

Figure 11. 0-40 mM sodium acetate + 0-56 % MeCN gradient, pH=6, a) 2 µL injection, 1.0 mM of salbutamol sulfate in methanol, and b) 100 µL injection, 0.05 mM of salbutamol sulfate in methanol.

a)

b)

(21)

Figure 12. 0-40 mM sodium acetate + 0-56 % MeCN gradient, pH=6, a) 2 µL injection, 1.0 mM of metoprolol base in methanol, and b) 100 µL injection, 0.05 mM of metoprolol base in methanol.

a)

b)

a)

(22)

Figure 13. 0-40 mM sodium acetate + 0-56 % MeCN gradient, pH=6, a) 2 µL injection, 1.0 mM of metoprolol succinate in methanol, and b) 100 µL injection, 0.05 mM of metoprolol succinate in methanol.

Figure 14. 0-40 mM sodium acetate + 0-56 % MeCN gradient, pH=6, a) 2 µL injection, 1.0 mM of propranolol HCl in methanol, and b) 100 µL injection, 0.05 mM of propranolol HCl in methanol.

b) a)

b)

(23)

Figure 15. 0-40 mM sodium acetate + 0-56 % MeCN gradient, pH=6, repeated 2, 5, 10, 20, 40, 60, 80 and 100 µL injections, 1.0, 0.5, 0.1, and 0.05 mM samples of propranolol HCl, 0.400 mL/min, 280.4 nm. The retention and peak width increases, while the peak height decreases as the injected volume increase.

Figure 16. The peak area was plotted against the amount of analytes injected. There is a linear relationship between the two.

y = 6E+14x - 462.47 y = 2E+14x - 165.98 y = 3E+14x + 93.445 y = 9E+14x - 108.78

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

0 2E-12 4E-12 6E-12 8E-12 1E-11 1.2E-11

Peak Area

Moles injected

Repeated Injections 2 - 100 µL

Metoprolol succinate Metoprolol base Salbutamol sulfate Propranolol HCl

Linear (Metoprolol succinate) Linear (Metoprolol base) Linear (Salbutamol sulfate) Linear (Propranolol HCl)

(24)

While no differences were noted between aqueous and methanolic samples, peak splitting and tailing occurred in some methanolic samples. This appeared when the method was rewritten so that the gradient continued for two more minutes resulting in 70 % MeCN and 50 mM Na⁺ inside the column (the previous method had a maximum of 56 % MeCN and 40 mM Na⁺ in the mobile phase). Formoterol fumarate dihydrate and salbutamol sulfate experienced peak distortions during the new gradient. The splitting of formoterol grew more intense as the injections were repeated, see figure 17. The distortion disappeared when the method was changed back to its previous settings, as seen in figure 18. Injection volumes of 100 µL worsened the peak splitting, while injection volumes of 2 µL reduced the splitting – this was true for both method gradients.

(25)

Figure 17. 0-50 mM sodium acetate + 0-70 % MeCN gradient, pH=6, repeated 2 µL injections, 1.0 mM samples. a) Formoterol fumarate dihydrate, the peak split over time, causing a shoulder that grew with every injection. Using the sodium acetate buffer directly after an acid wash of the system, the system stabilizes after the first injection (blue peak). b):

an enlargement of the deformed peaks in a), the main peak height decreases as the deformation worsens.

a)

b)

(26)

Figure 18. 0-40 mM sodium acetate and 0-56 % MeCN gradient elution, pH=6, 2 µL injections, 1.0 mM samples. Formoterol fumarate dihydrate five repeated injections. The system appears to be stable with this gradient.

(27)

3.5 Application

The neutral corticosteroid budesonide and the positively charged amine formoterol fumarate dihydrate are used together as active pharmaceutical ingredients in an inhalation product.

Figure 19 shows a chromatogram of a sample containing both substances. Budesonide is an epimer that produces two peaks in the chromatogram at 10.7 and 10.8 minutes, and formoterol was retained for 12.1 minutes. The fumarate ion elutes in the void at 0.48 minutes.

Figure 19. A two-component sample of budesonide and formoterol fumarate dihydrate, budesonide elutes as two peaks at 10.7 and 10.8 minutes.

(28)

4. Discussion and Conclusions

Acclaim Mixed-Mode WCX-1 retained and separated the subjects in the reversed-phase mode (pH 2.20), though the most polar substances, atenolol and salbutamol sulfate, were not

retained for very long. There was no noticeable difference in peak shape between methanol and water as the diluent for metoprolol succinate in this mode. In retrospect, a poorly retained polar substance such as salbutamol sulfate would have been the better test subject than the

"best looking" substance metoprolol succinate. A poorly retained substance in methanol would most likely have shown some measure of incompatibility with the eluent.

The switching of column modes from the reversed-phase mode to the mixed mode was finished after four and a half hours. It was executed by repeated injections with sodium acetate and acetonitrile gradient. The reversal from the salt form to the neutral form took approximately one hour. To hasten the deprotonation, one could use a narrower salt gradient, e.g., starting with 5-10 mM of sodium acetate in the A-line and proceed with the gradient as before with a higher flow rate. The hydrophobic and weak cationic exchange retention mechanisms coexist and separate the analytes efficiently in this mixed mode.

The very small and polar substance glycopyrronium bromide, a quaternary ammonium see table II in section 2.5, is one of the most retained ones in mixed mode; the reason may be that it creates a strong complex bond with the ion-exchange groups. On the other hand,

budesonide is neutral and had excellent retention in the mixed mode due to its hydrophobicity.

Both methanol and water work as the sample diluent in the mixed mode. Typically, there would be a diluent-eluent mismatch in reversed-phase chromatography when injecting such large volumes of neat methanol. The 20 µL injections gave the largest amount of moles injected. In hindsight, the 20 µL sample should have been injected from the 0.10 mM vial to mimic the others. Generally, the peak symmetry was improved when the injection volumes were < 10 µL or > 60 µL, which is not surprising considering that the 10 – 60 µL injections were the most concentrated ones. The injections showed a linear relationship between the peak area and the amount of injected substance.

A combined sample containing the neutral steroid budesonide and the positively charged amine formoterol fumarate dihydrate was successful, see figure 19. This makes the column interesting for further studies of combined pharmaceutical samples.

(29)

For the studies at pH 6, it was challenging to find a suitable buffer. Phosphoric acid did unfortunately not work since phosphates started to precipitate when the buffer was combined with the acetonitrile. For this study, it was decided to use acetic acid as the best non-toxic choice, but unfortunately, it had a significant absorbance in the wavelength area 210-250 nm.

The possibility of contaminated acetic acid was tested by preparing a new buffer from an unopened flask of acetic acid; the background absorption was still there and was caused by the acetate ion (UV cutoff = 240 nm). The background noise of the eluent saturated the detector at wavelengths lower than 250 nm, and the study had to be performed at higher wavelengths with the consequence that the absorbance of the analytes was poor, often < 50 mAUs. An inclination of the baseline was still visible at 280.4 nm for 0.1 mM samples, but new samples with ten times the strength gave a straight and smooth baseline and detection >

300 mAUs. The peak symmetry suffered for this.

Salbutamol tended to shift its retention; at times, it co-eluted with the metoprolol's, and at other times, salbutamol eluted a few minutes earlier than the metoprolols. Atenolol did also experience a shift in retention times; see figures 4 and 5. One possible explanation is that the substances are sensitive to changes in the pH of their environment; the substances are two of the most polar substances in this study (logP = 0.44 and logP = 0.57 for salbutamol sulfate and atenolol, respectively, see Table III). One could argue that a sample plug of salbutamol sulfate would be more acidic and polar than a sample plug of atenolol since two moles of salbutamol is accompanied by one mole of sulfuric acid.

In the project, pH 5.8-6.3 were used as high pH buffers, the range is generous, and if more attention were paid to attaining the same pH, perhaps this would not be an issue. Interestingly, the two most polar substances go through the same phenomenon while all the other

substances remain unchanged.

The peak deformation of formoterol in figure 17 may be caused by a sample stability problem causing the substance to precipitate, and the spectral information showed that all the peaks in figure 17 are formoterol. The splitting and tailing manifested themselves when a change of the gradient design was made, and the deformations were present for both the 2 and 100 µL injections. The splitting and tailing subsided little by little as the method was changed towards its starting point. As seen in figure 18, the old gradient design did not show the deformation and produced reproducible results.

In conclusion, the Acclaim Mixed-Mode WCX-1 analytical column showed promising results for the separations of samples with a combination of charged, polar amines, and neutral compounds, such as often are found in pharmaceutical products. The column has the potential of mitigating the problem of deformed peaks for samples dissolved in methanol, especially in chemical analyses of inhalation formulations that contain combinations of budesonide and salbutamol sulfate or formoterol fumarate dihydrate. Future studies are needed to investigate the substance-based precipitation phenomenon, confirm the robustness of the separation, and optimize the experimental conditions.

(30)

5. References

[1] John Daintith (18 August 2008). Biographical Encyclopedia of Scientists, Third Edition. CRC Press. p. 754. ISBN 978-1-4200-7272-3.

[2] Asimov, I. (1982). Asimov's biographical encyclopedia of science and technology: The lives and achievements of 1510 great scientists from ancient times to the present chronologically arranged. Second edition. p. 641. Garden City, N.Y:

Doubleday.

[3] Fornstedt, Torgny & Forssén, Patrik & Westerlund, Douglas. (2015). Basic HPLC Theory and Definitions: Retention, Thermodynamics, Selectivity, Zone Spreading, Kinetics, and Resolution. 10.1002/9783527678129.assep001.

[4] Lloyd R. Snyder, Joseph J. Kirkland, John W. Dolan. Introduction to Modern Liquid Chromatography. Third edition. John Wiley & Sons, Inc., Hoboken, New Jersey.

Chapter 7.

[5] Horváth, Csaba & Melander, Wayne & Molnár, Imre. (1976). Solvophobic Interaction in Liquid Chromatography with Nonpolar Stationary Phases. Journal of Chromatography A - J Chromatogr A. 125. 129-156. 10.1016/S0021-9673(00)93816-0.

[6] Snyder, Lloyd & Dolan, John & Marchand, Daniel & Carr, Peter. (2012). The Hydrophobic-Subtraction Model of Reversed-Phase Column Selectivity. Advances in chromatography. 50. 297-376.

[7] Snyder LR, Dolan JW, Carr PW. The hydrophobic-subtraction model of reversed-phase column selectivity. J Chromatogr A. 2004 Dec 10;1060(1-2):77-116.

PMID: 15628153.

[8] Horvath, C. Melander, W. Molnar, I. & Molnar, P. (1977). Enhancement of retention by ion-pair formation in liquid chromatography with nonpolar stationary phases. Analytical Chemistry. 49. 2295-2305.

[9] Veronika R. Meyer. Practical High-Performance Liquid Chromatography. 5th Edition. ISBN: 978-0-470-68218-0

[10] Zhang, Kelly & Liu, Xiaodong. (2016). Mixed-mode chromatography in pharmaceutical and biopharmaceutical applications. Journal of Pharmaceutical and Biomedical Analysis. 128. 10.1016/j.jpba.2016.05.007.

[11] D. Westerlund, K-E. Karlsson, A-M. Tivert, H. Ehrsson, K-G.Wahlund, O.

Gyllenhaal, R. Isacsson, C. Pettersson, J. Vessman, T. Arvidsson, P. Hartvig, B-A.

Persson. (2014). Separation Methods in Pharmaceutical and Biomedical Analysis.

Chapter 14. Uppsala University.

[12] Guiochon G., Felinger A., Shirazi D.G., & Katti A.M.. (2006). Fundamentals of Preparative and Nonlinear Chromatography. Second Edition. Elsevier academic press.

(31)

Appendix A

A performance test was performed on the chromatographic system (Agilent 1200) and the MMC stationary phase, see figure A.The sample contained thiourea 0.12 mM, methyl mandelate 0.5mM, phenol 0.5 mM, 3-phenyl-1-propanol 0.5 mM. This test result was

compared to the Xbridge C18 BEH Technology Waters results known to perform beautifully.

Both the Agilent 1200 and the MMC stationary phase passed the test. When the relocation was made, the very same sample was tested on the new instrument Agilent 1260 infinity series; see figure B for the result.

Stationary phase Acclaim mixed-mode WCX-1 3 x 5 mm, 3µm Flow (mL/min) 0.425

Injection volume (µL) 5.00

Eluent 25:75 acetonitrile and water Gradient (%B) Isocratic

Detection (nm) 210

Temperature (C) Room temperature

Figure A. A performance test on the Agilent 1200 series

Figure B. A performance test on the Agilent 1260 infinity series

(32)

Appendix B

Table I, injected amount of moles and peak area.

Moles Injected

Metoprolol succinate Area

Metoprolol base Area

Salbutamol sulfate Area

Propranolol HCl Area

2E-12 1034.6 439.5 567.2 1706.4

2E-12 1062 441.3 592.4 1761

5E-12 2648.6 1095.8 1472.3 4367.1

5E-12 2652.4 1093.8 1476.4 4376.9

5E-12 2722.3 1203.9 1293.5 4348.4

5E-12 2724.1 1202 1294.2 4356.4

1E-11 5429.8 2393.6 2581.9 8662.1

1E-11 5432.8 2395.3 2584.4 8664.6

4E-12 1826.2 780.1 1125 3116

4E-12 1827.9 780 1127.2 3121.4

6E-12 2734.1 1165.7 1685.3 4671.2

6E-12 2736.9 1167.5 1688.1 4674.3

4E-12 1445.4 783.4 1044.1 3420.7

4E-12 1445.5 784 1044.3 3425.2

5E-12 1805.8 981 1305.3 4281.4

5E-12 1806.4 981 1307.4 4286.3

Mixed-Mode Chromatography to Mitigate Diluent-Eluent Mismatch