Issue PDF

Click the title above to open the LCGC Europe October 2018 regular issue in an interactive PDF format.

Vol 31 No 10 LCGC Europe October 2018 Regular Issue PDF

	Glycosylation is a critical quality attribute (CQA) that can impact on product safety and efficacy of protein biopharmaceuticals. Characterization of N-glycans is therefore of paramount importance for the pharmaceutical industry. Hydrophilic interaction liquid chromatography (HILIC) combined with fluorescence detection (FLD) and 2-aminobenzamide (2-AB) labelling is the golden standard for the analysis of N-glycans enzymatically liberated from biopharmaceuticals. However, for phosphorylated N-glycans, that is, those attached on lysosomal enzymes, irreproducible data and recovery issues are observed on conventional liquid chromatography (LC) instrumentation and columns, which can be attributed to the interaction of the phosphate moieties with stainless steel components in the flow path. This article demonstrates the analysis of phosphorylated glycans with full recovery on a bio-inert LC system and PEEK-lined HILIC column.

	Protein biopharmaceuticals are a rapidly growing class of therapeutics that are widely used for the treatment of various lifeâthreatening diseases (1,2). Around 40% of protein biopharmaceuticals are glycosylated and total glycan mass can range from 2–3% for monoclonal antibodies (mAbs) and up to 50% for erythropoietin (EPO). Glycosylation can impact on product safety and efficacy and is therefore considered a critical quality attribute (CQA). For example, terminal sialic acid residues on complex N-glycans regulate the half-life of EPO in the blood stream, core fucosylation influences the effector function of mAbs, and mannose-6-phosphate moieties on therapeutic enzymes are essential for trafficking to the lysosomes where the enzymes need to be catalytically active (1,2). Characterizing these glycan structures is therefore an essential requirement. Glycosylation analysis of protein biopharmaceuticals can be performed at different levels, that is, at protein, peptide, and glycan levels (3).

	Hydrophilic interaction liquid chromatography (HILIC) combined with fluorescence detection (FLD) and 2-aminobenzamide (2-AB) labelling is the gold standard for analysis of N-glycans enzymatically liberated from biopharmaceuticals (1,2,3). While this method has a proven track record on neutral and sialylated N-glycans, irreproducible data with poor recovery is often encountered upon analyzing phosphorylated N-glycans originating from therapeutic lysosomal enzymes. This is attributed to the interaction of the phosphate groups with stainless steel components in the flow path. This phenomenon has also been observed for nucleotides and phosphopeptides (4–9). For N-glycans analysis, this issue is often alleviated by using mobile phases containing an ion-pairing reagent such as triethylamine (10,11) or by performing separations at high pH (pH > 12) using high-performance anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) (12,13).

	This article demonstrates that phosphorylated glycans can be successfully analyzed by HILIC with the standard 2-AB-labelling method using the commonly applied mobile phase composition, that is, 100-mM NH4-formate pH 4.5 and acetonitrile, when the entire sample flow path is metal-free. To achieve this, the use of a bio-inert liquid chromatography (LC) system as well as a PEEK-lined HILIC column is mandatory.

Acetonitrile and water were obtained from Biosolve. Ammonium formate was purchased from Sigma Aldrich. 2-AB-labelled RNase B and recombinant human acid alpha-glycosidase N-glycans (uncapped and capped) were provided by a local biotechnology company.

 HILIC–FLD measurements were performed on an Agilent 1260 Infinity II Bio-Inert LC system equipped with a quaternary pump (G5654A), a multisampler (G5668A), a multicolumn thermostat (G7116A) with heat exchanger, and fluorescence detector (G7121B) with bio-inert flow cell. For LC–mass spectrometry (MS) experiments the above LC configuration was hyphenated to an Agilent QTOF LC–MS system (G6545A) equipped with JetStream electrospray ionization (ESI) source. LC and LC–MS data were acquired and analyzed using OpenLAB CDS version 2.1, MassHunter for instrument control (B06.01) and MassHunter for data analysis (B07.00) (Agilent Technologies), respectively.

	Two columns were used in this study: a 2.1 mm × 100 mm, 2.7âµm AdvanceBio Glycan Mapping column (superficially porous particles-amide) (Agilent) in regular stainless steel housing and an equally dimensioned PEEKâlined column custom packed with identical superficially porous particles (Agilent). Elution was carried out with a linear gradient of (A) 100âmM NH

-formate pH 4.5 and (B) acetonitrile from 80% to 60%B in 38 min. The flow was set at 0.4 mL/min, the column temperature at 40 °C, and the injection volume was 1 µL. Excitation and emission wavelengths of the FLD were 260 nm and 430 nm, respectively. The operational parameters for the QTOF source were drying gas temperature: 300°C, drying gas flow: 8 L/min, nebulizer pressure: 35 psig, sheath gas temperature: 350 °C, sheath gas flow: 8 L/min, nozzle voltage: 1000 V, capillary voltage: 3500 V, and fragmentor voltage: 150 V. QTOF data were collected in centroid mode at a rate of 1 spectrum per second and acquisition range was 500–3200 

. The system was operated in the extended dynamic range mode (2 GHz).

	HILIC with FLD and 2-aminobenzamide (2-AB) labelling is currently the method of choice for the analysis of neutral and sialylated N-glycans liberated from biopharmaceuticals (1,2,3). Figure 1 shows the analysis of 2-AB-labelled neutral high mannose RNase B N-glycans on a stainless steel and PEEK-lined amide HILIC column installed on a bio-inert LC system. Very similar chromatograms are obtained on both columns for these neutral glycans. 2-AB-labelled glycans elute based on their polymerization degree, that is, the higher the number of glycosidic bonds, the more retention. Moreover, sufficient selectivity differences exist to separate compounds with the same polymerization, that is, M7 isomers which differ in the positioning of the mannose residue on the glycan tree. Both profiles were obtained on the bio-inert LC system, but similar results could be produced on a conventional stainless steel HPLC system.

The HILIC–FLD and HILIC–ESIâMS analysis of 2-AB-neutral and phosphorylated high mannose N-glycans originating from human acid alpha-glucosidase recombinantly expressed in yeast is shown in Figures 2 and 3, respectively. Human acid alpha-glucosidase catalyzes the hydrolysis of glycogen to glucose in the lysosomal compartment of the cell. Around 50,000 people worldwide have a deficiency of this enzyme, which leads to glycogen accumulation in the lysosomes causing muscle damage. These patients typically receive an enzyme replacement therapy with recombinant human acid alpha glucosidase (14). The presence of mannose-6-phosphate containing glycans is a CQA since these functionalities are responsible for targeting the enzyme to the lysosomes where it needs to break down glycogen (15).

	In contrast to the observations on the neutral RNase B N-glycans, an enormous discrepancy is noticed when analyzing the 2-AB-labelled glycans on a stainless steel or on a PEEK-lined column. The neutral glycans M2-8 behave equally well on both columns. The monophosphorylated glycans M2-8P show tailing peaks on the stainless steel column while biphosphorylated glycans M5-8P2 are not recovered at all from the latter column as a result of interaction between metal ions and phosphate moieties. The same recovery principle is known to apply for nucleotides, that is, AMP > ADP > ATP. On the PEEK-lined column, perfect Gaussian shaped peaks are observed for both mono- (M6P, for example) and biphosphorylated (M6P2, for example) structures showing the importance of removing all metal parts from the flow path (instrument, column including frits). Isomeric monophosphorylated glycans are observed as well, which differ in the positioning of the phosphate group on the α1-3 or α1-6 branch of the glycan tree. This is further supported by the extracted ion chromatograms of M6, M6P, and M6P2 presented in Figure 4. Some relevant MS/MS spectra are shown in Figure 5.

It is important to note that the above mentioned phenomena related to adsorption of phosphorylated species on metal ions of the LC system and column only occurs for the phosphate mono-ester containing N-glycans (PO

- and not for the phosphate di-ester containing N-glycans (M-PO

-. For the latter glycans, the profiles obtained on the stainless steel and PEEKâlined columns are exactly the same (Figure 6). Interestingly, these phosphate di-ester-carrying glycans are present on human acid alphaâglucosidase recombinantly produced in Yarrowia lipolytica, but the outer mannose residues are enzymatically removed during downstream processing to generate the active form of the enzyme, that is, carrying mannose-6-phosphate residues (15).


	
	The HILIC analysis of phosphorylated N-glycans released from therapeutic enzymes using a bio-inert LC and PEEK-lined column has been described. It was demonstrated that these challenging structures can successfully be analyzed when the entire flow path is devoid of metal parts, that is, instrument and column inertness.

	The authors acknowledge Sonja Schneider, Sonja Krieger, Linda Lloyd, and Udo Huber (Agilent Technologies).

		K. Sandra, I. Vandenheede, and P. Sandra, 

		S. Fekete, D. Guillarme, P. Sandra, and K. Sandra, 

		V. D’Atri, E. Dumont, I. Vandenheede, D. Guillarme, P. Sandra, and K. Sandra, 

		A. Wakamatsu, K. Morimoto, M. Shimizu, and S. Kudoh 

		S. Liu, C. Zhang, J.L. Campbell, H. Zhang, K.K.-C. Yeung, V.K.M. Han, and G.A. Lajoie, 

		R. Tuytten, F. Lemiere, E. Witters, W. Van Dongen, H. Slegers, R.P. Newton, H. Van Onckelen, and E.L. Esmans, 

		Y. Asakawa, N. Tokida, C. Ozawa, M. Osamu, M. Ishiba, O. Tagaya, and N.Asakawa, 

		H. Sakamaki, T. Uchida, L.W. Lim, and T. Takeuchi, 

		J.Y. Kang, O. Kwon, J.Y. Gil, and D.B. Oh, 

, J.P. Kamerling, Ed. (Elsevier, Amsterdam, The Netherlands, 2007), pp. 303–327.	

		S. Kandzia, N. Grammel, E. Grabenhorst, and H.S. Conradt, in Cells and Culture, ESACT Proceedings, T. Noll, Ed. (Springer, Dordrecht, The Netherlands, 2010), pp. 867–873.	

		P. Tiels, E. Baranova, K. Piens, C.D. Visscher, G. Pynaert, W. Nerinckx, J. Stout, F. Fudalej, P. Hulpiau, S. Tännler, S. Geusens, A. Van Hecke, A. Valevska, W. Vervecken, H. Remaut, and N. Callewaert, 

 is the editor of “Biopharmaceutical Perspectives”. He is the Scientific Director at the Research Institute for Chromatography (RIC, Kortrijk, Belgium) and R&D Director at anaRIC biologics (Ghent, Belgium). He is also a member of LCGC Europe’s editorial advisory board. Direct correspondence about this column to the editor-in-chief, Alasdair Matheson, at alasdair.matheson@ubm.com

 is President of RIC and Emeritus Professor at Ghent University (Ghent, Belgium).

Biopharmaceuticals Perspectives

Glycosylation is a critical quality attribute (CQA) that can impact on product safety and efficacy of protein biopharmaceuticals. Characterization of N-glycans is therefore of paramount importance for the pharmaceutical industry. Hydrophilic interaction liquid chromatography (HILIC) combined with fluorescence detection (FLD) and 2-aminobenzamide (2-AB) labelling is the golden standard for the analysis of N-glycans enzymatically liberated from biopharmaceuticals. However, for phosphorylated N-glycans, that is, those attached on lysosomal enzymes, irreproducible data and recovery issues are observed on conventional liquid chromatography (LC) instrumentation and columns, which can be attributed to the interaction of the phosphate moieties with stainless steel components in the flow path. This article demonstrates the analysis of phosphorylated glycans with full recovery on a bio-inert LC system and PEEK-lined HILIC column.

Analyzing Phosphorylated N-Glycans with Full Recovery on Bio-Inert LC Systems and  PEEK-Lined HILIC Columns

This instalment provides an overview of the modern trends and best practices in mobile-phase selection for reversed-phase liquid chromatography (LC). In particular, we focus on selection criteria and rationales for enhancing analytical performance and ease of preparation.

	Reversed-phase liquid chromatography (LC) is the dominant mode in high performance liquid chromatography (HPLC) for quantitative analysis and is used in ~80% of all HPLC applications (1–3). Reversed-phase LC uses a hydrophobic stationary phase and a polar mobile phase and analytes are retained primarily by hydrophobic interaction. Most analytical methods for purity assessments, quality control, and stability testing of pharmaceuticals employ reversedâphase LC with UV detection because of the excellent precision and reliability of the technique. Another reason for using reversed-phase LC in stabilityâindicating analyses is from a mass balance consideration. The retention in reversed-phase LC is highly correlated to solutes’ log P or partitioning coefficients between water and octanol. Here, the weaker interactive forces of the solutes with the stationary phase give assurance that all impurities within the injected sample are eluted from the column at the end of the purging gradient, thus accounting for all the components in the sample (4).

	The role of the mobile phase in controlling retention and selectivity in reversed-phase LC for neutral and ionizable solutes has been discussed extensively in textbooks, and the reader is referred to them for a detailed discussion (1–3). Retention in reversedâphase LC can be predicted by the linear solvent strength model (5) in which the logarithm of the retention factor of the solute is inversely proportional to the percentage of the strong solvent. However, the model is complicated by interactions between solutes and the surrounding polar solvent molecules in the mobile phase, termed the solvophobic interactions model by Horvath and associates (6), and the role played by the absorbed monolayers of the strong solvents to the hydrophobic stationary phase (1). Also, there can be secondary interactions, such as those between basic functional groups undergoing ionic strengthâdependent interactions with free residual silanols present on stationary phase surfaces, which can result in tailed peaks (7).

	In this instalment, we present a brief overview of mobile phase fundamentals in reversed-phase LC and describe modern trends in their selection, as summarized in Table 1. The highlighted trends focus on the stability-indicating analysis of pharmaceuticals, including both small molecule drugs and biotherapeutics, increased UV and mass spectrometer (MS) detection sensitivity, and the attainment of symmetrical peak shapes for basic analytes.

	In this section, we describe common selection criteria for the weak (aqueous) and strong (organic) mobile phases in reversed-phase LC and their additives.

Overall Trends Towards Simpler Mobile Phases: 

Although the column is the heart of a liquid chromatograph, the mobile phase can be used to manipulate the retention and selectivity of the separation. The column is packed with a high surface area support of variable porosity onto which the stationary phase is bonded. The stationary phase provides retention, and, in conjunction with the mobile phase, differential migration of the solutes. During the method development process, separation of coeluted peaks can be fine-tuned with infinite combinations of mobile phase factors, including solvent type, pH, additives, and operating conditions (such as temperature, flow rate, gradient time).

	Paradoxically, innovation in reversed-phase LC centres on new instruments and columns while scant research efforts are placed on the development of newer reagents for mobile phases. Looking at the historical practice of HPLC, there appears to be a trend towards the use of simpler mobile phases for several practical reasons.

	First, improved column technologies have reduced the need for mobileâphase additives or buffers to improve peak shapes or column batch-to-batch reproducibility (8). Second, the use of simpler binary solvents with linear gradient segments increases method robustness through reduction of method transfer issues for regulated assays (9). Finally, the rapid emergence of HPLC–MS as a standard technology in highâthroughput screening, in-process control, bioscience research, and clinical diagnostics also favours the use of simpler HPLC methods and mobile phases (10).

Selection of Organic Solvent or Mobile Phase B in ReversedâPhase LC: 

Mobile phase B is the conventional term for the strong mobile phase (organic solvent) in reversedâphase LC used in pump blending and gradient elution. Similarly, mobile phase A is the weaker, aqueous mobile phase.

	Historically, the three most common reversed-phase LC organic solvent choices are acetonitrile, methanol, and tetrahydrofuran. The order of eluotropic strengths are methanol<acetonitrile<tetrahydrofuran (1,2). For instance, a mobile phase of 44% methanol–water was found to have equivalent elution strength of 35% acetonitrile–water or 28% tetrahydrofuran–water for a reference application (11). These three solvents have distinctively different properties in terms of proton acceptor ability, proton donor ability, and dipole interactions (1,11). The selectivity differences of these three solvents in reversed-phase LC can be exploited effectively to develop optimized isocratic separations, as demonstrated by Glajch and associates with the aid of modelling software (12).

	Most practitioners prefer acetonitrile because of its strong eluotropic strength, low viscosity (0.37 cP) leading to higher column efficiency, and good UV transparency (to 190 nm). Acetonitrile is an aprotic solvent and is a proton acceptor with capability for  π	

	Methanol is a protic solvent and functions as a proton donor or proton acceptor. Methanol is less expensive but yields significantly higher pressure (viscosity of 0.55 cP) than acetonitrile particularly when mixed with water (for example, solution of 50% methanol–water has a viscosity of 1.62 cP) (11). Methanol has substantial UV endâabsorbance below 210 nm.

	Tetrahydrofuran is rarely used in reversed-phase LC, despite its strong solubilizing power and eluotropic strength. Toxicity and safety issues as a result of peroxide formation are problems preventing its widespread use except in gel permeation chromatography (GPC). Thus, the choice of mobile phase B is effectively narrowed to acetonitrile and methanol.

	Methyl tert-butyl ether can be a substitute for tetrahydrofuran in low concentrations because it has limited miscibility with water and does not form peroxides. Longer chain aliphatic alcohols such as ethanol, propanol, and butanol are not generally used because of their higher viscosities, with notable exceptions in separations of chiral compounds by normalâphase chromatography or in reversedâphase LC of larger proteins. Many publications have appeared in recent years using acetonitrile mixed with 

-propanol, or even n-butanol, which seems to be of benefit to improve mass recovery of some monoclonal antibodies (mAbs). Dimethyl sulfoxide has strong solubilizing power as a diluent but suffers from a very high viscosity and UV cutoffs (3), as well as incompatibility with specific polymeric materials (such as PEEK) used in some HPLC instruments.

The Use of Mobile Phase B Containing Water and Additives:

 One often sees HPLC methods with a stipulated mobile phase B such as 95% acetonitrile in water. The rationale is to help mixing efficiency by making the two mobile phases more similar in viscosity and surface tension. The cons are the reduction of solvent strength of mobile phase B and an extra step in its preparation. With the improvements of modern pumps and online mixers, this additional step appears to be less beneficial.

	There is also a common practice to use the same levels of additives in both mobile phases A and B, such as 0.1% trifluoroacetic acid in acetonitrile when using 0.1% trifluoroacetic acid (TFA) in water as mobile phase A. Technically, there are no perceivable differences between 100% acetonitrile or 0.1% trifluoroacetic acid in acetonitrile for separation reproducibility or elution order in reversed-phase LC except for reducing gradient shifts with UV detection at low wavelengths (to be covered in a later section).

Mobile Phase A: pH Modifiers and Buffers: 

As noted above, the weaker mobile phase in reversedâphase LC is termed mobile phase A which consists largely of water, often containing small amounts of modifiers, buffers, or salts for controlling pH and ionic strength. Pure water can be used for the separation of neutral molecules.

	In pharmaceutical analysis, most drugs are ionizable, that is, acidic, basic, or zwitterionic. Thus, the pH of mobile phase A must be controlled, as it can have a dramatic effect on solute retention. Ionizable solutes can exist in ionized or nonionized forms depending on the mobile phase pH, and the ionized forms have significantly lower retention than the non-ionized forms in reversed-phase LC.

	Table 2 lists the common mobile phase additives with their respective p

 and UV cutoffs. Volatile buffers labelled with an asterisk are used for LC–MS applications. The additives can be an acidifying or basifying agent or a buffer by the addition of its respective conjugate salt.

 An acidic pH of 2–4 is used for most pharmaceutical applications. The low pH suppresses the ionization of weakly acidic analytes, leading to higher retention (3). Basic analytes are ionized at low pH, though most basic drugs have sufficient hydrophobicity for adequate reversed-phase LC retention. An acidic pH also suppresses the ionization of acidic residual silanols to Si-O-ions, which can cause secondary interactions with basic analytes.

	Commonly used acids are trifluoroacetic acid, formic acid, and acetic acid at concentrations of 0.05 to 0.1% 

. In aqueous solutions, the pH values are: 0.1% trifluoroacetic acid (2.1), 0.1% formic acid (2.8), and 0.1% acetic acid (3.2). These simple mobile phases at 0.1% 

 are prepared by pipetting 1.0 mL of the acid into 1 L of purified water and can be used directly without further filtration. They are often used in LC–MS applications though their low ionic strengths may yield poor peak shapes for very basic drugs (7,13).

	The use of 0.1% phosphoric acid appears to be under-utilized because there is a traditional preference for phosphate buffer. It is readily prepared and is transparent down to 200 nm. This additive is useful for purity methods of raw materials or reagents using more universal detection at low UV.

 Buffers are required to tightly control the pH of mobile phase A for critical assays. According to the Henderson-Hasselbalch equation, buffers are most effective within ±1.0 pH units from their p

 values (1,11). A buffer is prepared by mixing a weak acid with the salt of its conjugate base (or a weak base with the salt of its conjugate acid).

	Historically, the most common buffer used in HPLC is phosphate. Because phosphoric acid has three ionizable hydrogens, phosphate buffers make effective buffer systems at pH values around 2, 7, and 10, respectively. It is UV-transparent to 200 nm, but it is not volatile, and thereby not MS-compatible. It also suffers from poor solubility in acetonitrile, particularly at high concentration (such as 50 mM) causing precipitation problems during pump blending. Changing mobile phase B to 85% acetonitrile in water or methanol can alleviate this issue.

	Buffers of ammonium salts of volatile acids are used for the development of MS-compatible HPLC methods. A common MS-compatible buffer system is 20 mM ammonium formate adjusted to pH 3.7 with formic acid. This mobile phase appears to yield excellent peak shapes for most basic drugs and peptides on modern columns, and appears to help retention of many basic or zwitterionic analytes, particularly as sample load increases (7,14). Higher concentrations can be used effectively to increase retention in reversed-phase LC by the “salting out” effect or the Le Chatelier’s principle to drive solutes into the hydrophobic stationary phase (1,11).

 Before the 1990s, silica-based columns could not be used with high pH mobile phases, because of the dissolution of silica support at pH values >8. The development of hybrid particles and improved bonding chemistries for silica particles has extended the usable pH range of these specific columns to a range of 1–12 (2–10 at column temperatures >40 °C) (3,15–16). Typical high pH mobile phases are 0.05–0.1% ammonia, and 10–25 mM ammonium carbonate, or phosphate buffers. Comparisons of mobile-phase buffers at elevated pH conditions revealed significant effects on silica solubility and expectations for stability under high pH and temperature conditions (16). Important observations were that ammonium hydroxide is perhaps the safest choice and has good compatibility with ESIâMS detection. The advantages of using high pH separations include higher retention of water-soluble bases (for example, for the analysis of opioid drugs, which traditionally require ion-pairing chromatography), better peak shapes for basic analytes, and MS-compatibility (15). The disadvantages are the instability of some drugs in basic mobile phases and potential method robustness issues if the mobile phase pH is close to the p

 values of the basic analytes (for example, most amines have pKa values in the range of 8–10) (3,15).

	Figure 1 shows a chromatogram of a test mixture containing two active pharmaceutical ingredients (APIs) spiked with process impurities and degradants for an MS-compatible stability-indicating assay of a combinational drug product (3). Note the excellent peak shapes of all peaks under high-pH conditions without any peak splitting. The disadvantage of this method is its sensitivity to mobile phase pH, which must be kept between 9.10 to 9.15 to prevent any peak coelution (3,15).

	Figure 2 further illustrates the effect of mobile-phase pH to dramatically alter the retention and elution orders of acidic and basic drugs (17). The use of column screening using mobile phase A at both acidic and basic pH has become a common practice in many laboratories for high-throughput screening and development of critical methods (3, 8–10).

Reduced Usage of SilanolâMasking, Ion-Pairing and Chaotropic Reagents:

 Prior to the 1990s, the primary complaint against silica-based HPLC columns was the difficulty in method transfer for quality control applications caused by batch-to-batch differences of raw silica support or bonded phase chemistry. The main cause of this non-reproducibility was the presence of active silanols, which are activated by metallic impurities (1,3). At that time, many bonded phases had such acidic silanols that basic solutes would not elute without the use of silanol-masking reagent additives in the mobile phase (such as triethylamine). Triethylamine was frequently used in drug analysis to yield better peak shapes and acceptable batch-toâbatch reproducibility in regulated testing (18).

	Today, the popularity of high-purity silica supports, and the elimination of metallic impurities in the base silica, leads to reduced silanophilic activities and much improved batchâto-batch reproducibility of bonded phases. The use of triethylamine as an additive has generally been abandoned for several reasons. First, triethylamine may permanently alter the selectivity of the column. Second, it would overwhelm all MS signals in the positive ionization mode. Third, new column technologies using charged surface and other approaches provide a better solution to improve peak shape of basic analytes.

	A fairly recent innovation has been the bonding of a positive charge (for example, an amine) to the surface of bonded phases (reviewed in Nawrocki, 30), such as the case of charged surface hybrids (CSH) introduced by Waters in 2010 (13). These columns yield excellent peak shapes for highly basic drugs even in mobile phases with low ionic strengths (such as 0.1% formic acid) as illustrated in Figures 3(a) and 3(b) showing comparative chromatograms of a test mix of 12 new chemical entities (NCEs) using a low ionic strength mobile phase A of 0.1% formic acid on columns packed with a CSH compared to those on a hybrid bonded phase (19,20). The CSH C18+ column shows excellent peak shapes for all NCEs while the other column displays considerable peak tailing for four of the more basic NCEs.

	Another approach is to use an advanced coating technique to reduce the activities of the residual silanols. An added benefit of this approach is to yield columns usable at high pH (such as the Agilent Poroshell HPH-C18) (20).

 Ion-pairing reagents are detergent-like molecules added to mobile phase A to provide retention of acidic or basic analytes (1). Long-chain alkyl sulfonates (C5 to C12) combine with basic solutes under acidic pH conditions to form neutral “ion-pairs” that are retained in reversed-phase LC. Retention is proportional to the length of the hydrophobic chain of the ionâpairing agent and its concentration of the ion-pairing reagent. Note that trifluoroacetic acid has some ion-pairing capability and is commonly used in reversedâphase LC of proteins and peptides. The use of ion-pairing reagents has decreased somewhat in recent years, because of issues such as lower column efficiencies, slow column equilibration, difficulties in performing gradient analysis, and the lack of MS-compatibility (1,2). Also, several companies have produced specialty surface-modified HPLC columns that reduce the need for ion-pairing reagents, by resolving the problem of “phase collapse” or “pore dewetting” at low or no organic solvent reversed-phase LC mobile phase conditions (3). Such materials can retain highly polar bases and acids, with less reliance on hydrophobic counterion additives, ensuring better compatibility with MS detection, and faster re-equilibration between separations.

	Another class of additives used to increase retention and selectivity of basic analytes is inorganic chaotropic agents (such as PF6-, BF4-, ClO4- ions) that form neutral ion pairs under acidic conditions in reversed-phase LC (1,2). Chaotropes are better suited for gradient analysis and can produce lesser baseline shifts, but are not MS-compatible.

	Note that availability of highâpHâcompatible volatile mobile phases (such as ammonium hydroxide), with suitable silicaâbased or hybrid columns, can offer a compelling alternative MS-compatible approach for the analysis of waterâsoluble basic drugs, and thereby reduce the need for ionâpairing or chaotropic reagents for such separations (15).

Other Mobile Phase Additives for Chiral Separations and  Ion Chromatography:

 It is possible to use a chiral additive as a pairing reagent to form a diastereomer, enabling chiral separations to be achieved on conventional reversedâphase columns (1). However, the availability of versatile and highly efficient chiral stationary phases (such as immobilized and chemically-modified polysaccharides) has reduced the need for these indirect methods for chiral separations (1–3).

	Ion chromatography sometimes employs a mobile phase additive (such as 3 mM sodium phthalate) to perform ion chromatography by conventional HPLC–UV equipment with indirect photometric detection of the displaced complex (21). However, the widespread availability of improved ion chromatography instruments with suppressed conductivity or MS detection has eliminated most of these approaches with limited applicability.

	The Use of Balanced Absorbance Mobile Phases: During gradient analysis using mobile phases with UV detection at low wavelengths (such as <230 nm using MS-compatible mobile phases), substantial gradient baseline shifts are observed as a result of an imbalance of absorbance and refractive indices of mobile phases A and B. The magnitude and direction of the baseline shift are strongly dependent on detector flow cell design, wavelength, and mobile phase composition. These imbalances can also be attributed to the chromophoric shifts of the UV spectra of the additives (such as carboxylic acids) in the two very different solvent environments (3). The differences in the apparent absorbance (absorbance detector signal) and the inadequate mixing of mobile phase A and B solvents can often also be observed as a periodic (sinusoidal) baseline disturbance. Again, the magnitude of such baseline disturbances is a complex function of the online mixer design, volume, and apparent absorbance differences between pure mobile phase A and mobile phase B.

	Figure 4(a) illustrates the gradient baseline shift observed during method development of a stability-indicating method for an NCE using mobile phase A (0.1% trifluoroacetic acid) in water and mobile phase B (0.1% trifluoroacetic acid in acetonitrile) (3). Because of the absorbance of trifluoroacetic acid in the far UV region (200–230 nm), significant baseline shift (~0.1 AU at 230 nm) was observed. These baseline problems were substantially reduced by lowering the concentration of trifluoroacetic acid to 0.05% (see the UV spectrum in Figure 4[b]). Further reduction of the gradient shift to 0.002 AU at 230 nm was achieved by adjusting the concentration of trifluoroacetic acid in mobile phase B to 0.03%, which has a similar apparent absorbance at 230 nm with 0.05% trifluoroacetic acid in water (Figure 4[c]).

	When using UV–vis absorbance detectors, suitable adjustments of baseline properties, as described above, are required to achieve the highest levels of UV sensitivity.

Modern Trends in Mobile Phase Preparation

	Modern trends in mobile phase preparation include the use of lower buffer concentrations, the elimination of the filtration process, and pH adjustment procedure for buffers.

Lower Buffer Concentrations in Mobile Phase A: 

Although 50 mM buffers are specified in many older methods, the modern trend is to use lower buffer strengths (such as 5–20 mM), which provide sufficient buffering capacity in most applications (3). Lately, the use of modern columns with lower residual acidic silanol activities, or CSH materials, has allowed the use of simplified mobile phase A with low ionic strength, such as 0.1% formic acid, while permitting good peak shapes for highly basic analytes (11,13,20)

Many laboratories have eliminated the filtration process with 0.2- or 0.5-µm membrane filters by using high-purity reagents (such as 99.995% ammonium formate from Aldrich, Cat# 516961), and HPLCâgrade solvents and water (22). The internal filters in most HPLC pumps (replaced during preventive maintenance programs) appear to be adequate to allow successful routine HPLC operation without filtering in most laboratories. This elimination of the filtration step can reduce potential mobile phase contaminations from the filtration process. Exceptions are mobile phases containing ion-pairing reagents and high-concentrations of salts, or when high-purity reagents are not readily available.

	Similarly, the use of vacuum filtration for dissolved gas reduction in mobile phases has been largely superseded by inline vacuum degassers, which are standard on most modern HPLC systems.

 Many analysts prepare buffers by dipping the pH electrode directly in mobile phase A and titrating the solution until the designated pH is reached. This practice can be problematic because trace contaminants adhering to the electrodes (such as preservatives from the pH calibration buffers to allow room temperature storage) can cause substantial ghost peaks in gradient elution (see the example in Figure 5). To prevent such contamination, the analyst should dispense a small amount of mobile phase A into a vial (such as a scintillation vial) to check the pH and continue this iteration until the correct pH is reached. The content of the vial should not be poured back into mobile phase A (3). With suitable care, the correct composition of mobile phase A can be maintained, with minimal risk of introducing contaminants that can readily appear in high sensitivity analyses.

	Modern trends in gradient methods include the use of simple binary mobile phases (such as 0.1% formic acid or 0.1% ammonia) with highâefficiency columns packed with superficially porous particles and low-silanophilic bonded phases (8). Robust stability-indicating methods for very complex pharmaceuticals can be accomplished by using multisegment binary linear gradients (23). Although quaternary pumps are widely used for method development, quality control methods that use ternary or quaternary mobile phases are rarely encountered, because of the difficulties with method reproducibility across different laboratories or equipment platforms. For that matter, the use of concave or convex gradients is strongly discouraged, even though they are available in some systems. Method transfer can be less problematic with simplified methods using easily prepared binary mobile phases.

Comments on the Stability of Mobile Phases

	There are no universal guidelines or scientific data on mobile phase stability after preparation. Obviously, fresh mobile phases can be prepared daily as needed, though this time-consuming process is neither a necessary or “green” practice. The shelf life of the mobile phase is difficult to generalize and is dependent on mobile phase composition, pH, storage container, storage conditions, and the sensitivity of the separation to small changes in mobile-phase composition. The potentials for growth of microorganisms (bacteria, algae, molds) poses the greatest risks to system contamination and column longevity. The use of a scavenger column (for example, C18, 33 × 4.6 mm, 10-µm) (24) can provide excellent protection against mobile-phase-borne particulates and contaminants. However, it does add system dwell volume and is rarely used in today’s laboratories.

	A common practice is to date all mobile phases and use aqueous buffers for one week, simple acidified water (such as 0.1% formic acid or trifluoroacetic acid) for 1–2 months, and organic solvents for at least 3 months. For weakly acidic buffered mobile phases or buffers at near neutral pH, storage of a concentrate (10–50×) can be useful. Depending on the specific buffer, refrigeration (5 °C) can permit the use of the concentrate for many months. This approach has worked well for acetate and formate buffers. In such cases, periodic checks of the pH and blank injections can verify the integrity and cleanliness of the mobile phase preparations. There may be standard operating procedures in some laboratories that stipulate reagent shelf life based on historical data. In any case, labelling of the preparation date is encouraged and required in many regulated laboratories.

Newer Additives for ReversedâPhase LC of Proteins and Peptides

	Separation of proteins and peptides is a challenging application for HPLC requiring bonded phases with low silanophilic activity, specific mobileâphase additives to allow detection at low UV wavelengths (210–220 nm), and higher column temperatures (>60 °C) to improve peak shapes, and mass recovery (2,3).

	The traditional mobile phase A for protein and peptide separations is 0.1% trifluoroacetic acid, which has ion-pairing capability for better peak shapes and offers reasonable UV sensitivity at 210–220 nm for peptide bonds. Although trifluoroacetic acid is considered MS-compatible from a volatility standpoint, MS sensitivity is seriously compromised as a result of ion suppression when trifluoroacetic acid is used during the electrospray ionization (ESI) process. Trifluoroacetic acid is an excellent ion pair reagent with primary Ïµ-amino lysine group, as well as the N-terminal α-amino group. In addition to supplying a strongly acidic modifier for reversed-phase LC separations, trifluoroacetic acid shows a favourable increment in retention of acidic peptides, because of the formation of an N-terminal amino ion pair, which is notably absent in formic acid mobile phases. Nevertheless, MS sensitivity will be impaired for protein and peptide detection using trifluoroacetic acid, when compared to formic acid alone, or compared to mixtures of trifluoroacetic acid and formic acid. The latter approach has some popularity, with the use of lowered concentration of trifluoroacetic acid in mobile phases A and B, in the presence of equal or greater concentration of formic acid to act
	as a competitor to trifluoroacetic acid, during gas phase exchange in the ESI process (25,26).

	Recent research by Boyes and associates has uncovered two promising additives that deliver a good compromise of performance for peak shape and MS sensitivity: difluoroacetic acid and 3-fluoropropionic acid (26). Figure 6 illustrates the comparative performance of four mobile phase additives (formic acid, ammonium formate/formic acid, trifluoroacetic acid, and difluoroacetic acid) in the HPLC–UV–MS analysis of a mixture of five peptides. The HPLC–UV and total ion chromatogram (TIC) data indicate that difluoroacetic acid offers excellent performance in peak shape and sensitivity under UV and MS detection.

	Figure 7 shows comparative chromatograms in a peptide mapping application showing excellent performance in peak shape and MS sensitivity for 10 mM difluoroacetic acid compared to those for formic acid, trifluoroacetic acid, and a mixture of formic acid and ammonium formate. Figure 8 shows an example of a protein separation using trifluoroacetic acid compared to difluoroacetic acid and formic acid in mobile phase A.

	Anecdotal evidence suggests that the advantages of difluoroacetic acid as a replacement are more apparent for ESI signal intensities of peptides and smaller proteins, compared to larger proteins up to the size of immunoglobins. Regardless, difluoroacetic acid is much easier to flush from LC systems than trifluoroacetic acid, which requires special and time-consuming procedures to completely flush out of LC–MS systems. Basically, trifluoroacetic acid appears more “sticky” towards components in the LC system than difluoroacetic acid. Over more than six years of use, difluoroacetic acid has shown no negative effects on instrument performance in standard or capillary column separations.

	Another approach that has emerged is the use of additives to mobile phase A to increase, or rescue, ionization efficiency lost to trifluoroacetic acid ionization suppression. Dimethyl sulfoxide has been suggested as an additive for improving MS detection in high complexity or high sensitivity proteomic sample LC–MS analyses (27,28). Several confirmatory examples are described for proteomics analysis, although some concerns remain about the long-term effects on LC–MS components, and the tendency of dimethyl sulfoxide (or impurities or break-down products) to accumulate in MS systems over time. Various weak “Brønsted” base compounds, shown to act as “supercharging agents”, have been studied as mobile phase A additives to improve MS signal intensity of proteins resolved in reversedâphase LC, using trifluoroacetic acid modified mobile phase A (29). The addition of such compounds does not hinder effective LC performance and has been shown to restore signal intensities heavily suppressed by trifluoroacetic acid. The effect of these additives on MS system stability and maintenance are still unknown.

	In this instalment, we present an overview of modern trends and best practices in mobile-phase selection and preparation in reversed-phase LC, with a focus on pharmaceutical applications and increasing peak shape and detection performance by UV and MS. These trends include the use of simple acidic mobile phase A or low-concentration MS-compatible buffers with acetonitrile or methanol in binary mode, using broad linear gradient or multi-segment gradient methods for complex separation, the elimination of the filtration process, and many less relevant mobile phase additives. Newer additives (difluoroacetic acid and 3-fluoropropionic acid) as alternatives to trifluoroacetic acid for an increased MS sensitivity are suggested.

	The authors thank the following colleagues for their review of the manuscript: Matt Mullaney, Tuty Norashikin of MSD International GMbH, Mike Shifflet of J&J Consumer Healthcare, M. Farooq Wahab of U. Texas at Arlington, Patrick Cronan of Agilent Technologies, Brian R. Holder of Merck, He Meng of Sanofi, and Justin Godinho at AMT.

		L.R. Snyder and J.J. Kirkland, Introduction to Modern Liquid Chromatography (Wiley, Hoboken, New Jersey, USA, 3rd Ed., 2010), Chapters 2, 6, and 7.	

		Y.V. Kazakevich and R. LoBrutto, Eds., HPLC for Pharmaceutical Scientists (Wiley, Hoboken, New Jersey, USA, 2007), Chapter 4.	

		M.W. Dong, Modern HPLC for Practicing Scientists (Wiley, Hoboken, New Jersey, USA, 2006), Chapters 1, 2, 5, 6, 8, and 10.	

		M.W. Dong, LCGC Europe 26(7), 393–399 (2013).	

		L.R. Snyder and J.W. Dolan, HighâPerformance Gradient Elution: The Practical Application of the Linear-Solvent-Strength Model (Wiley, Hoboken, New Jersey, USA, 2006).	

		W.R. Melander and C. Horváth, in HPLC, Advances and Perspectives Vol. 2, C. Horváth, Ed. (Academic Press, New York, USA, 1980).	

		D.V. McCalley, J. Chromatogr. A 1217(6), 858–880 (2010).	

		D. Guillarme and M.W. Dong, Amer. Pharm. Rev. 16(4), 36–43 (2013).	

		M.W. Dong, LCGC Europe 26(8), 455–461 (2013).	

		M. Wong, B. Murphy, J.H. Pease, and M.W. Dong, LCGC Europe 28(6), 343–352 (2015).	

		K. Schug and T. Taylor, The Essential CHROMacademy Guide: Mobile Phase Optimization Strategies for Reversed Phase HPLC, Webcast, http://www.chromacademy.com/essential_guide_webcast/mobile_phase_optimization_strategies_for_reversed_phase_hplc/mobile_phase_optimization_strategies_for_reversed_phase_hplc.pdf (accessed on 27/6/2018).	

		J.L. Glajch, J.J. Kirkland, K.M. Squire, and J.M. Minor, J. Chromatogr. 199, 57 (1980).

		K.J. Fountain, H.B. Hewitson, P.C. Iraneta, and D. Morrison, Waters Corporation, Milford, Massachusetts, USA, 720003720EN, Sep. 2010.	

Cover Story

Modern Trends and Best Practices in Mobile-Phase Selection in Reversed-Phase Chromatography

The use of ultrahigh-pressure liquid chromatography (UHPLC) is now commonplace among pharmaceutical laboratories. However, until depreciation cycles replace traditional high performance liquid chromatography (HPLC) systems that operate at a maximum pressure of 400 bar, the advantages of UHPLC cannot be realized worldwide. Thus, product methods developed using UHPLC capabilities cannot directly transfer these methods to receiving laboratories without qualified UHPLC availability. As scaling methods from traditional LC to UHPLC has been popular in recent years, the desire exists to address this LC limitation by potentially transferring methods back from UHPLC to available LC systems. This capability has not been shown to be effective on traditional LC columns. Today, pharmaceutical chromatographic methods often use superficially porous packed particle columns that enable UHPLCâlike separation under 400 bar. Unlike traditional stationary phases, which use a range of particle sizes with associated column dimensions, superficially porous particle stationary phases often use the same particle size in differing column dimensions. As the particle size now remains the same, this study looks to see if reverse scaling of chromatographic profiles of interest to the pharmaceutical industry can be routinely achieved.

Liquid chromatography (LC) instrumentation operating at greater than 400 bar is becoming more commonplace as newer equipment is making its way into commercial laboratories. Although these ultrahigh-pressure liquid chromatography (UHPLC) systems are becoming more prevalent, the industry still relies on transferring chromatographic methods throughout the world and, as such, the issue that not all receiving laboratories routinely operate with LC systems that are compatible with UHPLC methods still needs to be addressed.

	Since its inception, scaling low-pressure methods to UHPLC conditions has been prevalent. Many companies have scaling calculators available on their websites (1). Scaling of LC is becoming less of an issue as depreciation cycles replace laboratory LC systems with higher pressure capabilities. Yet, the issue of low-pressure systems still exists in pharma when using smaller contract and worldwide laboratories. With proper considerations for flow, injection, and system void volumes, methods developed on <400-bar systems are often made much more efficient using UHPLC conditions (2–4). However, the scaling is not perfect, and the UHPLC method should be revalidated before use. Although resolution may improve, the required lower injection volume may not place enough analyte on the column to detect low-level impurities. In addition, as demonstrated in our laboratory, columns may not behave identically as scaled (5). We found that scaling from UHPLC back to high performance liquid chromatography (HPLC) is not advantageous for impurity methods because of potential selectivity and performance issues. This observation makes sense; we don’t road test a Formula I automobile to race with a street car. 

	After publication of our reverse-scaling study, columns packed with superficially porous particles (SPPs) have become prevalent for pharmaceutical applications. Columns packed with these particles are often referred to as core–shell columns. Core–shell columns have the advantage of exhibiting near-UHPLC separations but at lower pressures (6). This capability enables laboratories to shorten their separation run times and efficiencies on traditional, nonâUHPLC, instrumentation. Relevant to scaling interests, core–shell columns are often packed with the same particle in differing column dimensions. Unlike in the original reverse-scaling studies, core–shell columns enable the same particle to be used in the differing column dimensions. This characteristic may allow separations to reverse scale more effectively from UHPLC to HPLC conditions and pressures. The original intent of this study was to only look at vertical scaling, but the data showed right off the bat that this approach would not work. The investigation moved to horizontal scaling as a follow-up series of experiments to further investigate core–shell column scaling. In this study, a representative separation profile using commercially available drug substances was used to investigate reverse scaling with SPP columns.

The LC data reported in this study was generated using a Thermo Fisher Scientific Ultimate 3000 system equipped with a photodiodeâarray detector. System volume changes are critical in method scaling. A single instrument platform was used to negate this affect. Using a single system for all column investigations allowed us to operate at both traditional and UHPLC conditions and normalize the system effects and void volume differences that we would have using different systems. Thus, this study focused solely on column differences. The LC system was controlled and the resulting data were processed using Atlas Version 9.00.0.10711 (Thermo Fisher Scientific).

 Mobile phases and LC conditions were consistent with those used in a previous study (11). A single vendor for the traditional phase UHPLC column and two separate vendors for the SPP phases were used. All target analytes and mobile phases were prepared with chemicals purchased from SigmaâAldrich. The LC column dimensions that were used are listed in Table 1.

 Scaling calculations were calculated as described in the manuscript and correlated with spreadsheets received from Grace Discovery Sciences. The use of method translation calculators was recently reviewed by Majors (8).

	Scaling Calculations: Scaling calculators were created for projecting the run conditions moving from traditional HPLC column formats to UHPLC column formats. In this regard, the calculators work well in speeding up run times from traditional HPLC methods and oftentimes improving, or at least maintaining resolution, throughout the profile (2,3). Because many laboratories still use equipment limited to the traditional HPLC platform, laboratories have attempted to scale back from UHPLC to HPLC with the goal of rapid HPLC method development on the UHPLC platform and transferring the method on traditional platforms that are still being used. Webster and Elliott showed why this approach may not be a good idea (5). This issue is limited in scope as depreciation cycles equip more and more laboratories with UHPLCâcompatible instruments. Scaling calculators are designed with the conversion from HPLC to UHPLC in mind. The critical parameters in scaling from UHPLC to HPLC include column length, particle size, and injection volume (5,9–11). Neue and colleagues (12) demonstrated the scaling of separations using 5-μm, 15-cm columns to 1.7-μm, 5-cm columns. Dramatic changes should be made only if the chromatographic profile and resolution of the critical pair is maintained (13).

	Scaling calculations are based on standard chromatographic concepts (14). The primary equations used by the scaling calculators are

 is the injection volume for methods 1 and 2; dc is the column diameter used in methods 1 and 2; and 

is the column length used for the method 1 and 2 columns.

 is the column diameter used in methods 1 and 2; and 

 is the particle size used for the method 1 and 2 columns.

 is the gradient time for methods 1 and 2.

	The calculations are designed to provide starting chromatographic conditions followed by further optimization. The critical parameter in reverse scaling is whether these conditions yield a pressure under the pressure maximum for HPLC. Most traditional HPLC platforms tolerate pressure of up to 400 bar. A UHPLC method that is to be converted to an HPLC platform must first be optimized to run at conditions that project that the new method pressure will remain under 400 bar. This often entails slowing down the optimized UHPLC method (5,14).

 The goal of this study was to evaluate whether columns packed with core–shell particles reverse scale more effectively than traditional column packings. A separation based on earlier work (5) was used for actual UHPLC separation using compounds that vary in their chemical nature. In that study, the UHPLC column in a traditional particle format was used as the basis to investigate scaling back to HPLC.

 Our first attempt is looking at UHPLC to short SPP columns. In Figure 1(a), the rapid resolution of varying drug substances on a traditional UHPLC phase is presented. If we cannot scale a representative profile from traditional UHPLC to core–shell UHPLC phases, going to core shell phases packed in traditional lower pressure column dimensions is not warranted. Using the horizontal scaling conditions from Table 2, the resulting profile for the 2.6-µm Core–shell A phase is shown in Figure 1(b). In this separation, a peak reversal for the carbamazepine and furosemide pair is seen as well as a loss in resolution. Additionally, ibuprofen is coeluted with the indomethacin peak. It is not possible to conclude that scaling yielded an equivalent profile with the Core–shell A column and conditions. Looking at the results obtained using a second core–shell column (from a different manufacturer) in Figure 1(c), the resulting profile again exhibits peak reversal and loss of resolution for the carbamazepine and furosemide peak pair. In addition, a coelution of the ibuprofen and indomethacin peak pair can be seen in the profile. Again, an equivalent profile was not produced. The effect of scaling may not be the culprit here as much as the older, traditional stationary phase (type A silica) likely has more exposed silica. It should be noted that using calculators specifically from the Core–shell A and B vendors for their phases did not solve this issue. Looking into this stationary phase effect a little further, the mobile phase was changed from formic acid to trifluoroacetic acid to investigate whether the addition of an ion-pairing reagent would improve the core–shell profile. Figure 2(a) is the resulting profile adding 0.1% trifluoroacetic acid in place of formic acid used with the traditional phase UHPLC column. Essentially the same profile is produced. However, the profile improves for both SPP phases (Figures 2[b] and 2[c]). The ibuprofen peak resolution with the core shell particles remains unacceptable. It is likely with further optimization that the profile could be obtained on the SPP phases. Method optimization is not direct method scaling of the separation. What this data shows is not so much that scaling cannot be achieved, but in more-complex separations the differences in stationary phase chemistries still remain significant. As with all LC separations, the nature of the stationary phase and the chemistry of the analytes must be considered. The United States Pharmacopeia (USP) may traditionally allow the substitution of one C18 column for another, but this equivalence is seldom realized in industrial pharma (16).

 One of the advantages of superficially porous particles, according to their vendors, is that because the same particle is used in different column dimensions, scaling with these phases should be straightforward. Rather than looking at a horizontal scaling from the UHPLC short column as we did before, now we are changing the optimal core–shell conditions and scaling back to traditional HPLC column dimensions. We are going to start with the separation on the short SPP column and see how it vertically scales back to columns with dimensions compatible with lower pressure systems (Table 3). Figure 3(a) is our initial separation for the Core–shell A phase. Ibuprofen was left in the sample matrix to see if the conditions applied will ultimately resolve this peak from the indomethacin peak. Increasing column dimensions, shown in Figures 3(b)–3(e), effectively maintains the original profile. Interestingly, going down to the 1.7-µm particles did not enhance resolution of the original profile. The resolution data for the separations is presented in Table 4 and the relative retention times (RRT) are listed in Table 5. The RRTs are close, but not acceptably equivalent to say that the profiles scaled equivalently for both the 2.6- and 5-µm particles. From this data, we conclude that the scaling was a good estimate of the conditions needed for the different column geometries. However, there is still enough error in particle composition and packing differences so that direct method scaling was not achieved. 

	The vertical scaling study was repeated with columns from a second core–shell column manufacturer, designated as the Core–shell B phase (Table 6). Again, illustrated in Figure 4(a), an initial resolution profile for this phase was established. The increasing column dimensions seen in Figures 4(b)–4(e) actually improve the profile for the carbamazepine and furosemide peak pair as well as the indomethacin and ibuprofen peak pair to a lesser extent. The resolution and RRT data for this study are presented in Tables 7 and 8. The RRT data are much improved with this series of profiles. It remains difficult to argue that the scaling profile was maintained.

	Column scaling in liquid chromatography is an interesting concept that may work for series of analytes with consistent chemistry. However, the real value in scaling chromatographic methods is attempting to improve challenging lower-pressure separations using UHPLC capability. In our laboratory we have not seen stationary phases scale theoretically with column dimension. Scaling has not proven to be an achievable opportunity for most complex pharmaceutical separations. It is our suggestion to improve chromatographic efficiencies using core–shell and UHPLC technologies in a standalone fashion that can be optimized using scaling principals. However, revalidation is required to establish equivalence because direct scaling is not adequate in the cases studied.

	This project was supported by AbbVie, Inc. AbbVie, Inc. participated in the interpretation of data, writing, reviewing, and approving the publication. Gregory K. Webster is
	an employee of AbbVie and Matthew A. Gragg is a contractor with AbbVie, Inc.

		Sigma, http://www.sigmaaldrich.com/analytical-chromatography/hplc/method-transfer-calculator.html, Sciences Analytiques: https://epgl.unige.ch/labs/fanal/hplc_calculator:en; ACD Labs: http://www.acdlabs.com/resources/freeware/translator/.	

		M. Swartz, “HPLC to UPLC Method Migration: An Overview of Key Considerations and Available Tools,” presented at the Pittsburgh Conference & Exposition on Analytical Chemistry and Applied Spectroscopy (Pittcon), Chicago, Illinois, USA, 2007.	

		V. González-Ruiz, A.I. Olives, and A. Martín, 

		D. Guillarme, D.T.T Nguyen, S. Rudaz, and J.L. Veuthey, 

		M.M. Dittmann, “Approaches Towards Method Compatibility Between HPLC and UHPLC Systems,” Paper 1970-6, presented at the Pittsburgh Conference & Exposition on Analytical Chemistry and Applied Spectroscopy (Pittcon), Orlando, Florida, USA, 2010.

		U.D. Neue, D. McCabe, V. Ramesh, V. H. Pappa, and J. DeMith, 

		G.K. Webster, T.F. Cullen, and L. Kott, Ultra-High Performance Liquid Chromatography and Its Applications, (Wiley Interscience, 2013), pp. 31–54.

 is a senior principal research scientist in Development Sciences at AbbVie, Inc. Dr. Webster received his PhD in analytical chemistry at Northern Illinois University in 1991. Prior to joining Abbott/AbbVie in 2007, he worked at Alpharma, Chemsyn Laboratories, Bayer Corporation, and Pfizer.

 is currently a chemist working within the NCE Analytical R&D – LC group at AbbVie, Inc. He graduated with a BS in Chemistry from Truman State University in 2015.

Features

The use of ultrahigh-pressure liquid chromatography (UHPLC) is now commonplace among pharmaceutical laboratories. However, until depreciation cycles replace traditional high performance liquid chromatography (HPLC) systems that operate at a maximum pressure of 400 bar, the advantages of UHPLC cannot be realized worldwide. Thus, product methods developed using UHPLC capabilities cannot directly transfer these methods to receiving laboratories without qualified UHPLC availability.

Scaling LC Methods Using Superficially Porous Particle Stationary Phases

What types of mixers are needed with different types of liquid chromatography (LC) pumps? Are different mixers needed for different applications?

	With some topics I have written about in the past year, I have tried to draw attention to aspects of our practical work with liquid chromatography (LC) that I think are often underappreciated-for example, the value of using inline mobile phase filters, and best practices around reducing contaminants when using LC–mass spectrometry (MS). There are other topics where the issues I observed in talking with people about their LC work are more about misunderstandings, misconceptions, and myths. This month, I have chosen to tackle a topic of this kind, focused on mobile-phase mixing and mixers. I think there is some confusion among users of LC about what happens when we combine two (or more) different flow streams in LC systems, and when mixers are needed to blend the different fluids. Hopefully, the discussion will enable a more detailed understanding of why they are needed, and when and how one might consider using something other than the default mixer that is used at the time of installation.

Why Do We Need Mixers in LC Pumps in the First Place?

 The vast majority of LC pumps in use today can be approximately characterized by one of two designs, shown schematically in Figure 1. These designs have been discussed in the “LC Troubleshooting” column in the past (1), and there are good educational resources for learning about them elsewhere (see www.chromacademy.com). Readers interested in much more detail will find an extensive discussion of modern pumps by Frank Steiner in “The HPLC Expert II”, edited by Stavros Kromidas (2). For the sake of discussion here, we need to establish that we will discuss the designs at a particular level of detail to appreciate why and when a “mixer” is needed. The design in Figure 1(a) is typically referred to as a high-pressure mixing system, meaning that each of the mobileâphase constituents is first pushed with enough force to meet the inlet pressure of the column, and then these flows are brought together in a fluidic junction (referred to here as the convergence point) upstream from the point where the sample is injected. These systems are usually binary pumps, meaning that they have two independent high-pressure pumps that can deliver two different solvents to the convergence point at high pressure. The design shown in Figure 1(b) is typically referred to as a low-pressure mixing system, meaning that two or more fluids are brought together at the convergence point while at low pressure, and then the single stream of combined fluid is pushed by a single high-pressure pump to meet the inlet pressure of the column. The major advantage of the high pressure mixing design in Figure 1(a) is that the time it takes for a change in solvent composition made by the pump (as in gradient elution) to appear at the column is typically much shorter than it is for low-pressure mixing systems. We refer to this as the gradient delay time, and we’ll discuss this more later on the in the article. On the other hand, the major advantage of the low-pressure mixing system illustrated in Figure 1(b) is that these typically come with a lower purchase price, simply because there is one high-pressure pump involved instead of two. The fact that this type of pump can also produce ternary (three fluids) and even quaternary (four fluids) mobile phases can also be a great asset for some applications (for example, making pH gradients online using the pump).

Imagining the Convergence of Two Fluids in the Pump:

 Now, to the question at hand: Why do we need a mixer? First, it is helpful to define the meaning of mixer in the context of this discussion. Here, I will use the term mixer to describe any device that is deliberately used in a LC system for the purpose of mixing two or more fluids. Of course, in real LC systems there are many elements of the system that promote mixing of the mobile phase (and sample), whether or not this is the intent of that element. For example, some mixing can occur in a straight, open connecting capillary, and a fair amount of mixing happens in all chromatography columns (although we try to prevent this by design!). Now, the ways that different fluid streams converge in a LC system can be very different, depending on the context in which this happens. This month we will consider two scenarios: 1) convergence in a highâpressure mixing pump (Figure 1[a]); and 2) convergence in a low -pressure mixing pump (Figure 1[b]). Next month, we will consider the convergence of sample and mobile phase when a sample is injected into a column. Figure 2 illustrates, albeit in an idealized way, the major differences between these three scenarios in terms of how two fluids will converge. In the first scenario, the two fluids are brought very close together in a single smallâdiameter tube such that molecules of fluid A only have to travel a short distance (hundreds of micrometres at most) by diffusion or convection, or both, to actually mix with molecules of fluid B. In this case, there is probably quite a lot of disordered flow that happens in the tube as it makes turns and goes through connections to valves and other fittings. This disordered flow is probably sufficient to mix two fluid streams if they really are consistent over time, as shown in scenario 1 of Figure 2. This ideal is never exactly achieved in pumps that use reciprocating pistons; we’ll address this in a bit more detail below.

Moving on to scenario 2 in Figure 2, we see that the way the fluids converge here is entirely different from the scenario in Figure 1. In the low-pressure mixing design, the composition of the mobile phase is determined by the duration of opening and closing of solenoid valves that are mounted on the solvent proportioning valve. For example, in the case where the desired mobile-phase composition is 25:75 A:B (as in Figure 2), solenoid A may open for 150 ms and then close, followed by solenoid B opening for 450 ms and then closing. These two events constitute one cycle of the valve that allows the pump to draw in a packet of the 25:75 A:B mixture, and this process repeats over time. The important difference between scenarios 1 and 2 is that in scenario 2 the fluids are brought together in a serial fashion, like slices of bread assembled into a loaf, whereas in scenario 1 the two streams are brought together in a parallel fashion. In scenario 2, this means that for the two fluids to actually mix to produce a homogeneous mixture, the molecules of fluid A have to be moved relatively long distance to mix with the molecules of fluid B, as shown by the curved arrow in Figure 2. This distance depends on a number of factors, including flow rate, the desired mixture composition, and the diameter of the tube connecting the solvent proportioning valve to the high pressure pump. Quite a bit of mixing will occur as the mixture is drawn into and pumped out of the high-pressure pump head, but generally speaking pumps with this low-pressure mixing design require mixers with larger volumes, mainly because this is what is required to effectively make the mixture of fluids homogeneous. There is one additional and important detail relevant to scenario 1 that is relevant here. As stated above, most pumps in use today that use the high-pressure mixing design use dual reciprocating pistons in each high-pressure pump. Although there are many variations on this design (2), the basic idea is that one piston draws fluid in from the solvent bottle while the other piston pushes fluid out at high pressure towards the column. These pistons accumulate and displace a finite volume, and so at the end of a piston stroke the travel of both pistons is reversed. At the end of the stroke, a check valve is used to prevent flow of mobile phase backward from the column into the pump. While this all works very well in principle, the important point here is that short-term deviations in the flow from each highâpressure pump are very hard to avoid. A simple view of the consequence of this is shown in Figure 3. When the flow of fluid A is temporarily lower at the end of a stroke, the composition of the combined fluid will be enriched in fluid B, and vice versa. In cases where the mobileâphase components are transparent to the detector (for example, water in the case of UV detection, or methanol in fluorescence detection), we are blind to these deviations in mobile phase composition. However, if the detector is even slightly sensitive to these changes (for example, when using formic acid in water with UV detection at 214 nm, or acetonitrile and water with refractive index detection), these deviations will be detected as a wavy pattern in the baseline. Here again, as was the case with low-pressure mixing systems, a mixer is needed in this case to move fluid molecules from one enriched region of the stream to another as shown by the arrow in Figure 3. And so while very little mixing is needed for high-pressure mixing designs in the ideal case (Figure 2, scenario 1), non-ideal behaviour of this type of pump requires that a mixer be added to smooth out the short-term variations in mobile phase composition.

Impact of Inadequate Mixing on Chromatographic Performance:

 The two major practical consequences of inadequate mixing are noisy detector baselines and poor retention time precision. The extent to which each of these is actually a problem in practice depends on a large number of experimental details, and a thorough discussion of all of them is far beyond the scope of this piece. However, a comparison of detector signals under typical operating conditions with and without mixing will illustrate the point that the extent of deliberate mixing in the system can have a big effect on the observed signal. Figure 4 shows absorbance signals from a UV detector when pumping 25:75 acetonitrile:50 mM ammonium formate in water from a high-pressure mixing pump, both with and without a mixer installed. The effect of the mixer is to smooth out short-term variations in mobile phase compositions, which obviously makes the detector baseline appear much less noisy. It is easy to imagine how these kinds of variations without mixing could lead to poor precision in retention time as well.

When Should We Consider Using a Different Mixer?

	As with a lot of other method development decisions in LC, the analyst must decide whether the mixer in use yields a level of performance for the method that is sufficient for the application at hand, and explore alternatives when it is not. Some applications are considerably more demanding than others in this regard. One that is particularly challenging, yet very important and in widespread use, involves the use of trifluoroacetic acid (TFA) in the eluent, reversedâphase LC columns, and gradient elution. Such conditions are very commonly used for separations of complex peptide mixtures, as well as mixtures of other hydrophilic amine-containing compounds. This is a particularly challenging case because not only does TFA absorb UV light at the same wavelengths that are used to detect peptides (most commonly 214 nm), but TFA itself is slightly retained by reversed-phase columns in water-rich mobile phases. This means that shortâterm variations in the mobile phase like that depicted in Figure 3 can lead to waves of TFA moving through the column as pulses of acetonitrile-rich mobile phase cause local decreases in retention of TFA and thus transient increases in elution of TFA from the column. Figure 5 shows a comparison of the UV absorbance baselines observed under these conditions, with two different mixers in use. This example clearly shows the benefit of using a larger mixer in this case, as the baseline becomes much smoother.

	One thing I have not touched on specifically in this article is mixer design. This is, in part, because there are too many different designs in use to cover in one article! However, although some designs undoubtedly are superior to others, it is wise in most cases to stick with the design supported by the manufacturer of your pump where possible. In cases where the pump manufacturer has limited offerings for different mixers, then in principle it is possible to mix and match pumps and mixers because almost all mixers in use today are passive devices-that is, they do not require power and do not have any moving parts. Readers interested in more details associated with specific mixer designs will find quite a lot of discussion on this point in Kromidas’ book (3).

	Although I just made the case that there should not be anything standing in the way of adding a large mobileâphase mixer to an LC system, it is very important that the user understands the consequence of adding a substantial mixing volume to the system. Any element that adds volume to the flow path between the fluid convergence point in a pump and the LC column will increase the gradient delay time. Again, this is the difference between the time that a change in mobile-phase composition is made at the convergence point and the time that this change arrives at the column inlet. Adding a mixer with a volume of 35 µL to a system that is running at 1 mL/min will not be very consequential because it will only change the gradient delay time by 0.035 min (~2 s). But, adding a larger volume (say, 200 µL) to a system running at 100 µL/min will have a much bigger effect, as this will increase the gradient delay time by 2 min! The effects of a large change like this will be numerous when gradient elution is used, and so should be made with care. A short list of changes that should be expected includes: significant (absolute) increases in retention times, significant changes in selectivity (relative retention), and increases in analysis time because of the slower arrival of the gradient start at the column, and longer time required to flush strong solvent from the system at the end of the gradient elution program.

	The performance of LC pumps has been improved dramatically in terms of mobile-phase composition precision and accuracy compared to what was possible in the earlier days of modern LC. This has led to a significant reduction in mixer volume and concomitant decreases in gradient delay time, which has been very helpful for increasing throughput of LC methods, reducing solvent waste, and enabling higher performing two-dimensional LC methods. However, even with the best available pumps, mixers are still needed, and the size of the mixer that is optimal varies somewhat by application. A more detailed understanding of why mixers are needed and when a change in mixers is warranted should lead to better method performance and improved method transferability in the future.

		F. Steiner, in The HPLC Expert II: Find and Optimize the Benefits of Your HPLC/UHPLC, S. Kromidas, Ed. (Wiley-VCH, Weinheim, Germany, 2017), pp. 101–170.	

		S. Kromidas, Ed., The HPLC Expert II: Find and Optimize the Benefits of Your HPLC/UHPLC (Wiley-VCH, Weinheim, Germany, 2017).

is the editor of “LC Troubleshooting”. Stoll is a professor and co-chair of chemistry at Gustavus Adolphus College in St. Peter, Minnesota, USA. His primary research focus is the development of 2D-LC for both targeted and untargeted analyses. He has authored or coauthored more than 50 peer-reviewed publications and three book chapters in separation science and more than 100 conference presentations. He is also a member of LCGC’s editorial advisory board. Direct correspondence to: 

LC Troubleshooting

Mixing and Mixers in Liquid Chromatography—Why, When, and How Much?  Part 1, The Pump

LCGC Europe spoke to Guilherme L. Alexandrino from the University of Copenhagen, Denmark, about the benefits of combining a modern “pixel-based” chemometrics technique with comprehensive twoâdimensional gas chromatography (GC×GC) in petroleomics applications.

Q. Do you think analysts are unnecessarily “scared” about using GC×GC?

Yes. The initial impression of GC×GC is that method development and data handling are both very complex. However, an understanding of the fundamental concepts of GC×GC, and particularly how the modulator works, makes method development simpler than it may first seem, and the number of publications on GC×GC has greatly increased in recent years. The data handling for two-dimensional chromatograms is also getting easier because the software from vendors and alternative open-source programs has become more “user-friendly”.

: One dimensional (1D) gas chromatography–mass spectrometry (GC–MS) is generally well-established and commonly used for targeted analysis. However, there are some applications where GC×GC can overcome the analytical limitations of 1D GC. For example, in petroleum analysis 1D GC–MS still requires SARA (saturates, aromatics, resign, and asphaltenes) fractionation of the crude oils before the analysis to separate saturates and aromatic compounds. Without this step chromatographic interferences can still occur, even when using more selective techniques, such as GC–MS with selected ion monitoring (SIM).

	This fractionation step can be laborious because it requires the use of organic solvents and is prone to cross-contamination. However, this limitation can be overcome by using GC×GC–MS with SIM (1) because the fractionation step for crude oils can be eliminated or, at least, simplified to remove all the asphaltenes because they cannot be analyzed by GC or GC×GC. In my view, GC×GC could also be used more routinely in other applications where the possibility of simplifying the sample preparation steps would be beneficial, for example, when associating GC×GC with “green” micro-extractions in metabolomics (2).

Q. What other benefits does GC×GC offer in petroleomics?

 In analytical chemistry, the “omics” are generally based on the analysis of a wide range of chemical features in samples to discover specific compounds that have correlation with an external property. For petroleomics, it is useful to be able to identify a specific geochemical property. Petroleum is an extraordinary complex substance that contains hundreds of thousands of different organic and inorganic compounds.

	The discovery of new petroleum biomarkers that can enhance the understanding of the physicochemical and geochemical properties of crude oils is seriously affected by coelutions when using 1D GC–MS. Chromatographic coelutions can often be handled quite satisfactorily when using selective MS techniques, such as SIM and tandem mass spectrometry  (MS/MS), but for the discovery of new petroleum biomarkers it is very difficult when a selective analysis is required because everything else in the crude oils becomes “invisible” in the analysis.

	The enhanced peak resolution and peak detection obtained by GC×GC with SIM has successfully handled this limitation, and a deeper understanding of the complex composition of the crude oils is achieved. For example, GC×GC has allowed the identification of new groups of petroleum biomarkers for improved geochemical diagnostics of crude oils (3,4,5) and has also  increased the chromatographic resolution of the so-called unresolved complex mixture (UCM) in crude oils. The UCM comprises numerous branched-, cyclic-alkanes, and aromatic hydrocarbons that cannot be resolved by conventional GC, resulting in a hump in the chromatograms that can be clearly observed, especially for degraded oils (6).

Q. You recently developed a novel GC×GCÐMS method with SIM to investigate the inter-reservoir geochemical features of crude oil. What was the aim of this research?

This method aimed to combine the chromatographic advantages of GC×GC with selective detection to identify the most important groups of petroleum biomarkers in the nonpolar fraction of crude oils incorporating a novel data analysis strategy based on chemometrics (1).

	We applied our new approach to the inter-reservoir discrimination of crude oils. The geochemical characterization of a petroleum reservoir is very important to understand its origin and its potential for economic exploitation. Although the “golden” method for geochemical investigation of petroleum is based on the peak integration of individual biomarkers obtained by GC–MS or GC–MS/MS, this method can be very laborious when trying to extract all the information potentially obtained by GC×GC. Additionally, the discovery of new patterns in the biomarkers “fingerprint” that distinguish crude oils from specific locations is more difficult and more “analyst-dependent”.

	We aimed to overcome these limitations by using chemometrics to mathematically extract the main chromatographic patterns from wellâknown classes of biomarkers that distinguish crude oils from specific reservoirs from the entire 2D-SIM chromatograms.

	We use all the chemical information potentially extracted by GC×GC because we directly handle the chromatographic signals (pixels) instead of user-extracted peak tables. This approach is called pixelâbased analysis, and when combined with chemometrics it successfully provides wider biomarker signatures that describe important geochemical features in crude oils.

	 In this case the chemical differences of crude oils originating from marine or lacustrine paleoâenvironments, and the discovery of biomarker patterns that described the inter-reservoir discrimination of similar lacustrine crude oils placed at specific locations were identified.

Q. Can you tell us more about what is novel about this“pixel-based” approach?

The pixel-based approach performs the data analysis directly from the chromatographic signals (pixel-by-pixel), instead of first processing the chromatograms through the conventional peak picking and integration procedures. Therefore, the amount of chemical information that can be assessed from the chromatograms by the pixelâbased approach is potentially bigger, because the data analysis is not restricted to the peaks selected by the conventional method. The term “pixel-based” appeared approximately 10 years ago, however, this concept has already been used without this specific name, for example, for comparing pairs of chromatograms (7). The benefits of the pixel-based approach in GC×GC is that enormous amount of peaks potentially resolved by this two-dimensional technique can be evaluated in a broader way. This is very interesting in petroleomics, because the alternative manual selection of individual peaks within complex two-dimensional chromatograms would be a very timeâconsuming task.

Q. Could this pixel-based chemometrics approach be useful in other areas outside petroleomics?

Yes, the pixel-based analysis has already been used in other approaches. In GC×GC for example, the pixel-based approach has already been used to perform untargeted identification of adulterants in diesel using pixels-by-pixel oneâway ANOVA (8). Furthermore, the pixel-based approach has also been proposed to perform forensic investigation of oil spill in the environment from GC–MS data (9,10). We are currenty investigating the use of the pixel-based approach for untargeted analysis of anthropogenic environmental pollution in the urban area of Copenhagen.

Q. Do you think chemometrics is being used to its full potential by the chromatography community?

I think the use of chemometrics within the chromatography community is still a growing field, which should be stimulated by the need for an efficient strategy to handle the big datasets provided by modern analytical instruments. In my view, there is still some resistance to using chemometrics in chromatography because it demands additional knowledge about data preprocessing and modelling that are not so common in traditional chromatography. However, I am very enthusiastic about the use of chemometrics in chromatography, and I believe there is a lot of potential for chemometrics in analytical chemistry.

	As well as the examples previously mentioned the combination of chromatography with chemometrics has already found many applications in the literature for the analysis of environmental samples, fuels, food, and flavours and fragrances (11,12). Chemometrics has also been an important tool for peak deconvolution in chromatography. Computational methods are able to resolve mathematically coeluting peaks, which also includes the resolution of the “pure” mass spectra of the individual resolved peaks when chromatography is hyphenated with mass spectrometry (13). Computational methods are able to resolve mathematically coeluting peaks, which also includes the resolution of the individual mass spectra of the resolved peaks when chromatography is hyphenated with mass spectrometry. therefore, less ambiguous identification of the resolved peaks can be done when interpreting the MS fragments and searching for potential hits in MS libraries.

		J.R.B. de Souza, F.F.G. Dias, J.D. Caliman, F. Augusto, and L.W. Hantao, 

		A.P. Kiepper, A. Casilli, and D.A. Azevedo, 

		J. Laakia, A. Casilli, B.Q. Araújo, F.T.T. Gonçalves, E. Marotta, C.J.F. Oliveira, C.A. Carbonezi, M.R.B. Loureiro, D.A. Azevedo, and F.R. Aquino Neto, 

LCGC Europe spoke to Guilherme L. Alexandrino from the University of Copenhagen, Denmark, about the benefits of combining a modern “pixel-based” chemometrics technique with comprehensive two‑dimensional gas chromatography (GC×GC) in petroleomics applications.

LCGC Europe-10-01-2018