Received: 2 March 2019 Revised: 5 May 2019 Accepted: 6 May 2019 DOI: 10.1002/sscp.201900021 RESEARCH ARTICLE Comparative study of different chemistries and particle properties, high-performance liquid chromatography stationary phases in separation of escitalopram oxalate and its impurities in different pharmaceutical dosage forms Wafa. F. S. Badulla1,2 Nafiz Öncü Can1 Zeki Atkosar1 Göksel Arli1 Hassan Y. Aboul-Enein3 1 Department of Analytical Chemistry, Faculty of Pharmacy, Anadolu University 26470, Eskisehir, Turkey 2 Department of Analytical Chemistry, Faculty of Pharmacy, Aden University, Aden, Yemen 3 Pharmaceutical and Medicinal Chemistry Department, Pharmaceutical and Drug Industries Research Division, National Research Centre, Cairo, 12622, Egypt Correspondence Professor Hassan Y. Aboul-Enein, Pharmaceutical and Medicinal Chemistry Department, Pharmaceutical and Drug Industries Research Division, National Research Centre, Dokki, Cairo 12622, Egypt. Email: haboulenein@yahoo.com The liquid chromatographic separation performance of 12 different columns was evaluated for the estimation of the escitalopram oxalate and six of its impurities in different pharmaceutical formulations. All columns exhibited a reasonable separation power. Six columns with three different chemistries and two different particle morphology (i.e. core-shell and fully porous particle) were nominated for further evaluation. The objectives of the current study were to develop a simple method for detection of escitalopram oxalate and its impurities and to compare the separation power of the stationary phase of different chemistry and particle properties. The separation was carried out under isocratic elution by using acetonitrile/methanol/water/phosphate buffer solution (pH = 3.5, 50 mM) (25:5:20:50, v/v/v/v) mixture as mobile phase, with a flow rate of 1.2 mL min−1 . The core-shell phenyl-hexyl column showed the best separation performance over the traditional octadecyl and pentafluorophenyl phases. The developed method was applied for detection of escitalopram oxalate and its impurities in valid and expired dosage forms. KEYWORDS core shell columns, escitalopram oxalate, impurities, pentafluorophenyl columns, porous particles 1 I N T RO D U C T I O N Escitalopram oxalate (ESC-OX) was approved by the Federal Drug Agency (FDA) as a selective serotonin reuptake inhibitor in 2002. The S enantiomer of the racemic mixture of the antidepressant drug citalopram which possessed more Article Related Abbreviations: ANOVA, analysis of variance; CIT A, citalopram A; CIT B, citalopram B; CIT C, citalopram C; CIT D, citalopram D; CIT E, citalopram E; DAD, photo diode array detector; ESC-OX, escitalopram oxalate; FDA, Federal Drug Agency; K’, capacity factor; MeOH, methanol; octadecyl, C18; OX-A, oxalic acid; PFP, pentafluorophenyl; USP, United State Pharmacopeia; α, selectivity of columns. Sep Sci plus 2019;1–16. potency and low side effect compared to the R enantiomer [1]. Control of the pharmaceutical impurities became a critical demand for the determination of drug safety and efficacy [2]. The analytical methods relating to the estimation of ESCOX impurities are very limited including HPLC, LC-MS/MS and capillary electrophoresis techniques [3–11]. Most of these methods were based on quantification and qualification of the R-enantiomer or other impurities that were not listed in the official USP monograph of the ESC-OX. The latest version of United State Pharmacopeia (USP 2016) provides three different methods for the determination of the ESC-OX impurities in raw material and different dosage forms [12]. The main aims of this work were to develop simple, isocratic analytical method for the determination of ESC-OX www.sscp-journal.com © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1 BADULLA ET AL. 2 and five most common its impurities named as CIT A, B, C, D, E and oxalic acid (OX-A) and to study the selectivity of different columns with diverse characteristics as an effective alternative to the current official USP used and other developed methods in order to achieve better resolution at a short analysis time. Using columns of different chemical properties provides separation of unresolved peaks of two compounds [13]. Evaluation and comparison of the separation performances of 12 columns were performed by using a system suitability test according to USP 2016. Six columns of three different stationary phase chemistries (C18 (octadecyl), pentafluorophenyl (PFP) and phenyl-hexyl) and two different particle properties (i.e. core-shell and fully porous particle) were selected for further comparative study. The validation was carried out by using the three core-shell columns. The selectivity of the six columns was evaluated by using correlation coefficients between the K values of all analytes [14]. The ESC-OX and its impurities are basic compounds. The analysis of these compounds with conventional C18 column requires acidic mobile phase, which may be problematic due to the interaction of the residual silanol on the silica material. These interactions may result in poor separation with peak tailing which affect the resolution of the basic analytes. For this reason, the separation power of stationary phases of different chemistry was examined. The effect of stationary phase chemistry was evaluated by applying the same analytical and instrumental parameters. The applicability of the established method was evaluated by analyzing valid and expired dosage forms. 2 2.1 M AT E R I A L S A N D M E T H O D S Drugs and reagents The ESC-OX and its impurities (Figure 1) were supplied by the United States Pharmacopeia Convention, USA, with 99.2% purity. All mobile phase compositions were of HPLC grade (Sigma-Aldrich, Switzerland). 2.2 Instrumentation and chromatographic conditions The method development was carried out in HPLC system A, containing a gradient pump with the LC-20AT model LPGE unit, SIL-20AC automatic sampler, CBM-20A model communication unit, CTO-10-ASvp model column furnace, SPD 20A model UV/VIS detector and DGU-20ASR model degassing. An Intel Pentium 4 processor computer and Shimadzu LabSolutions, LC solution version 1.25 data analysis program was used in addition to the HPLC product of Shimadzu (Japan). The pressure changed with column types, it was between 258–285 bar for core-shell columns and between 142–180 bar for fully porous columns. The dead volume is 300 μL. Other instruments used for ruggedness and specificity studies consist of the following parts: the gradient pump, SIL20A automatic sampler, CBM-20A model communication unit, CTO-10-ASvp model column oven, SPD-M20A model photo diode array detector (DAD) and DGU-20A5 model The chemical structures of escitalopram and its impurities FIGURE 1 BADULLA ET AL. 8% Reference: a (17), b (18), c (19), d (20), e (21). 11% 2–9 2–9 9% Core-Shell particles Core-Shell particles 200 m2 g−1 200 m g 1.5–8.5 1.5–8.5 100 Å 100 Å 2.6 μm 15 cm × 4.6 mm 15 cm × 4.6 mm c Kinetex PFP (phenomenex) Kinetex® C18 (phenomenex) 2.6 μm −1 2 c ® Core-Shell particles Fused-Core particle platform 150 200 m2 g−1 1.5–8.5 2–9 90 Å 100 Å 2.6 μm 2.7 μm 10 cm × 4.6 mm 15 cm × 4.6 mm c Express Phenyl-Hexyl (Supelco Analytical) Kinetex® Phenyl-Hexyl (phenomenex) g−1 m2 e Ascentis® 19% 2–8 Fused-Core particle platform Fused-Core particle platform 150 m2 g−1 150 m g 1–9 2–9 90 Å 90 Å 2.7 μm 10 cm × 4.6 mm 10 cm × 4.6 mm e Ascentis Express F5 (Supelco Analytical) Ascentis® Express C18 (Supelco Analytical) 2.7 μm −1 2 e ® 15% Porous Silica Fully Porous Silica 475 m2 g−1 1.5–7.5 80 Å 4 μm 15 cm × 4.6 mm Synergi® Hydro-RP (Phenomenex) 350 m2 g−1 2–8 100 Å 3 μm 10 cm × 4.6 mm d Nucleosil® C18 (Teknokroma) c 11.5 Ultra-pure, metal-free silica (99.99 % purity) 1.5–9.0 10% 11.5% Ultra-pure, metal-free silica (99.99 % purity) 400 m g 400 m2 g−1 100 Å 100 Å 15 cm × 4.6 mm 3 μm 15 cm × 4.6 mm c Luna PFP (Phenomenex) Luna® Phenyl-Hexyl (Phenomenex) 3 μm 1.5–9.0 −1 2 c ® Matrix High Purity, base deactivated 100 m2 g−1 450 2–7.5 2–8 100 Å 100 Å 3 μm 15 cm × 4.6 mm 5 μm g−1 m2 Surface Area pH Range Pore Size Particle Size L x I. D. 10 cm × 4.6 mm b The comparison of the performances of 12 columns was performed by evaluation of the system suitability test according to the USP 2016. The specificity of the established method was determined according to the ICH Q2(R1) guideline [20]. The chromatograms of the blank oral solution and the tablet matrix were compared with the chromatograms of the cited analytes by using three core-shell columns. The linearity of the developed method was evaluated by mean of three core-shell columns. The calibration curve for ESC-OX and its impurities was carefully chosen in subsequent range ESC-OX (0.236–70.70 μg mL−1 ), CIT A (0.0241–72.30 μg mL−1 ), CIT B (0.208–62.40 μg mL−1 ), CIT C (0.202–60.60 μg mL−1 ), CIT D (0.208–62.40 μg mL−1 ) and ODS-3 (GLSciences) Method validation Fluophase® RP (Thermo Scientific) 2.5 Physicochemical Properties of the Tested Columns The mobile phase composition contains ACN/methanol (MeOH)/H2 O/phosphate buffer (50 mM, pH = 3.5) (25:5:20:50, v/v/v/v). The phosphate buffer solution (50 mM, pH = 3.5) was prepared by dissolving 6.8 g of KHPO4 .2H2 O and 115.6 μL of H3 PO4 (d = 1.695 g mL−1 and 86%) in ultrapure water (V = 1000 mL). The pH of the buffer solution was controlled by using a calibrated pH-meter. After preparation, the solution was filtered through a 0.45 μm pore size cellulose membrane filter and sonicated for 15 min to remove dissolved gases. Mixing of the quantified solvents at the anticipated percentages was carried out via the low-pressure gradient mixing unit of the HPLC device. a Preparation of mobile phase TABLE 1 2.4 Inertsil® To prepare the standard stock solutions of the analytes, appropriate amount of standard was weighed and dissolved in 25 mL of ACN/water (30:70, v/v) solution. For this purpose, 3.01 mg of ESC-OX, 2.41 mg of CIT A, 2.08 mg of CIT B, 2.02 mg of CIT C, 2.08 mg of CIT D, 2.06 mg of CIT E and 2.57 mg of OX-A were used. Accordingly, the stock solutions were mixed and diluted using the same diluent to obtain the test solution used in the method development studies. The final concentrations of the analytes were as follows: ESC-OX 15.05 μg mL−1 , CIT A 12.05 μg mL−1 , CIT B 10.40 μg mL−1 , CIT C 10.10 μg mL−1 , CIT D 10.40 μg mL−1 , CIT E 10.30 μg mL−1 and OX-A 12.85 μg mL−1 . 3 Series High Purity Silica Gel Standards preparations Column (Brand) 2.3 Carbon Load degassing equipped with LP. The Intel Pentium 4 processor computer and the Shimadzu Lab Solution LC solution v1.11 SP1 data analysis program were used in addition to the HPLC product of Shimadzu (Japan). Twelve columns were applied in this study as it is represented in the Table 1 [15–19]. 15% 3 BADULLA ET AL. 4 The effect of mobile phase pH on the retention of ESC-OX and its impurities by using (column: Kinetex® PFP core-shell column (3 μm, 150 × 4.6 mm), mobile phase; buffer: ACN: MeOH: H2 O (50:25:5:20, v/v/v/v). (a) pH = 2, (b) pH = 3, (c) pH = 3.5, all phosphate buffer, (d) pH = 4(acetate Buffer), (e) pH = 5(acetate Buffer), (f) pH = 6 (phosphate Buffer) FIGURE 2 CIT E (0.206–61.80 μg mL−1 ). The linearity was estimated by linear regression analysis with intra- and interday replicates. The LOD and LOQ values were calculated by using signal to noise ratio. For each column, the (S/N) ratio attained from the chromatograms is multiplied by 3 for the LOD and 10 for the LOQ. Moreover, probable variations between the analytical columns were studied by one-way ANOVA test. Statistical calculations were completed with Graph Pad Prism v 6.0 b software. The standard addition method was used to evaluate the accuracy of the current method. Three different percentages (50, 100, 150%) of each analyte was added to the preanalyzed tablet and oral solution. All the concentration levels were prepared three times, spiked to the pharmaceutical dosage forms and then the mean recovery was calculated. The intraday and interday (intermediate) precision evaluations were performed at the lowest concentration by using standard solutions for each analyte. In order to estimate the method robustness. The changes in mobile phase composition, concentration, buffer pH, wavelength, flow rate and temperature were followed according to Vander Heyden [21]. The standard solutions of ESC-OX and its impurities were examined for the short term (1 day), freeze and thaw and long term (30 days). 3 RESULTS AND DISCUSSION 3.1 Selection of the maximum UV absorbance wavelength The scanning of the maximum absorbance at UV-region was carried out between 200–400 nm for ESC-OX and its impurities. All cited analytes had maximum absorbance in 196–210 nm region. Accordingly, 210 nm was selected for further study. 3.2 Chromatographic conditions The selection of mobile phase composition was performed by using Phenomenex Kinetex® PFP core-shell column. All tested analytes are basic compounds; the pKa value of ESCOX is 9.5 while the predicted pKa of its impurities is in the following ranges 9.48–16.06. In order to get a reproducible result, the ionization of the cited analytes should be controlled by adding the buffer solution to the mobile phase. Several buffer salts in the pH range of 2–6 were tested. The retention time of the ESC-OX and its impurities were extremely affected by the mobile phase buffer pH. The examination of pH effect was conducted by using mobile phase that contains, BADULLA ET AL. The chromatograms related to the method development by using Kinetex® PFP core-shell column (a) buffer: ACN: H2 O (50:30:20) v/v/v/v, (b) buffer: ACN: H2 O (50:25:25, v/v/v/v), (c) buffer: ACN: MeOH: H2 O (50:30:5:15, v/v/v/v), (d) buffer: ACN: MeOH: H2 O (50: 20:10:20, v/v/v/v), (e) buffer: ACN: MeOH: H2 O (50:25:5:20, v/v/v/v) FIGURE 3 5 BADULLA ET AL. 6 TABLE 2 Results of system suitability tests (n = 3) Kinetex® C18 Core-Shell Column. Parameters CIT A CIT B CIT C CIT D Inertsil ODS-3® C18 Fully Porous Column. ESC-OX CIT E CIT A CIT B CIT C CIT D ESC-OX CIT E N 3189.13 803.68 12157.45 14817.70 14472.58 17438.25 3682.75 1006.07 9758.81 10339.27 10900.16 11265.25 tR (min) 2.105 3.404 4.882 5.753 6.194 7.228 2.992 5.443 8.265 9.802 10.614 12.865 T 1.263 1.001 1.234 1.215 1.254 1.169 1.237 1.011 1.362 1.380 1.416 1.417 As 1.255 1.004 1.217 1.202 1.239 1.160 1.230 1.012 1.326 1.343 1.370 1.364 K’ 0.655 1.678 2.843 3.526 3.872 4.678 1.060 2.748 4.691 5.75 6.309 7.859 Rs 4.901 3.656 4.013 4.601 2.180 4.773 7.117 4.874 4.866 4.025 1.922 4.695 α 2.562 1.694 1.240 1.098 1.210 - 2.952 1.707 1.225 1.097 1.246 - RSD% of tR 0.41 0.34 0.19 0.20 0.19 0.30 0.19 0.25 0.29 0.33 0.32 0.27 Kinetex® PFP Core-Shell Column. Luna® PFP Fully Porous Column Parameters CIT A CIT B CIT C CIT D ESC-OX CIT E CIT A CIT B CIT C CIT D ESC-OX CIT E N 4415.28 769.69 15178.22 15987.17 14922.74 17851.06 5288.87 808.78 7407.11 7798.72 6884.89 6839.233 tR (min) 2.618 4.446 6.802 7.380 8.302 9.586 4.051 8.363 14.024 15.326 17.676 19.775 T 1.327 1.018 1.325 1.318 1.485 1.352 1.332 1.0389 1.359 1.357 1.416 1.398 As 1.297 1.023 1.293 1.290 1.447 1.314 1.294 1.034 1.320 1.319 1.374 1.360 K’ 0.829 2.107 3.754 4.159 4.805 5.700 1.789 4.747 8.655 9.547 11.163 12.609 Rs 7.071 4.580 5.465 2.547 3.648 4.588 11.202 6.145 6.190 1.922 3.005 2.292 α 2.540 1.781 1.108 1.155 1.186 - 2.653 1.823 1.103 1.169 1.130 - RSD% of tR 0.187 0.214 0.239 0.244 0.367 0.251 0.324 0.360 0.435 0.422 0.469 0.445 Kinetex® Phenyl-Hexyl Core-Shell Column. Parameters CIT A CIT B CIT C CIT D Luna® Phenyl-Hexyl Fully Porous Column. ESC-OX CIT E CIT A CIT B CIT C CIT D ESC-OX CIT E N 3155.52 614.36 11998.06 14284.38 14990.14 15107.56 5878.93 834.011 15508.38 16635.45 16877.16 16249.16 tR (min) 2.272 3.697 5.242 6.072 6.712 7.861 3.399 6.353 9.585 11.307 12.632 14.620 T 1.335 1.001 1.277 1.252 1.336 1.470 1.237 1.023 1.108 1.109 1.149 1.151 As 1.310 1.003 1.257 1.241 1.315 1.394 1.214 1.010 1.105 1.105 1.147 1.140 K’ 0.745 1.853 3.048 3.684 4.173 5.606 1.386 3.461 5.731 6.941 7.871 9.266 Rs 5.695 3.758 3.920 4.207 3.027 4.836 10.301 5.588 5.441 5.232 3.581 4.689 α 2.487 1.043 1.208 1.133 1.343 - 2.497 1.656 1.211 1.134 1.177 - RSD% of tR 0.067 0.041 0.029 0.025 0.031 0.048 0.158 0.286 0.354 0.336 0.355 0.348 buffer: ACN: MeOH: H2 O (50:25:5:20, v/v/v/v). Since ESCOX and its impurities are basic compounds, these compounds are relatively ionized by using a buffer solution of pH 2–6 and so their retention times were reduced. By using phosphate buffer of pH = 2 the shape of CIT B was odd. This can be explained by interconversion between chiral isomers of CIT B [22]. By changing the buffer pH to 3 the peak shape of CIT B became broad. Using a buffer solution of pH 4, the peak of CIT B was sharp; however, the ESC-OX and CIT E eluted together. The retention times of the analytes in the buffer pH range from 2–4 depended mainly on their polarities (log p). At pH 5 and pH 6, CIT C and CIT E co-eluted. The elution order was changed at pH 5 and 6, the CIT E eluted between CIT C and CIT D, ESC-OX was the latest eluting compound in contrast to the pH 2 to 3.5, where CIT E was the latest analyte. Alteration of the sequence of elution can be clarified by variation in the ionization degree of the compounds and electrostatic interaction between the F atom in PFP stationary phase and amine lone pair in ESC-OX and its impurities [23]. Using buffer of higher pH leads to delay in the retention time of all analytes. This can be clarified by the decrease in the degree of ionization, and may be due to slight π-π interaction between the aromatic rings of analytes and stationary phase [24] and some electrostatic interaction or by strong adsorption of compounds with the stationary phase [25]. However, the particular mechanism for the multipart interaction of these compounds with stationary phase by changing pH is complex and yet unsolved [26]. It is clear from the pH scanning that the retention times of all analytes and the peak shapes are strongly influenced by the pH value as shown in Figure 2. By using potassium phosphate buffer of pH 3.5 (50 mM), the peak shape of CIT B was sharper with a good resolution and reasonable retention time. Therefore, further analyses were carried out at this pH. Various percentages of ACN were tested and the best result was obtained at 25% by volume, but the resolution between the CIT C and CIT D BADULLA ET AL. 7 Chromatogram of standard solution of ESC-OX and its impurities by using (a) Kinetex® C18 (2.6 μm, 150 × 4.6 mm) core-shell column, (b) Inertsil® C18 (3 μm, 150 × 4.6 mm) fully porous column, (c) Kinetex® PFP (2.6 μm, 150 × 4.6 mm) core-shell column, (d) Luna® PFP (2.6 μm, 150 × 4.6 mm) fully porous column, (e) Kinetex® phenyl-hexyl (2.6 μm, 150 × 4.6 mm) core-shell column, (f) Luna® phenyl-hexyl (2.6 μm, 150 × 4.6 mm) fully porous column FIGURE 4 was low; thus, to improve the resolution between these compounds and peak symmetry, and reduce tailing of CIT E peak, MeOH was added. MeOH has an UV cut-off of 205 nm. For getting absorbance of MeOH less than 0.05 AU (absorbance unit), the working wavelength must be chosen to be greater than 235 nm or the MeOH percentage must be lower than 15%. The best result was attained by using the buffer of ACN: MeOH: H2 O (50:25:5:20, v/v/v/v). In addition, various gradient elution was also tried, however, the best result obtained by isocratic elution which is preferred over the gradient elution due to not requiring special instrument, highly trained technician and easy to prepare. The effect of flow rate is relatively small on the resolution, but it affects the overall analysis time and last peak’s tailing. The flow rates of 1.1, 1.2 and 1.3 mLmin−1 were tried. The reasonable analysis time with the best peak shape was acquired for the 1.2 mLmin−1 flow rate. The temperature should be kept constant during analysis time because it affects selectivity. The column oven temperature was varied as 35, 40 and 45◦ C. The column oven temperature kept at 40◦ C because no valuable decrease in run time was observed at the higher temperature. This temperature was best for maintaining low viscosity of the mobile phase and back pressure, providing fast flow rate and lower the analysis time. In addition, two buffer concentrations were tried (25 and 50 mM). 50 mM was selected because the peaks were sharper, especially for CIT B. The chromatograms related to the method development are given in Figure 3. 3.3 Selection of the columns Selection of the columns plays important role in getting a powerful separation with reasonable resolution. In this work, 12 columns with different length and particle properties were 8 examined (fully porous and core-shell). After evaluation of the separation power of all columns by using SST (Table 2), six best performing columns were selected for further study. The choice of PFP column was established on the existence of fluorine (F) in the stationary phase to enhance the retention and selectivity of the halogenated polar compounds [19]. In addition, this stationary phase acts as Lewis acid (electron acceptor) and offers π-π interaction, dipole-dipole interaction, steric and shape/size selectivity [27]. Choosing the phenyl-hexyl stationary phase was established on having variant selectivity for aromatic and hydrophobic interactions. This selectivity comes from the existence of six-carbon chain and aromatic phenyl group. Phenyl is considered as a Lewis base (electron donor) [19], a good source of selectivity for the amine and heterocyclic containing aromatic compound and π-π interaction [28]. Two different chemistry stationary phases (PFP and phenyl-hexyl) were selected and compared with the traditional C18 column. The idea of using different BADULLA ET AL. physical and chemical stationary phases was found useful to evaluate the selectivity of columns. The elution order of all compounds was the same in all of the columns; only a trivial superiority of phenyl hexyl stationary phase was detected in term of resolution. This may be originated from different multisite interactions of the compounds with the phenyl ring. Relatively longer retention in the PFP stationary phase in comparison with the other two stationary phases is due to the presence of the F atom which can form dipole-dipole interactions and H-bonds with the solutes, resulting in slightly longer retention of the solutes. Also, the longer retention may be due to the larger size of fluorinatedphenyl ring in comparison to the C18 and phenyl-hexyl ligands of the stationary phase [29]. This leads to increase steric retention of the solutes. The peak symmetry was relatively better by using phenyl-hexyl stationary phase, especially for CIT B. The chromatograms relating to the 12 columns are shown in Figure 4 and 5. The result of comparing the system suitability test of the six-selected columns (Table 2) indicated Chromatogram of ESC-OX and its impurities by using;(a) Nucleosil® 100 C18 (3 μm, 10 × 0.46 cm), (b) Synergi Hydro-RP® (4 μm, 15 × 0.46 cm), (c) Fluophase RP® (3 μm, 10 cm × 4.6 mm), (d) Ascentis® Express C18 (2.7 μm, 10 cm × 4.6 mm), (e) Ascentis® Express F5 (2.7 μm, 10 cm × 4.6 mm), (f) Ascentis® Express Phenyl-hexyl (2.7 μm, 10 cm × 4.6 mm) FIGURE 5 BADULLA ET AL. that all core-shell columns provide excellent separation of the ESC-OX and its impurities in regard to analysis time, theoretical plate and resolution. The core-shell columns have the highest number of theoretical plates with the preference of PFP column, the retention times were almost similar in the three core-shell columns. Comparing, the capacity factor (K’) and selectivity of columns (α) the best results obtained by using fully porous columns that provide suitable analysis time with good selectivity but the peaks were slightly broader. The difference in the tailing factor of all columns was not great and all values were within the recommended value (T < 2). The resolution values between the CIT D, ESC-OX and CIT E were close to each other in most columns with the slight superiority of phenyl-hexyl column. In concern to the entire columns’ performance, all core-shell columns provide excellent separation of the ESC-OX and its impurities in regard to analysis time, theoretical plate and resolution. The dissimilarity factor of the core-shell column with the fully porous column was estimated by mean of regression analysis and correlation coefficients (r) between the K’ values of all analytes. The low correlation coefficient shows the high dissimilarity and vice versa [30]. A plot of K’ achieved from Inertsil® C18 fully porous column versus the K’ achieved from Kinetex® C18 core-shell column refers a high degree of correlation (r = 0.99959), i.e. high degree of similarity as 9 shown in Figure 6 (a). By relating the correlation between K’ achieved from Luna® PFP fully porous column versus the K’ achieved from Kinetex® PFP core-shell column and Luna® phenyl-hexyl fully porous column versus the K’ achieved from Kinetex® phenylhexyl core-shell column the result displays that the values are, r = 99872 and r = 0.99110, respectively (Figure 6 b, c). Consequently, it can be inferred that the similarity in the selectivity between core-shell column and fully porous column for ESC-OX and its impurities order as follows: C18 > PFP > phenyl-hexyl. The selectivity of stationary phase chemistry of conventional C18 with the PFP and phenyl-hexyl stationary phases were compared by the similar manner and the correlation coefficient between the C18 with the PFP and C18 with phenyl-hexyl stationary phases indicate that the similarity of PFP column selectivity is close to the C18 more than the phenyl-hexyl column. The correlation coefficients are, r = 0.9970 and r = 0.9921 (Figure 6 (d)). The result of the selectivity approves the slight superiority of phenyl-hexyl column in the separation for the ESC-OX and its impurities over the conventional C18 and PFP. However, there is an insignificant dissimilarity between the C18 traditional column and the other two stationary phases which refers to the ability of using traditional C18 in separation of ESC-OX and its impurities in case of unavailability of other columns with specific chemical ligands. k’ of ESC-OX and its impurities on (a) C18 fully porous column versus C18 core-shell column, (b) PFP fully porous column versus PFP core-shell column, (c) phenyl-hexyl fully porous column versus phenyl-hexyl core- shell column, (d) K’ of ESC-OX and its impurities on C18 column versus PFP and phenyl-hexyl core-shell column versus PFP Column FIGURE 6 10 TABLE 3 Statistical Evaluation of Linearity and precision studies Kinetex® Phenyl-Hexyl Core-Shell Column. Linearity range (μg mL−1 ) Slope ± SDa (intraday, n = 27) CIT A CIT B CIT C CIT D ESC-OX CIT E 0.0241–72.30 0.208–62.40 0.202–60.60 0.208–62.40 0.236–70.70 0.206–61.80 36667 ±52.60 24005 ± 165.8 63780 ± 191.50 30544 ± 43.41 37942 ± 72.95 27851 ± 41.07 Intercept ± SDa (intraday, n = 27) 11214 ± 1598 3304 ± 4346 4445 ± 4876 2547 ± 1138 1672 ± 2167 -5770 ± 1066 Regression coefficient (intraday, n = 27) 0.9999 0.9988 0.9998 0.9999 0.9999 0.9999 95% CI of Slope 36558 to 36775 23664 to 24347 63385 to 64174 30455 to 30634 37792 to 38092 27767 to 27936 95% CI of Intercept 7922 to 14506 −5649 to 12257 −5598 to 14489 201.9 to 4892 −2792 to 6137 −7967 to −3573 LOQ 0.005 0.008 0.003 0.006 0.006 0.007 LOD 0.016 0.025 0.009 0.019 0.017 0.021 Repeatability (inter-day, mean, n = 18) 13835.56 4344.83 10302.94 6322.72 7249.17 3631.72 Repeatability (inter-day, RSDc %, n = 18) 0.71 0.86 0.62 0.87 0.76 1.15 Repeatability (inter-day, SDa, n = 18) 97.95 37.13 63.86 55.01 55.16 41.81 Repeatability (intraday, SEMd, n = 18) 39.99 15.16 26.07 22.46 22.52 17.07 Repeatability (inter-day, CIe, n = 18) 13732.79 to 13938.32 4301.93 to 4301.93 10235.95 to 10369.94 6265.01 to 6380.43 7191.30 to 7307.04 3587.85 to 3675.59 Kinetex® PFP Core-Shell Column. CIT A CIT B CIT C CIT D ESC-OX CIT E Linearity range (μg mL−1 ) 0.0241–72.30 0.208–62.40 0.202–60.60 0.208–62.40 0.236–70.70 0.206–61.80 Slope ± SDa (intraday, n = 27) 40898 ± 54.94 25068 ± 93.82 64661 ± 96.28 30168 ± 51.71 38535 ± 35.59 27162 ± 39.32 Intercept ± SDa (intraday, n = 27) 1874 ± 1669 −467.8 ± 2088 −5340 ± 2452 207.7 ± 1356 −454.2 ± 1057 2158 ± 1021 1.000 0.9996 0.9999 0.9999 1.000 0.9999 95% CI of Slope −1564 to 5312 24874 to 25261 64463 to 64860 30062 to 30275 38462 to 38609 27081 to 27243 95% CI of Intercept −0.1301 to 0.0382 −4769 to 3833 −10391 to -289.6 −2586 to 3001 −2632 to 1724 54.51 to 4262 LOQ 0.013 0.020 0.008 0.017 0.017 0.018 LOD 0.038 0.061 0.024 0.050 0.050 0.055 Repeatability (inter-day, mean, n = 18) 11611.50 5618.56 11114.39 5654.22 9349.67 6683.22 Repeatability (inter-day, RSDc %, n = 18) 1.22 1.77 0.63 1.26 1.19 1.31 Repeatability (inter-day, SDa, n = 18) 141.92 99.37 70.14 71.38 111.71 87.51 Repeatability (intraday, SEMd, n = 18) 57.94 40.57 28.63 29.14 45.61 35.72 Repeatability (inter-day, CIe, n = 18) 11497.94 to 11760.40 5514.30 to 5722.81 11040.80 to 11187.98 5579.33 to 5729.12 9232.46 to 9466.88 6591.41 to 6775.03 (Continues) BADULLA ET AL. Regression coefficient (intraday, n = 27) BADULLA ET AL. TABLE 3 (Continued) Kinetex® C18 Core-Shell Column. CIT A CIT B CIT C CIT D ESC-OX CIT E Linearity range (μg mL−1 ) 0.0241–72.30 0.208–62.40 0.202–60.60 0.208–62.40 0.236–70.70 0.206–61.80 Slope ± SDa (intraday, n = 27) 41326 ± 32.75 23086 ± 91.56 67498 ± 101.80 29412 ± 201.0 37468 ± 54.35 24007 ± 75.61 Intercept ± SDa (intraday, n = 27) −4896 ± 995.1 1435 ± 2401 −1067 ± 2592 3013 ± 5271 35.50 ± 1615 274.4 ± 1964 Regression coefficient (intraday, n = 27) 1.000 0.9996 0.9999 0.9988 0.9999 0.9998 95% CI of Slope 41258 to 41393 22897 to 23274 67288 to 67708 28998 to 29826 37356 to 37580 23851 to 24163 95% CI of Intercept −6946 to -2847 −3511 to 6381 −6406 to 4272 −7844 to 13871 −3290 to 3361 −3771 to 4319 LOQ 0.004 0.008 0.003 0.006 0.006 0.007 LOD 0.014 0.023 0.008 0.018 0.018 0.021 Repeatability (inter-day, mean, n = 18) 10309.40 5437.06 12678.17 6207.33 8710.16 6175.66 Repeatability (inter-day, RSDc %, n = 18) 0.79 1.73 0.73 1.18 1.21 0.98 Repeatability (inter-day, SDa, n = 18) 80.99 94.02 92.66 73.08 105.02 60.81 Repeatability (intraday, SEMd, n = 18) 33.06 38.38 37.83 29.83 42.87 24.82 Repeatability (inter-day, CIe, n = 18) 10224.43 to 10394.37 5338.41 to 5535.70 12580.95 to 12775.38 6130.66 to 6284.01 8599.98 to 8820.35 6058.20 to 6185.80 ANOVA (between columns) F (1.004, 8.035) = 3.276, P = 0.1077 F (1.011, 8.084) = 3.818, P = 0.0858 F (1.010, 8.084) = 4.238, P = 0.0729 F (1.019, 8.153) = 3.683, P = 0.0902 F (1.173, 9.384) = 4.026, P = 0.0702 F (1,010, 8.080) = 4.166, P = 0.0750 11 BADULLA ET AL. 12 TABLE 4 Results of Accuracy Studies in Tablet Matrix and Oral Solution by Using Kinetex® Phenyl-Hexyl Core-Shell Column Kinetex® Phenyl-Hexyl Core-Shell Column. Tablet Matrix Compounds CIT A CIT B CIT C CIT D ESC-OX CIT E TABLE 5 Added (𝛍g mL−1 ) Oral solution Found (𝛍g mL−1 ) Recovery % Average Recovery 98.87 2.41 2.36 98.04 4.82 4.80 99.53 7.23 7.16 99.04 2.08 2.14 103.04 103.44 Added (𝛍g mL−1 ) Found (𝛍g mL−1 ) Recovery % Average Recovery 97.92 2.41 2.40 99.53 4.82 4.66 96.65 7.23 7.055 97.58 2.08 2.11 101.24 4.16 4.34 104.23 4.16 4.34 104.22 6.24 6.43 103.06 6.24 6.48 103.84 2.02 2.02 99.80 4.04 4.08 101.04 6.06 6.05 99.92 2.08 2.04 98.29 100.2 100.77 2.02 1.96 97.03 4.04 4.06 100.48 6.06 6.03 99.50 2.08 2.04 97.86 4.16 4.32 103.86 4.16 4.11 98.84 6.24 6.25 100.18 6.24 6.25 100.12 2.35 2.33 98.73 4.71 4.83 102.41 7.07 7.03 99.39 2.06 2.09 101.45 100.17 101.08 2.35 2.39 101.36 4.71 4.82 102.38 7.07 7.45 105.36 2.06 2.08 100.97 4.12 4.22 102.42 4.12 4.15 100.62 6.18 6.14 99.39 6.18 6.28 101.62 103.10 99.00 98.94 101.62 101.07 Results of Accuracy Studies in Tablet Matrix and Oral Solution by Using Kinetex® PFP Core-Shell Column Kinetex® PFP Core-Shell Column Tablet Matrix Compounds CIT A CIT B CIT C CIT D ESC-OX CIT E Added (𝛍g mL−1 ) Oral solution Found (𝛍g mL−1 ) Recovery % Average Recovery 99.99 2.41 2.44 101.14 4.82 4.84 100.37 7.23 7.12 98.45 2.08 2.12 101.88 101.10 Added (𝛍g mL−1 ) Found (𝛍g mL−1 ) Recovery % Average Recovery 99.86 2.41 2.46 101.92 4.82 4.77 99.02 7.23 7.13 98.63 2.08 2.08 100.22 4.16 4.15 99.73 4.16 4.03 100.91 6.24 6.33 101.36 6.24 62.8 101.24 2.02 2.06 101.74 4.04 3.97 98.21 6.06 6.02 99.35 2.08 2.10 100.56 99.77 99.54 2.02 2.04 101.16 4.04 4.02 99.49 6.06 6.01 99.17 2.08 2.12 101.74 4.16 4.12 99.15 4.16 4.13 99.22 6.24 6.17 98.90 6.24 6.24 99.93 2.35 2.34 99.50 4.71 4.86 103.11 7.07 7.20 101.86 2.06 2.11 102.25 101.49 101.15 2.35 2.43 103.14 4.71 4.85 102.96 7.07 7.25 102.65 2.06 2.11 102.25 4.12 4.21 102.11 4.12 4.21 102.11 6.18 6.19 100.18 6.18 6.23 100.77 100.79 99.94 100.30 102.92 101.71 BADULLA ET AL. TABLE 6 13 Results of Accuracy Studies in Tablet Matrix and Oral Solution by Using Kinetex® C18 Core-Shell Column Kinetex® C18 Core-Shell Column Tablet Matrix Compounds CIT A CIT B CIT C CIT D ESC-OX CIT E 3.4 Added (𝛍g mL−1 ) Oral solution Found (𝛍g mL−1 ) Recovery % Average Recovery 104.17 2.41 2.54 105.27 4.82 4.96 103.00 7.23 7.54 104.25 2.08 2.21 106.47 104.85 Added (𝛍g mL−1 ) Found (𝛍g mL−1 ) Recovery % Average Recovery 105.58 2.41 2.59 106.58 4.82 5.02 104.21 7.23 7.66 105.96 2.08 2.16 103.82 4.16 4.35 104.45 4.16 4.33 104.18 6.24 6.47 103.62 6.24 6.68 107.00 2.02 1.99 98.29 4.04 3.90 96.58 6.06 5.84 96.30 2.08 2.15 103.41 97.06 102.41 2.02 2.04 100.83 4.04 3.92 96.98 6.06 5.87 96.93 2.08 2.08 99.76 4.16 4.21 101.27 4.16 4.12 99.08 6.24 6.40 102.56 6.24 6.31 101.15 2.35 2.38 100.84 4.71 4.59 97.48 7.07 6.92 97.84 2.06 2.11 102.56 98.72 103.06 2.35 2.37 100.45 4.71 4.63 98.33 7.07 7.05 99.78 2.06 2.18 105.89 4.12 4.28 103.79 4.12 4.20 101.83 6.18 6.35 102.82 6.18 6.37 103.00 Method validation The result of the specificity test showed that there were no interfering peaks at the retention time of the examined compounds. As a complement to the specificity tests, stability studies were carried out. Short term (1 day), freeze and thaw and long term (30 days) stability studies and the monitoring of the stability showed that all compounds were stable at all studied conditions. The linearity was estimated by linear regression analysis by using nine concentrations for all analytes by using the three core-shell columns. The statistical data on the intraday and inter-day linearity of the method for the three coreshell columns are given in Table 3. In addition, possible differentiations between the analytical columns were examined by one-way ANOVA test. The results revealed that there were no noteworthy statistical differences among the analytical columns for all analytes. The precision of the methods was studied at two steps: intraday (repeatability) and interday (intermediate) precision. The results showed that the method was analytically adequate (RSD% < 2.0) in relation to the precision. The precision data are outlined in Table 3. The developed method accuracy was evaluated by the standard addition method for both dosage form matrixes and the results were revealed that the current method is sufficiently accurate (Table 4–6). 105 98.25 100.00 99.52 103.57 The robustness and raggedness evaluation outcome revealed that the developed method was sufficiently robust with RSD% below 2%. Moreover, the ruggedness of the method was evaluated by the other HPLC system. There was no significant difference between the chromatograms of the two HPLC systems. The applicability of the developed method was assessed by using it for the determination of ESC-OX and its impurities (if any) in different pharmaceutical dosage forms (tablet and oral solution). The statistical result of different dosage forms analyses by three core-shell columns are summarized in Table 7. At the same time, an expired oral solution (expiration date: April 2015) and tablet dosage form (expiration date: August 2016) were examined to determine the presence of impurities. The results obtained from different pharmaceutical dosage forms indicated the presence of CIT C impurity in all dosage forms with a very low quantity. The impurity CIT C is considered as one of the critical impurities that can arise during cyclization step and it might remain after conventional purification steps [31]. In the tablet dosage form, CIT C and CIT A were detected. Impurity CIT A may be resulted from hydrolysis [32] and also monitored during synthesis of the ESC-OX [33]. In the evaluation of the expired oral solution, dosage form impurity CIT D was detected alongside impurity CIT C. This CIT D was detected and identified as a degradation product of BADULLA ET AL. 14 TABLE 7 Assay Results of ESC-OX pharmaceutical dosage forms ESC-OX 20 mg Film Tablet (n = 6) C18 column PFP column phenyl-hexyl column Average (mg) 20.06 20.07 20.12 Minimum (mg) 20.00 20.00 20.07 Maximum (mg) 20.10 20.10 20.19 SD (mg) 0.032 0.034 0.046 RSD % 0.16% 0.17% 0.23% SEM (mg) 0.0104 0.0107 0.0189 Bias% −0.30 −0.35 −0.60 95% CI of Average 20.03 to 20.08 20.04 to 20.09 20.07 to 20.17 % Content Uniformity 100.3 100.35 100.6 ANOVA F (2, 27) = 3.066, P = 0.0631 C18 column PFP column phenyl-hexyl column 10.34 10.30 10.29 ESC-OX 10 mg Film Tablet (n = 6) Average (mg) Minimum (mg) 10.27 10.16 10.20 Maximum (mg) 10.39 10.42 10.33 SD (mg) 0.042 0.072 0.037 RSD % 0.41% 0.70% 0.36% SEM (mg) 0.0133 0.0228 0.0117 Bias% −3.4 −3.0 −2.9 95% CI of Average 10.31 to 10.37 10.25 to 10.35 10.27 to 10.32 % Content Uniformity 100.3 100.35 100.2 ANOVA F (2, 27) = 2.009, P = 0.1537 ESC-OX Oral Solution (10 mg mL−1 ) C18 column PFP column Average (mg) 10.23 10.21 phenyl-hexyl column 10.25 Minimum (mg) 10.18 10.14 10.21 Maximum (mg) 10.33 10.27 10.29 SD (mg) 0.054 0.043 0.025 RSD % 0.53% 0.42% 0.24% SEM (mg) 0.0170 0.0135 0.0079 Bias% −2.3 −2.1 −2.5 95% CI of Average 10.19 to 10.27 10.18 to 10.24 10.23 to 10.27 % Content Uniformity 102.3 102.1 102.5 ANOVA CIT (racemic mixture) [34]. The examination of the expired tablet, identified only impurity CIT A (Figure 7). The detected quantities of all impurities were below the maximum allowed limit of USP 2016. 4 CONC LU SI ON In this study, a simple isocratic, reversed phase HPLC method with a run time of 12 min was developed for quantitative F (2,27) = 1.835, P = 0.1789 determination of ESC-OX and its impurities. Moreover, the method can be successfully applied for detection of impurities in the pharmaceutical dosage forms. The developed method was applied by using 12 different columns. Thereafter, six different columns (core-shell and fully porous) were examined with three different stationary phases. The best separation was obtained from core-shell columns with good peak symmetry, resolution and short analysis time. The phenyl hexyl stationary phase with middle hydrophobic selectivity in comparison with the C18 and PFP stationary phase provided the best BADULLA ET AL. 15 Chromatograms represent impurities in (a) an expired oral solution, (b) oral solution by using phenyl-hexyl column, (c) an expired tablet dosage form 20 mg, (d) tablet dosage form 20 mg by using phenyl-hexyl column FIGURE 7 resolution. The PFP fully porous stationary phase proved to be more retentive with poor resolution and peak shape; however, the developed method can be utilized by using C18 fully porous column if they are the only available choice with relatively longer run time about 15 min. The method was fully validated and the results obviously show that it is accurate and precise with low LOD, which makes it a preferable alternative from the official USP method. By the development of this method, both objectives of this study were achieved with satisfactory results. ACKNOW LEDGMENTS The authors acknowledge the financial support provided by the Anadolu University project No: 1502S091, Eskişehir, Turkey. CO NFLICT OF I N T E R E ST The authors have declared no conflict of interest. ORC ID Hassan Y. Aboul-Enein https://orcid.org/0000-0003-0249-7009 REFERENCES 1. Brown, W. A., Khan, A., Which depressed patients should receive antidepressants? CNS Drugs. 1994, 1, 341–347. https://doi.org/ 10.2165/00023210-199401050-00004. 2. Görög, S., Laukó, A., Herényi, B., Estimation of impurity profiles in drugs and related materials. J. Pharm. Biomed. Anal. 1988, 6, 697–705. https://doi.org/10.1016/0731-7085(88)80081-5 3. Rao, R. N., Raju, A. N., Nagaraju, D., Development and validation of a liquid chromatographic method for determination of enantiomeric purity of citalopram in bulk drugs and pharmaceuticals. J. Pharm. Biomed. Anal. 2006, 41, 280–285. https://doi.org/10.1016/j.jpba.2005.10.025 4. Sungthong, B., Jac, P., Scriba, G. K. E., Development and validation of a capillary electrophoresis method for the simultaneous determination of impurities of escitalopram including the R-enantiomer. J. Pharm. Biomed. Anal. 2008, 46, 959–965. https://doi.org/10.1016/j.jpba.2007.05.029 5. Dhaneshwar, S. R., Mahadik, M. V., Column liquid chromatography-ultraviolet and column liquid chromatography/mass spectrometry evaluation of stress degradation behavior of escitalopram oxalate. J. AOAC Int. 2009, 92, 138–147. 6. Ramana, B., Sharmaa, B. A., Ghugare, P. D., Nandavadekar, S., Singh, D., Karmuse, P. K., Kumar, A., Structural elucidation of process-related impurities in escitalopram by LC/ESIMS and NMR. J. Pharm. Biomed. Anal. 2010, 53, 895–901. https://doi.org/10.1016/j.jpba.2010.06.019 7. Kaleemullah, T., Ahmed, M., Sharma, H. K., Rajput, P., Reverse phase liquid chromatographic method for the quantification of dip-toluoyl-d-tartaric acid in escitalopram oxalate drug substance. Eurasian J. Anal. Chem. 2011, 6, 197–205. 8. Soliman, S. M., Youssef, N. F., Enantiomeric Thin-Layer Chromatographic Assay of Escitalopram in Presence of “In-Process Impurities”. J. Plan Chromatogr. 2011, 24, 475–481. 9. Dighe, V. V., Pawaskar, P., Adhyapak, S., Shambhu, N., Mestry, D., Development of normal phase chiral liquid chromatographic method for estimation of escitalopram oxalate and determination of R-citalopram enantiomer from escitalopram BADULLA ET AL. 16 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. oxalate in bulk drug and tablet. J. Chem. Pharm. Res. 2012, 4, 4804–4809. Deng, X., De Wolf, J., Vervoort, R., Pamperin, D., Adams, E., Schepdae, A. V., Development and validation of a capillary electrophoresis method for the determination of escitalopram and sensitive quantification of its enantiomeric impurity in formulations. Electrophoresis. 2012, 33, 1648–1651. Vaghela, B. K., Rao, S. S., Development and validation of stability indicating RP-LC, short runtime method for the estimation of escitalopram in escitalopram dosage form. World J. Pharm. Res. 2013, 2, 1018–1030. U.S. Pharmacopoeia-National Formulary (USP 39 NF 34), Volume 2. Escitalopram Oxalte monograph. (p. 3765–3770). USA: Book Press, Inc., Baltimore, MD 2016, http://www.usp.org/. Ng, L., Lunn, G., Faustino, P., in: R. J., Smith and M. L., Webb (Eds.), Organic impurities in drug substance: origin, control, and measurement in Analysis of Drug Impurities. Blackwell Publishing, UK 2007, pp. 12. Van Gyseghem, E., Van Hemelryck, S., Daszykowski, M., Questier, F., Massart, D. L., Vander Heyden, Y., Determining orthogonal chromatographic systems prior to the development of methods to characterize impurities in drug substances. J. Chromatogr. A. 2003, 988, 77–93. Inertsil ODS-3. http://www.glsciences.com/c-product/hplc/lccolumns/inertsil-ods-3. (accessed, December 8, 2016). Lupo, S., Carroll, F., Method development and column Selection: How the fluorophenyl phase provides the power of HILIC and reversed-phase modes in One Column. https://www.restek.com/ pdfs/GNAR2613-UNV.pdf (accessed, Nonember18, 2016). Gritti, F., Guiochon, G., Kinetic Core shell column. LCGC (North America) 2012, 30, 586–595. https://www.phenomenex. com/Kinetex/CoreShellTechnology. (accessed, December 2, 2016). Nucleosil standard C18 Phases. https://www.mnnet.com/HPLC Start/NUCLEOSILphases/NUCLEOSILC18/tabid/6128/language/ en-US/Default.aspx (accessed, October 9, 2016). https://www.drugs.com/sfx/escitalopram-side-effects.html. (accessed, December 23, 2016). ICH-Topic Q2 (R1), Validation of Analytical Procedures: Text and Methodology. Harmonized Tripartite Guideline, 2005, pp. 1–13. http://www.ich.org/pdfICH/Q2A. Vander Heyden, Y., Nijhuis, A., Smeyers-Verbeke, J., Vandeginste, B. G. M., Massart, D. L., Guidance for robustness: ruggedness tests in method validation. J. Pharm. Biomed. Anal. 2001, 24, 723–753. Kern, W. J., Kirkland, K., in: R. L. Snyder, J. Kirkland, J. L. Giajch (Eds.). Chiral separations in Practical HPLC method development. A Wiley-Interscience publication, New York 1997, pp. 562. Marin, A., Barbas, C., Systematic comparison of different functionality columns for a classical pharmaceutical problem. J. Pharm. Biomed. Anal. 2006, 40, 262–270. 24. Przybyciel, M., Santangelo, M. A., Paper Presented at the 54th Pittsburg Conference on Analytical Chemistry and Applied Spectroscopy. Orlanda, Florida, USA 2003, 914, 210–215. 25. Sun, M., Liu, D. Q., in: P. G. Wang and W. He (Eds.), Analysis of pharmaceutical impurities using hydrophilic interaction liquid chromatography in Hydrophilic Interaction Liquid Chromatography (HILIC) and Advanced Applications. CRC Press. Tayler and Francis Group, USA 2011, pp. 279. 26. Przybyciel, M., Novel Phases for HPLC Separations. Recent Developments in LC Column Technology. 2003, 2–5. 27. Yamamoto, F. M., Rokushika, S., Retention properties of the fluorinated bonded phase on liquid chromatography of aromatic hydrocarbons. J. Chromatogr. A. 2000, 898, 141–151. 28. Goss, J. D., Improved liquid chromatography of salicylic acid and some related compounds on a phenyl column. J. Chromatogr. A. 1998, 828, 267–271. 29. Petruczynik, A., Wróblewski, K., Dzioba, K., WaksmundzkaHajnos, M., Retention, separation selectivity and system efficiency of selected basic psychotropic drugs on different RPLC columns. Open Chem. 2015, 13, 943–950. 30. Wilson, N. S., Gilroy, J., Dolan, J. W., Snyder, L. R., Column selectivity in reversed-phase liquid chromatography VI. Columns with embedded or end-capping polar groups. J. Chromatogr. A. 2004, 1026, 91–100. 31. Vedantham, R., Vetukuri, V. P. R., Boini, A., Khagga, M., Bandichhor, R., Improved one-pot synthesis of citalopram diol and its conversion to citalopram. Org. Process Res. Dev. 2013, 17, 798–805. 32. Sharma, M., Jawa, P. R., Gill, R. S., Bansal, G., Citalopram hydrobromide: degradation product characterization and a validated stability-indicating LC-UV method. J. Braz. Chem. Soc. 2011, 22, 836–848. 33. Gangula, S., Kolla, N. K., Elati, C. R., Dongamanti, A., Bandichhor, R., Identification, synthesis and spectral characterization of impurities in process development of Escitalopram. GERF Bulletin of Biosciences. 2011, 2, 1–5. 34. Tadic, S., Nikolić, K., Agbaba, D., Development and optimization of an HPLC analysis of citalopram and its four nonchiral impurities using experimental design methodology. J. AOAC Int. 2012, 95, 733–743. How to cite this article: Badulla WFS, Can NÖ, Atkosar Z, Arli G, Aboul-Enein HY. Comparative Study of Different Chemistries and Particle Properties, High-Performance Liquid Chromatography Stationary Phases in Separation of Escitalopram Oxalate and Its Impurities in Different Pharmaceutical Dosage Forms. Sep Sci plus. 2019;1-16. https://doi.org/10.1002/sscp.201900021