G Model

CCLET 3663 1–7

Chinese Chemical Letters xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Chinese Chemical Letters

journal homepage: www.elsevier.com/locate/cclet

1
2 Original article

3 Quantification of flupirtine maleate polymorphs using X-ray

4 powder diffraction
5 Q1 Yu-Mei Zhao a,b, Zhi-Bing Zheng a,*, Song Li a,c,**
6 a
Laboratory of Computer-Aided Drug Design & Discovery, Beijing Institute of Pharmacology and Toxicology, Beijing 100850, China
7 b
Laboratory of Structure Identification, Beijing Institute of Pharmacology and Toxicology, Beijing 100850, China
8 c
State Key Laboratory of Toxicology and Medical Countermeasures, Beijing Institute of Pharmacology and Toxicology, Beijing 100850, China

A R T I C L E I N F O A B S T R A C T

Article history: Flupirtine maleate, a pharmaceutical compound for treating psychotic disease in clinics, has seven
Received 18 February 2016 polymorphs. Form A, with better crystal stability and bioavailability, has been widely used as the
Received in revised form 14 March 2016 pharmaceutical crystal form. Unfortunately, it is usually found in a polymorphic mixture with form B. In
Accepted 17 March 2016
this study, pure crystal forms of A and B were prepared and characterized by X-ray powder diffraction
Available online xxx
(XRPD), Fourier transform infrared spectroscopy (FT-IR) and thermal analysis. An XRPD–based method
for the quantitative determination of the amount of the flupirtine maleate polymorphs form A and form
Keywords:
B was also established through a systematic optimization of instrumental parameters. The results of the
Flupirtine maleate
X-ray powder diffraction
analytical methodology validation showed that the XPRD method had a broad quantitative range of 0–
Quantitative analysis of polymorphs 100% (w/w), good linear relationship, with R2 = 0.999, excellent repeatability and precision and low
Preferred orientation limits of detection (LoD) of 0.15% (w/w) and quantification (LoQ) of 0.5% (w/w). The results also showed
Transmission that the single-peak method was not as good as the whole pattern in reducing the influence of the
preferred orientation, but this can be compensated for by a systematic optimization of instrumental
parameters and validating the analytical methodology to reduce errors and obtain a good, repeatable,
sensitive, and accurate method. This XRPD method can be used to analyze mixtures of flupirtine maleate
polymorphs (forms A and B) quantitatively and control the quality of the bulk drug.
ß 2016 Chinese Chemical Society and Institute of Materia Medica, Chinese Academy of Medical Sciences.
Published by Elsevier B.V. All rights reserved.

9
10 1. Introduction Many analytical techniques, including infrared (mid- and near- 24
IR), FT-Raman, solid-state NMR spectroscopy, thermal methods, 25
11 Polymorphism creates challenges during drug development and X-ray powder diffraction (XRPD) [2–7] have been used to 26
12 and manufacturing because different polymorphs of a compound determine the polymorphic content of mixtures or the amorphous 27
13 have different physicochemical properties such as density, content of crystalline materials. However, XRPD has become the 28
14 morphology, solubility, dissolution rate, stability, and hygroscop- most preferred and extensively used technique for quantitatively 29
15 icity. In addition, sometimes different polymorphs of the same analyzing the purity of a polymorphic drug because of its 30
16 drug exhibit differences in bioavailability, efficacy, and drug advantages, including the uniqueness of the X-ray powder patterns 31
17 product performance in clinical situations. So the identification of different compounds, non-destructive nature, simplicity, and the 32
18 and specification of polymorphs has become an important part of ability to make the measurements of both the active ingredient and 33
19 the quality assurance process for pharmaceuticals [1]. In order to the final commercial product at room temperature [8–14]. Single- 34
20 control the polymorphic impurities of the final product, developing peak and whole pattern fitting are the primary quantification 35
21 an accurate quantification method for detecting low-level methods of XRPD. The single-peak method is suitable for the 36
22 polymorphic impurities in pharmaceuticals has become an quantitative analysis of crystals due to its advantages, such as 37
23 important aspect of drug development and manufacture. requiring less information about the sample in advance, its simplicity, 38
and its high sensitivity. But this method relies heavily on having a 39
pure standard sample and is dependent on the orientation of the 40
crystal. Therefore, the single-peak method often requires validation 41
* Corresponding author.
in practical applications. The whole pattern fitting method has a 42
Q2 ** Corresponding author at: Laboratory of Computer-Aided Drug Design &
Discovery, Beijing Institute of Pharmacology and Toxicology, Beijing 100850, China. higher signal to noise ratio (SNR), greater sensitivity, and a higher 43
E-mail addresses: zzbcaptain@aliyun.com (Z.-B. Zheng), lis@bmi.ac.cn (S. Li). level of specificity compared to the single-peak method, and it is not 44

http://dx.doi.org/10.1016/j.cclet.2016.03.042
1001-8417/ß 2016 Chinese Chemical Society and Institute of Materia Medica, Chinese Academy of Medical Sciences. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

2 Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx

45 dependent on the purity of a standard sample, and the influence of the samples in all the DSC experiments weighed between 2.35 and 95
46 orientation is also reduced. However, the application of whole 3.24 mg, with an accuracy of 0.01 mg. TGA: The thermogravimetric 96
47 pattern fitting method is still limited by its requiring prior measurements were performed using a Q500 TGA (TA, UK) system. 97
48 information about the sample’s structure. The mass loss of the sample as a function of temperature was 98
49 Flupirtine maleate (FPTM) (Fig. 1), 2-amino-3-carbethoxya- determined. 2.8590 mg of form A and 1.8480 mg of form B were 99
50 mino-6-(4-fluorobenzylamino) pyridine maleate, an antipsychotic weighed, respectively, with an accuracy of 0.0001 mg, separately 100
51 drug, has seven polymorphs. Forms A and B are the most common placed in an open alumina crucible, and then heated at a rate of 101
52 crystalline forms and usually coexist in a mixture [15]. Form A is 10 8C min 1 under nitrogen purge (60 mL min 1). And then a 102
53 the more stable anhydrous form at room temperature and is the recorded TGA spectrum was obtained. 103
54 one used as a medicine. Form B is unsuitable due to its (FT-IR): The FT-IR spectra for each of the FTPM forms were 104
55 metastability and is rapidly transformed into form A in concen- obtained by averaging 32 scans performed using a Thermo Nicolet 105
56 trated isopropanol suspensions or at higher temperature [16]. Thus, 6700 FT-IR spectrometer. About 2 mg of sample was gently ground 106
57 it is necessary to develop a simple, highly sensitive and accurate with 200 mg of KBr and pressed into a 13 mm-diameter pellet with 107
58 technical method for quantifying the amount of form B in a hydraulic press at 700 MPa for 20 s. The spectrum for each 108
59 polymorphic (forms A and B) mixtures of flupirtine maleate. To sample was recorded over the 4000–400 cm 1 spectral region at a 109
60 the best of our knowledge, the crystal structures of forms A and B of resolution of 4 cm 1. 110
61 FPTM have not been published. In this work, the single-peak-based XRPD: XRPD patterns for samples of different percentages of B/ 111
62 XRPD method was utilized to quantify the polymorphic forms of A were recorded at room temperature on a Bruker D8 Advance 112
63 FPTM (forms A and B). As mentioned before, an authentic and diffractometer (Karlsruhe, West Germany) that utilizes Cu Ka 113
64 validated single-peak-based XRPD calibration curve requires an radiation (1.54 Å) at 40 kV, 40 mA passing through a nickel filter 114
65 accurate identification and measurement of parameters, such as with a 0.58 variable slit, a 2.5 mm solar slit, and a 1 mm receiving 115
66 the intensity, height, and area of the diffraction lines, which is the slit to obtain both reflection and transmission measurements. The 116
67 most critical factor in developing any assay errors for solid-state diffractometer had a 2u compensating slit and the accuracy of the 117
68 forms. To reduce these errors, the instrument and sample peak positions was calibrated with a-Al2O3 as standard 118
69 preparation parameters, type of sample holders, sample rotation, sample. One hundred milligrams of the powder mixture was 119
70 particle size, and powder packing, all of which influence the loaded into the 0.2 mm deep hollow of an aluminum sample holder 120
71 quantification results by affecting the diffraction peak intensities, equipped with a quartz monocrystal zero background plate. To 121
72 areas, and balance, must be considered [9,13,14]. ensure a flat surface that was continuous with the holder surface, a 122
73 This study focused on three objectives: i) characterizing the clean glass slide was used to compress the sample into the hollow 123
74 inherent nature of samples using differential scanning calorimetry of the holder plate. The samples were analyzed by a continuous 124
75 (DSC), thermogravimetric analysis (TGA), Fourier transform mode X-ray powder diffraction analysis with a step size of 0.018 125
76 infrared spectroscopy (FT-IR), and X-ray powder detection (XRPD) and a step time of 0.6 s over an angular range of 4–168. During the 126
77 to test the purity and the choice of quantification methods; ii) measurements, the sample holder was rotated in the surface plane 127
78 optimizing the instrument and sample preparation parameters at 15 rpm. DIFFRACplus EVA (ver. 9.0) diffraction software was used 128
79 with the goal of minimizing the errors; and (iii) developing a to analyze the resulting diffractograms. 129
80 quantification calibration curve, which has been validated and
81 checked for assay errors, for quantifying the amount of form B in
82 polymorphic of FPTM using data obtained by XRPD. 3. Results and discussion 130

3.1. Solid-state characterization of crystal forms A and B 131

83 2. Experimental
3.1.1. Thermal analysis 132
84 2.1. Materials The DSC curve (Fig. 2) for form A showed a melting endotherm 133
at 164.88–168.37 8C, a subsequent recrystallization exotherm at 134
85 FTPM form A and FTPM form B were prepared and supplied by 169.86–170.2 8C, and a final melt at 178.44–179.81 8C. Form B only 135
86 the Hong de Pharmaceutical Co., Beijing and were used without had a single melting endotherm at 182.81–183.45 8C. The melting 136
87 any further purification. All other reagents and solvents obtained point and melting enthalpy are listed in Table 1. 137
88 from commercial suppliers were used as received.

89 2.2. Instrumentation

90 Thermal analysis: DSC: A differential scanning calorimeter

91 (DSC-Q2000; TA, UK) was used. The samples were heated from 40
92 to up to 200 8C at a heating rate of 10 8C
min 1 under a nitrogen
93 purge flow rate of 50 mL min 1. The temperature end point was
94 determined by the melting point of the less fusible component. The

Fig. 1. Flupirtine maleate. Fig. 2. DSC curves of forms A and B.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx 3

Table 1
1
Quantitative data of melting of flupirtine maleate polymorphs (10 8C min ).

Polymorph Te(8C) DH(J g 1)

A 178.44  0.5 109.9  3
B 182.81  0.5 119.8  3

138 In order to evaluate the reversible character of the thermal events

139 between form A and B, cyclic measurements were taken (Fig. 3). The
140 first step was to heat the FPTM at a rate of 10 8C min 1 to 167 8C (the
141 temperature corresponding to the end of the first endothermic peak)
142 or up to 169 8C (the temperature corresponding to the end of the
143 exothermic peak).The FPTM samples were then cooled to room
144 temperature and, finally, reheated to 190 8C. When a sample was
145 heated to 167 8C, an exothermic peak with a size that corresponded to
146 that of the exothermic peak present in the original rise in temperature
147 occurred immediately after the cooling began. However, when the
148 temperature reached in the first heating was 169 8C, no thermal event
Fig. 4. TGA curves of form A and form B.
149 occurred during the cooling period and in the second heating, only an
150 endothermic peak, which corresponds to the endothermic peak of
151 form B. These results indicate that form A transforms to form B at high
152 temperatures. So form A and form B are pure polymorphs.
153 The TGA analysis results indicated that FPTM crystal forms A
154 and B were anhydrous, as they exhibited no weight loss up to the
155 melting peak (Fig. 4).

156 3.1.2. FT-IR spectroscopy

157 The spectral regions between 3500–3100 cm 1 and between
158 1800–1000 cm 1 are important for the FT-IR identification of
159 forms A and B. Fig. 5 showed that the most characteristic peaks of
160 form A were at 3429, 1709, 1670, 1643, and 1527 cm 1, which
161 differed from those of form B (3319, 1695, 1655, and 1513 cm 1).
162 In particular, the spectral region between 3500–3100 cm 1
163 corresponds to the OH group or amino group; The spectral peak
164 at 1709 cm 1 corresponds to the carbonyl group; The spectral
165 region between 1500–1700 cm 1 corresponds to the aromatic
166 carbonyl group; Compared to that of form A, the FT-IR spectrum of
167 form B is red shifted .The most obvious reason is that the formation
168 of hydrogen bonds in form B. These results agree well with the
169 claims in US patent 2011/0184030 Al [15]. Fig. 5. FT-IR of form A and B.

170 3.1.3. The XPRD analysis

171 The XRPD patterns (Fig. 6) for form A and form B showed several the two crystal forms. The characteristic peaks for the two forms 173
172 differences in their characteristic peaks that can be used to identify are listed in Table 2 and corresponds well with the values reported 174
in the US patent 2011/0184030 Al [15]. 175

3.2. The quantitative analysis of flupirtine maleate polymorphs A and 176

B 177

3.2.1. The choice of a scanning mode 178

Two scanning modes of XRPD technology, reflection and 179
transmission, are commonly used in analyzing the purity of 180

Table 2
XRPD characteristic peaks of form A and form B.

Form 2u (8) d Rel. intensity (%) Area

A 6.87 12.84 24.4 50.60

9.16 9.60 6.7 13.56
12.39 7.14 27.6 51.85
17.82 4.97 100 200
18.46 4.80 13.9 27.94
B 5.43 16.35 100 51.56
7.2 12.26 4.6 2.51
10.84 8.15 27.8 14.90
14.46 6.12 79 64.76
14.99 5.90 75 48.79
Fig. 3. DSC curves of the modifications as well as of preheated (a) form A heated up
16.34 5.43 30.7 21.70
to 170 8C; (b) cooling to 80 8C;(c) preheated up to 190 8C; and (d) form B.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

4 Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx

Fig. 6. X-ray diffractograms of form A and B.

181 polymorphs. Errors due to the preferred orientation are the most relative intensity of each characteristic peak was more balanced 193
182 widely studied factor affecting XRPD. The mode of packing of the and reflective of the actual crystal structure. For form A, the 194
183 powder sample in the holder can lead to the crystals arranging in a resolution of the peaks was improved, especially between 208 (2u) 195
184 preferred orientation because of a non-random distribution of the and 308 (2u). However, the captured film used in the transmission 196
185 crystal orientations [9]. A crystallographic orientation of the crystal mode has a diffraction peak at 5.48, which coincides with the 197
186 particles can affect the intensity as much as 100%, and the characteristic peak of form B at 5.438/2u (I/I0 100%). Therefore, 198
187 consequences are more prominent for acicular and irregularly even though the reflection mode has more inherent difficulties 199
188 shaped crystals. In theory, using the transmission mode can with resolution, it is better suited for analyzing polymorphic 200
189 decrease the preferred orientation and improve the resolution. In FTPM. In order to find ways to overcome the disadvantages of the 201
190 this work, form A was initially studied using the transmission reflection mode, studies of the preparation of the sample, the 202
191 mode. Fig. 7 shows that the transmission mode reduced the optimization of the instrumental parameters, and the choice of 203
192 preferred orientation and increased the resolution. Thus, the quantifiable characteristic peaks are necessary. 204

Fig. 7. X-ray reflection and transmission diffractograms of forms A and B.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx 5

205 3.2.2. Preparation of polymorphic mixtures for XRPD study mixtures. The other characteristic peaks at 7.28, 10.848, 14.468, 240
206 Before mixing, the two forms of FTPM were passed through a 14.998, and 16.348 2u were low in resolution and intensity and 241
207 200-mesh sieve to reduce the effect of particle size on the preferred exhibited low R2 values (less than 0.926) and, therefore, were not 242
208 orientation [9]. A mixed sample of forms A and B was prepared by used in the quantification process. 243
209 accurately weighing out specific amount of the previously ground Form A had five characteristics peaks at 6.878, 9.168, 12.398, 244
210 pure form A (200-mesh) after which, form B was physically mixed 18.468, and 17.828 2u. Choosing the quantitative characteristic 245
211 into form A in an agate mortar with a pestle to form separate peaks was important. The peak at 6.878 2u could not be selected for 246
212 mixtures with the following percentages of form B/form A: 0, 0.5, 1, the quantification because the peak at 6.878 2u (from 6.458 to 247
213 2, 5, 10, 15, 20, 25, 30, 40, 60, 80, and 100 w/w%. A qualitative 7.188) overlapped with the peak at 7.28 2u (from 6.98 to 7.58) of 248
214 crystal study of FTPM by XRPD showed that grinding did not form B. The other low area and intensity peak (9.168, 18.468 2u) and 249
215 induce a polymorph transformation between the crystal forms A the high area and intensity peak (17.628 2u) exhibited low R2 250
216 and B. The samples were made in triplicates and 100 mg powder values (less than 0.914) and, thus, were not used for quantification. 251
217 mixtures with an accuracy of 0.01 mg were loaded into a sample Therefore, for form A, the peak at 12.398 2u (I/I0 = 27.6%, 252
218 holder. area = 51.85), which exhibited higher R2 values (0.999), was 253
selected to quantify the amount of form A in mixtures. The choice 254
219 3.2.3. Optimization of instrumental parameters showed that if the area of the peak was closer to that of the 255
220 Earlier studies reported that the scan rate and chopper increment characteristic peak of form B, the linear correlation of the 256
221 critically affect the area of the diffraction peaks [10].In this study, six calibration curve was better. 257
222 groups with different scan rates and chopper increments were
223 investigated using a 5%, w/w polymorphic mixture of form A and 3.2.5. Calibration curve development 258
224 form B. Fig. 8 showed that when the scan rate was 128 min 1, only In general, the height and area of peaks are both considered for 259
225 one peak was identified at 5.438, whereas when the scan rates were analysis. Peak shape and height vary more with changes in particle 260
226 2, 1, 0.6 and 0.38 min 1, a maximum of four peaks of form B were size, but the peak area tends to be less variable [13,14]. Therefore, 261
227 identified at 2u 5.43o, 10.84o, 14.46/14.99o, and 16.308. Because the in this study, peak area was used for the quantitative analysis. The 262
228 7.28 and 13.358 2u peaks were of very low intensity (I/I0 = 6.6%, peak area at 5.438 2u was useful for monitoring the amount of form 263
229 12.3%), they were not observed in any of the combinations. B in the samples. Fig. 9 showed that the calibration curve was quite 264
230 Therefore, in order to keep a balance between the peak resolution linear over a wide range (0–100%, w/w) of form B, with a linear 265
231 and recording time, a scan rate of 28 min 1 with about 21.5 min equation of y = 0.987  0.711 and a high correlation coefficient of 266
232 recording time was selected for the experiments. 0.999. These results confirmed that XRPD is a very good method for 267
quantifying FPTM mixtures of polymorphs (forms A and B). 268
233 3.2.4. The choice of quantifiable characteristic peaks
234 In previous studies of the quantification of polymorphic forms 3.2.6. Validation of the analytical method 269
235 of drugs by XRPD, the highest intensive peak (I/I0 = 100%) was The analytical method developed for quantifying the amount of 270
236 usually used to detect the amount of the different polymorphs in FPTM form B in form A was validated using accuracy, precision, 271
237 polymorphic mixtures [17]. In this study, FTPM form B had six ruggedness, LoD, and LoQ [10,17,18]. The results are provided in 272
238 characteristic peaks, so the highest peak (5.438 I/I0 = 100%, Table 3. The accuracy of the result was confirmed by an average 273
239 area = 51.56) was used to detect the polymorphic content of the recovery of 97.0–102.0%, and the precision was demonstrated in 274

Fig. 8. Effect of scan rate on X-ray diffractogram of 5% (w/w) of form B. (a) 128 min 1, (b) 68 min 1, (c) 28 min 1, (d) 18 min 1, (e) 0.68 min 1
, and (f) 0.38 min 1
with the solid
line boxes depicting the identifiable peaks and the dotted boxes showing the peaks that were not detected in 5% (w/w) mixture.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

6 Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx

100 Table 4
Assay error evaluation for 0–100% (w/w).

Assay error Validation data

y=0.987x-0.711
80 2
R =0.999 Instrument reproducibility (% R.S.D.) 0.048
Inter-day variability (% R.S.D.) 0.64
Intra-day reproducibility (% R.S.D.) 0.05
60 Sample positioning (% R.S.D.) 0.54
Sample packing (% R.S.D.) 1.74
Area

40
excellent, we also investigated the impacts of potential sources of 285
error, as reported in the next section. 286
20
3.2.7. Estimation of assay errors and sources of error 287
Many potential assay errors, including instrument, inter-day, 288
0 intra-day, sample position, and sample packing errors, which are 289
0 20 40 60 80 100 associated with an XRPD quantitative analysis [14]. So, all the assay 290
% w/w form B in form A errors were investigated in this work. In order to determine the 291
effect of these errors on the quantification of forms A and B, a single 292
Fig. 9. Calibration curve for determination of FTPM polymorphic form B in form A by
mixture (5 wt% of form B) was used to calculate the size of error. 293
XRPD.
The data from the assay error evaluation are shown in Table 4. 294
The instrument reproducibility was tested by repeatedly 295
Table 3 measuring a single sample without disturbance throughout the 296
Validation parameters.
whole assay process. The results showed good reproducibility with 297
Validation parameters Validation data an RSD value of 0.048%. We also found a low intra-day variation of 298
Recovery (%) 97–102% 0.05% and a higher inter-day variation of 0.64%. The day-to-day 299
Precision (% R.S.D.) 1–2% variability seemed to have resulted from a combination of several 300
R2 of correlation curve 1 factors, such as the daily removal of the sample from the 301
Slope of correlation curve 0.999 instrument and the resulting re-positioning of the sample each 302
Intercept of correlation curve 0.0324
day, which in combination may have disturbed the sample. The 303
LoD (%) 0.15
LoQ (%) 0.5 observed variability from changing the sample position and re- 304
packing the same sample was 0.54% and 1.74%, respectively. 305
Taking into account of the potential sources of error investigated in 306
this study, sample packing seemed to be the most critical factor 307
275 that the %RSD was between 1.0% and 2.0%. The detection and influencing the accuracy of the quantitative analysis. This, in turn, 308
276 quantitation limits were calculated using ICH guidelines with implies that the geometric orientation, especially the adopting of a 309
277 consideration of both the background and blank responses to preferred orientation by the crystals, may be the most important 310
278 determine the noise level. The results showed that the linear source of error, while sample particle size, sample rotation, and 311
279 relationship covered the whole range (0–100%), and the LoD and transmission mode can reduce the probability of the crystals 312
280 LoQ were 0.15% and 0.5%, respectively. The curve showing the adopting a preferred orientation a process that can aid in reducing 313
281 relationship between the actual and the predicted content (%, w/w) the assay error. 314
282 of form B was plotted (Fig. 10) with an R2 value of 1, a fitted slope of
283 0.999, and a small intercept of 0.0324. Although the correlation 4. Conclusion 315
284 between the actual and calculated concentrations of form B was
An XRPD method for quantifying FPTM polymorphs was 316
established by systematically optimizing the instrumental param- 317
100
eters and validating the analytical methodology. The reflection- 318
scanning mode was found to be the more practicable XRPD method 319
predicted concentration of form B % w/w

y=0.999x-0.0324 for quantifying FPTM polymorphs. A 200-mesh sample particle 320

80
size and a scan rate of 28 min 1 were selected in determining the 321
R2=1 proportions of the polymorphs in the FPTM mixtures. The 322
calibration curve was found to be a linear fit across the entire 323
60
range from 0–100% (w/w) with an LoD as low as 0.15% and an LoQ 324
of 0.5%. Although errors in assays can result from a multiplicity of 325
factors, such as instrument variability or other instrument 326
40
difficulties, inter-and intra-day variability, and/or sample packing 327
issues, a systematic optimization of the instrumental parameters 328
reduced the size of the errors. This systematic validation of the 329
20
analytical methodology strongly indicated that the determination 330
of the content of the FPTM mixtures was reliable. Therefore, this 331
method was confirmed as an effective and practical method for the 332
0
0 20 40 60 80 100 quantitative determination of FPTM polymorphs and could be used 333
in the quality control of polymorphic FPTM in bulk drug samples to 334
Actual concentration of form B % w/w
ensure the clinical therapeutic effects of FPTM. The strategies for 335
Fig. 10. Correlation curve of observed vs. theoretical percentage of form B in form A, systematically optimizing the instrumental parameters outlined in 336
obtained using peak areas in XRPD. this article may be helpful in resolving similar technical problems 337

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042
 G Model
CCLET 3663 1–7

Y.-M. Zhao et al. / Chinese Chemical Letters xxx (2016) xxx–xxx 7

338 in the quantification of the polymorphic forms of other drugs by [7] R.D. Vickery, G.A. Nemeth, M.B. Maurin, Solid-state carbon NMR characterization 363
of the polymorphs of roxifiban, J. Pharm. Biomed. Anal. 30 (2002) 125–129. 364
339 XRPD. [8] K.H. Hestnes, B.E. Sørensen, Evaluation of quantitative X-ray diffraction for 365
possible use in the quality control of granitic pegmatite in mineral production, 366
340 Acknowledgments Miner. Eng. 39 (2012) 239–247. 367
[9] S.N.C. Roberts, A.C. Williams, I.M. Grimsey, S.W. Booth, Quantitative analysis of 368
mannitol polymorphs. X-ray powder diffractometry—exploring preferred orien- 369
341 This research was supported by the Major Program of Ministry of tation effects, J. Pharm. Biomed. Anal. 28 (2002) 1149–1159. 370
342 Science and Technology of China (No: 2015ZX09J15104-003002). [10] M. Tiwari, G. Chawla, A.K. Bansal, Quantification of olanzapine polymorphs using 371
powder X-ray diffraction technique, J. Pharm. Biomed. Anal. 43 (2007) 865–872. 372
343 The authors would like to acknowledge the English and content [11] L. Alexander, H.P. Klug, Basic aspects of X-ray absorption in quantitative diffrac- 373
344 editing of Rhoda E. and Edmund F. Perozzi, PhDs. tion analysis of powder mixtures, Anal. Chem. 20 (1948) 886–889. 374
[12] X.M. Chen, J.G. Stowell, K.R. Morris, S.R. Byrn, Quantitative study of solid-state 375
345 References acid-base reactions between polymorphs of flufenamic acid and magnesium 376
oxide using X-ray powder diffraction, J. Pharm. Biomed. Anal. 51 (2010) 377
866–874. 378
346 [1] M. Saifee, N. Inamda, D. Dhamecha, A. Rath, Drug polymorphism: a review, Int. J. [13] V.J. Hurst, P.A. Schroeder, R.W. Styron, Accurate quantification of quartz and other 379
347 Health Res. 2 (2009) 291–306. phases by powder X-ray diffractometry, Anal. Chim. Acta 337 (1997) 233–252. 380
348 [2] Y. Li, P.S. Chow, R.B.H. Tan, Quantification of polymorphic impurity in an enantio- [14] R. Takehira, Y. Momose, S. Yamamura, Quantitative analysis of crystalline phar- 381
349 tropic polymorph system using differential scanning calorimetry, X-ray powder maceuticals in tablets by pattern-fitting procedure using X-ray diffraction pat- 382
350 diffraction and Raman spectroscopy, Int. J. Pharm. 415 (2011) 110–118. tern, Int. J. Pharm. 398 (2010) 33–38. 383
351 [3] N. Chieng, S. Rehder, D. Saville, T. Rades, J. Aaltonen, Quantitative solid-state [15] M. Ernest, M.D. Bogdanovic, A. Danilovski, et al., Pharmaceutically acceptable salt 384
352 analysis of three solid forms of ranitidine hydrochloride in ternary mixtures using and polymorphic forms of flupirtine maleate, US 2011184030 A1 385
353 Raman spectroscopy and X-ray powder diffraction, J. Pharm. Biomed. Anal. 49 [16] K.F. Landgraf, A. Olbrich, S. Pauluhn, et al., Polymorphism and desolvation of 386
354 (2009) 18–25. flupirtine maleate, Eur. J. Pharm. Biopharm. 46 (1998) 329–337. 387
355 [4] V. Uvarov, I. Popov, Development and metrological characterization of quantita- [17] A. Nageswari, K.V.S.R. Reddy, K. Mukkanti, Stability-indicating UPLC method for 388
356 tive X-ray diffraction phase analysis for the mixtures of clopidogrel bisulphate determination of imatinib mesylate and their degradation products in active 389
357 polymorphs, J. Pharm. Biomed. Anal. 46 (2008) 676–682. pharmaceutical ingredient and pharmaceutical dosage forms, J. Pharm. Biomed. 390
358 [5] J. Ibáñez, O. Font, N. Moreno, et al., Quantitative Rietveld analysis of the crystalline Anal. 66 (2012) 109–115. 391
359 and amorphous phases in coal, Fuel 105 (2013) 314–317. [18] N. Vadera, G. Subramanian, P. Musmade, Stability-indicating HPTLC determina- 392
360 [6] F.L.F. Soares, R.L. Carneiro, Evaluation of analytical tools and multivariate methods tion of imatinib mesylate in bulk drug and pharmaceutical dosage form, J. Pharm. 393
361 for quantification of co-former crystals in ibuprofen-nicotinamide co-crystals, J. Biomed. Anal. 43 (2007) 722–726. 394
362 Pharm. Biomed. Anal. 89 (2014) 166–175.

Please cite this article in press as: Y.-M. Zhao, et al., Quantification of flupirtine maleate polymorphs using X-ray powder diffraction,
Chin. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.cclet.2016.03.042