Research article

Received: 14 March 2014 Revised: 1 May 2014 Accepted: 30 May 2014 Published online in Wiley Online Library: 18 June 2014

(wileyonlinelibrary.com) DOI 10.1002/mrc.4098

480

Quantitative determination and validation ofavermectin B1a in commercial products usingquantitative nuclear magnetic resonancespectroscopyZhuoni Hou, Xianrui Liang, Liping Du, Feng Su and Weike Su*

Nuclear magnetic resonance is defined as a quantitative spectroscopic tool that enables a precise determination of the num-ber of substances in liquids as well as in solids. There is few report demonstrating the application of NMR in the quantificationof avermectin B1a (AVB1a); here, a proton nuclear magnetic resonance spectroscopy (1H NMR) using benzene [1-methoxy-4-(2-nitroethyl) (PMN)] as an internal standard and deuterochloroform as an NMR solvent was tested for the quantitative deter-mination of AVB1a. The integrated signal of AVB1a at 5.56ppm and the signal of PMN at 8.14ppm in the 1H NMR spectrumwere used for quantification purposes. Parameters of specificity, linearity, accuracy, precision, intermediate precision, range,limit of detection (LOD), limit of quantification (LOQ), stability and robustness were validated. The established method wasaccurate and precise with good recovery (98.86%) and relative standard deviation (RSD) of assay (0.34%) within the linearityof the calibration curve ranging from 5.08 to 13.58mg/ml (R2 = 0.9999). The LOD and LOQ were 0.009 and 0.029mg/ml, whichindicated the excellent sensitivity of the method. The stability of the method was testified by a calculated RSD of 0.11%. Therobustness was testified by modification of four different parameters, and the differences among each parameter were all lessthan 0.1%. Comparing with the assay described by the manufacturer of avermectin tablets, there was no significant differencebetween the assay obtained by HPLC and quantitative NMR (qNMR), which indicated qNMR was a simple and efficient methodfor the determination of AVB1a in commercial formulation products. Copyright © 2014 John Wiley & Sons, Ltd.

Keywords: NMR; 1H NMR; quantitative NMR; avermectin B1a; drug validation; commercial products; quality control

* Correspondence to: Weike Su, Key Laboratory for Green Pharmaceutical Tech-nologies and Related Equipment of Ministry of Education, College of Pharma-ceutical Sciences, Zhejiang University of Technology, Hangzhou 310014, China.E-mail: pharmlab@zjut.edu.cn

Key Laboratory for Green Pharmaceutical Technologies and Related Equip-ment of Ministry of Education, College of Pharmaceutical Sciences, ZhejiangUniversity of Technology, Hangzhou 310014, China

Introduction

In the pharmaceutical industry, nuclear magnetic resonance(NMR) spectroscopy can be generally exploited for quantitativeanalysis owing to the proportional relationship between theNMR signal peak area and the resonant nuclei number. Com-pared with the traditional quantitative methods, quantitativeNMR (qNMR) spectroscopy shows the following advantages: (i)no calibration standard of the analyte is needed, (ii) high selectiv-ity can be obtained under appropriate acquisition conditions, (iii)more than one analyte in a mixture can be determined simulta-neously, (iv) measuring time is relatively shortened and (v) sam-ple preparation is easy. For the first time since 1963 that qNMRwas presented in the literature issued by Forbes[1] and Hollis,[2]

a great deal of application reports for qNMR have been pub-lished.[3–5] It was particularly worth mentioning that 1H NMR asa precise quantitative tool with a validated protocol was an-nounced by Malz and Jancke in 2005.[6] Liang (2014) developeda qNMR method for determining the content of clindamycinphosphate tablets, which was approximately 45%.[7] Therefore,in order to prove that qNMR can also be applied for determina-tion of low-content drug tablets, an attempt is made to testavermectin B1a (AVB1a) tablets whose content is less than 5%.Avermectin B1a is a macrolide antibiotics from Streptomyces

avemitilis.[8] It is widely employed in agriculture according toits efficient broad spectrum anti-parasitic activity by affectingon glutamate-gated chloride channel of worms.[9] With the

Magn. Reson. Chem. 2014, 52, 480–485

increasing applications of avermectin in agriculture,[10–15] it isinteresting to develop an accurate and sensitive quantitativemethod for residues determination of the environment and theassay of commercial products. Few analytical techniquesincluding GC, LC, HPLC method with fluorescence detectionand LC/MS have been successfully applied for AVB1a determina-tions.[16–20] However, to the best of our knowledge, the applica-tion of qNMR on AVB1a has not been notified so far. Hence, theaim of this study is to develop an advantageous, competitiveand selective NMR method for the assay determination of AVB1ain commercial products.

Experimental

Reagents and standards

Avermectin B1a (99.6%, HPLC) was provided from Shenghua BIOKCo. Ltd., (Zhejiang, China). 1-Methoxy-4-(2-nitroethyl) benzene

Copyright © 2014 John Wiley & Sons, Ltd.

An effective method utilized on AVERMECTIN B1a assessment

(PMN; 99.5%, HPLC) was supplied from Zhejiang Hisoar Pharma-ceutical Co., Ltd and used as the internal standard.Deuterochloroform (CDCl3; 99.8% with 0.03% TMS) used forNMR analysis was purchased from Qingdao Tenglong WeiboTechnology Co., Ltd., (Qingdao, China). Avermectin (AV) tablets(uncoated, 120 ± 3.0mg per tablet) were purchased from PekingZhongnongda Biotechnique Co. Ltd., (Peking, China).

Instrument

All 1H NMR spectra were obtained using a Varian Mercury plus400 spectrometer (Varian, CA, USA) operating at 400.13MHzequipped with 5mm ID probe and 5mm ASW probe. An Agilent1200 HPLC (Agilent, CA, USA) equipped with a G1314B UV–visdetector was also applied in this study for verification purposes.

Preparation of test solutions

Internal standard solution preparation: 5.21mg of PMN was accu-rately weighed and transferred into an NMR tube with 0.60ml ofCDCl3; standard solution preparation: AVB1a and PMN (1 : 1, m/m)were weighed accurately and transferred into an NMR tube with0.60ml of CDCl3; sample solution preparation: the powder of AVtablets and PMN (1 : 1, m/m) was weighed accurately and trans-ferred into an NMR tube with 0.60ml of CDCl3. All the solutionswere sonicated at room temperature until completely dissolved.

Data acquisition and processing

Typical 1H NMR experimental parameters were shown as follows:32 scans were collected for each sample with 39 840 data pointsby using a 30° pulse length; the spectral width 3906.20 Hz; acqui-sition time 3.5 s; a relaxation delay of 30 s that was more than fivetimes the longest T1 was employed to ensure full T1 relaxation ofprotons; an exponential line-broadening function of 0.3 Hz wasapplied to FID. The temperature was regulated at 298 K. All thedata processing was performed by using MestReNova.

Phase and baseline corrections were performed manually. Thismanual mode was set also for the signal integration [choice of in-tegration limits (generally without the 13C satellites) and, if it wasnecessary, the BIAS and SLOPE functions were utilized to recalcu-late the integrals owing to improper baseline corrections].Furthermore, each measurement was repeated five times for sta-tistical purposes. Chemical shifts were referenced internally bythe TMS signal obtained at 0.00 ppm. The integrated signal ofAVB1a was obtained at 5.56 ppm (doublet of doublet, H-23) withrespect to the signal of PMN at 8.14 (doublet, H-3 and H-5) inthe 1H NMR spectrum. The structures of AVB1a and PMN areshown in Fig. 1.

Figure 1. Structure of avermectin B1a and 1-methoxy-4-(2-nitroethyl) benze

Magn. Reson. Chem. 2014, 52, 480–485 Copyright © 2014 John

Quantitative NMR analysis

The most important fundamental relation of qNMR is signal re-sponse (integrated signal area) Ix in a spectrum that is directlyproportional to the number of nuclei Nx generating the corre-sponding resonance line:[6,21,22]

Ix ¼ KsNx (1)

Ks is a spectrometer constant that is unknown, whereas it is aconstant for all resonance lines in the same 1H single-pulseNMR spectrum. Accordingly, the determination of relative arearatios Ix/Iy is the most efficient way to obtain quantitative resultsby using Eqn (2) when Ks cancels for the ratio:

IxIy¼ Nx

Ny(2)

For the purity determination of a substance, an internal stan-dard with known purity is needed. On the basis of Eqn (2), thecomponent purity can be calculated from the NMR intensity viaEqns ((3) and (4)):

WX ¼ IX�NStd�MX�mStd

IStd�NX�MStd(3)

PX ¼ IX�NStd�MX�mStd�PStdIStd�NX�MStd�m

(4)

WX and PX represent the mass and purity of the analyte.MX andMStd are the molar masses of the analyte (873.09 g/mol) and thestandard (181.19 g/mol). m is the weighed mass of the investi-gated sample. mStd and PStd are the weighed mass and the purityof the standard (99.5%). NStd and IStd correspond to the number ofprotons for the standard (2.0) and the integrated signal area of atypical NMR line obtained at 8.14 ppm. Nx and Ix correspond tothe number of protons for the analyte 1H in drug (1.0) and the in-tegrated signal area of analyte obtained at 5.56 ppm.

Results and Discussion

Determination of relaxation time (T1)

As a necessary prelude to all the measurements, analyte and in-ternal standard (AVB1a and PMN) were analyzed qualitatively byroutine 1H experiments to determine longest spin–lattice relaxa-tion time. The T1 relaxation is described by Eqn (5):

ne.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/mrc

481

Z. Hou et al.

482

Mz ¼ M0 1� e� τ=T1ð Þ� �

(5)

Mz andM0 are the magnetization along the z-axis after the relaxationdelay τ and at thermal equilibrium, respectively. The value of T1 wasmeasured by the inversion recovery pulse sequence experiment forall the protons of AVB1a and PMN. The longest T1 of AVB1a andPMN were 3.103 and 4.069 s, respectively. The delays of 5T1 are re-quired to reestablish complete recovery after full excitation of NMRresponse. Accordingly, the relaxation delay of 30swas sufficient to en-sure full T1 relaxation between pulses for the determination of AVB1a.

Validation

Specificity and selectivity

The specificity and selectivity are key prerequisites that must beevaluated for avoiding possible interference according to thepresence of the components in the sample solution includingsolvent and the excipients in the pharmaceutical formulations.Figure 2 manifests 1H NMR spectra of internal solution, drug stan-dard solution, placebo solution and sample solution individually.It was obvious that the signals obtained at 5.56 and 8.14 ppmfor AVB1a and PMN were not disturbed by solvent and excipients.Moreover, the signals of AVB1a and PMNwere well separated fromeach other in standard and sample preparations.

Linearity

The intensity of the response signal is directly proportional to theamount of nuclei, as described in Eqn (1). To check the linearity ofthe method, linearity tests were performed by standard solutionof AVB1a in different concentration ratios (from 5.08 to13.58mg/mL). Consequently, the linearity regression yielded acorrelation coefficient of 0.9999 and a regression Eqn (6):

Y ¼ 1:002Xþ 0:036 (6)

where X is themass of AVB1a (true weight) and Y is the mass of ob-tained AVB1a (calculated by Eqn (3)).

Figure 2. 1H NMR spectra of internal standard solution (a), drug standard s

wileyonlinelibrary.com/journal/mrc Copyright © 2014 Joh

Accuracy

The accuracy of an analytical method is described as the correla-tion of a theoretical value and the value found. The recom-mended method is to utilize a minimum of nine determinationsover a minimum of three concentration levels. In this study,80%, 100% and 120% levels with respect to the sample were cho-sen in triplicate at each level. The accuracy was estimated bymeasuring the mean recovery and relative standard deviation(RSD). The recovery was determined by comparing the calculatedvalue of added reference material and the true value via Eqn (7):

Recovery %ð Þ ¼ mx �mo

ms�100% (7)

where mx is the calculated mass of analyte after reference addi-tion, mo is the calculated mass of original analyte before refer-ence addition and ms is the true mass of reference addition.

According to the results exhibited in Table 1, it was concludedthat the method for assay content was accurate between theranges of selected levels. The mean recovery was 98.86%, andRSD was 0.34%.

Precision and intermediate precision

Precision is defined as the degree to which repeatedmeasurementsunder unchanged conditions demonstrate the same results. Inter-mediate precision expresses variations within laboratories. Theyare evaluated by RSD for repeatability. Otherwise, the SNR plays asignificant role in the precision of experimental measurements. Inthe previous research, an SNR of at least 150 : 1 is necessary for everyresonance line. ICH (The International Conference onHarmonisationof Technical Requirements for Registration of Pharmaceuticals forHuman Use) guidelines recommend that the precision should beassessed by using six repeated determinations. In this study, theprecision was assessed by six separate sample preparations, andintermediate precision was evaluated by a second analyst on analternative probe (5mm ASW probe, Varian, Peking, China). Thecalculated content of drug in mg and assay for each sample and

olution (b), placebo solution (c) and sample solution (d).

n Wiley & Sons, Ltd. Magn. Reson. Chem. 2014, 52, 480–485

Table 1. Accuracy test results

Accuracy level AV tablet (mg) m PMN (mg) mstd Obtained AVB1a (mg) Wx Assay (%) Px Recovery (%)

80% Set 1 129.76 1.14 4.86 3.73 98.99

80% Set 2 131.24 0.98 4.94 3.75 98.63

80% Set 3 129.66 1.20 4.88 3.74 99.57

100% Set 1 161.49 1.31 6.07 3.74 99.62

100% Set 2 161.79 1.30 6.05 3.72 98.94

100% Set 3 161.99 1.55 6.07 3.73 98.05

120% Set 1 194.94 1.50 7.30 3.73 99.21

120% Set 2 194.62 1.63 7.27 3.72 98.79

120% Set 3 195.28 1.48 7.28 3.71 97.92

Mean 3.73 98.86

SD 0.01

RSD (%) 0.34

AV, avermectin; AVB1a, avermectin B1a; SD, standard deviation; RSD, relative standard deviation.

Table 2. Precision and intermediate precision test resultsa

Study Precision Intermediate precision

Preparation AV tablet(mg)

Internal standard(mg)

Obtained AVB1a(mg)

Assay(%)

AV tablet(mg)

Internal standard(mg)

Obtained AVB1a(mg)

Assay(%)

1 161.80 1.31 6.06 3.73 161.95 1.23 6.10 3.75

2 161.80 1.38 6.05 3.72 161.58 1.36 6.05 3.73

3 161.75 1.38 6.06 3.73 161.54 1.35 6.05 3.73

4 161.80 1.22 6.08 3.74 161.81 1.42 6.08 3.74

5 162.13 1.18 6.09 3.74 161.63 1.42 6.09 3.75

6 162.15 1.20 6.10 3.74 161.51 1.25 6.05 3.73

Mean 3.73 Mean 3.74

SD 0.01 SD 0.01

RSD (%) 0.21 RSD (%) 0.27

AV, avermectin; AVB1a, avermectin B1a; SD, standard deviation; RSD, relative standard deviation.aPrecision studies were determined on 5mm ID probe, and intermediate precision studies were tested on an alternative NMR probe (5mm

ASW probe), by a different analyst on different days.

An effective method utilized on AVERMECTIN B1a assessment

483

statistical results are shown in Table 2; there was no significantdifference between results of precision and intermediate precision.

Range

Range testing aims to assess the saturated concentration of theanalyte in solution. The saturated solution was prepared byadding excess AVB1a standard amount and analyzing super-natant solution for determining the dissolved concentrationof analyte. The saturation concentration was 54.74mg per0.60ml CDCl3.

LOD and LOQ

Limit of detection is the lowest concentration of analyte in a sam-ple that is adequate for detected but not necessarily quantified.LOQ is the lowest concentration of analyte in a sample, whichcan be quantitatively determined with suitable precision and ac-curacy. They are two fundamental elements of method validationdefining the limitations of an analytical method. In qNMR, theLOD and LOQ cannot be determined by means of SNR becauseof the fact that the NMR response signals are Lorentzian lines.Hence, the LOD and LOQ are calculated by the standard

Magn. Reson. Chem. 2014, 52, 480–485 Copyright © 2014 John

deviation of the response σ (0.003, calculated by results of threeblank samples) and the slope S (1.002) of a calibration curve ob-tained in the linearity study by using Eqns ((8) and (9)):

LOD ¼ 3:3σ

S(8)

LOQ ¼ 10σ

S(9)

LOD and LOQ were calculated to be 0.009 and 0.029mg/mL,respectively.

Stability

Stability testing is crucial for estimating the allowed time lapsebetween sample collection and analysis. It is also essential toevaluate whether the results from test related to the original sam-ple are meaningful or not. The solution is stable when the differ-ence in assay compared with initial value is not more than 1.0%and the RSD less than 3.0%. If the solution is not stable at roomtemperature, a same study repeated at refrigeration temperature(~2–8 °C) is recommended.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/mrc

Table 3. Results for stability study

Standard solution Sample solution

Time (h) Obtained AVB1a (mg) Assay (%) Difference (%) Obtained AVB1a (mg) Assay (%) Difference (%)

0 6.00 99.38 NA 6.13 3.77 NA

8 6.00 99.38 0.00 6.14 3.78 0.01

16 6.00 99.39 0.01 6.13 3.77 0.00

24 6.00 99.32 0.06 6.13 3.77 0.00

RSD (%) 0.03 0.11

AVB1a, avermectin B1a; RSD, relative standard deviation.

Z. Hou et al.

484

The sample preparation and standard solution were analyzedat ambient temperature (~25 °C) at 0-h (initial), 8-h, 16-h and24-h intervals, and assay was calculated for each interval. The dif-ference and RSD of both sample and standard solutions wereshown in Table 3, which means no major change was observedin both solutions during the testing period.

Robustness

ICH defines the robustness of an analytical procedure as a mea-surement of its capacity to remain unaffected by small but delib-erate variations in method parameters listed in the proceduredocumentation. It supplies an indication of its suitability duringnormal routine usage. The robustness of the method was esti-mated by varying four parameters independently: (i) the numberof scans, (ii) the relaxation delay, (iii) the acquisition time and (iv)different analyte proton chosen. The result expressed that thismethod was quite robust in terms of the aforementioned param-eters (Table 4).

Comparison with HPLC method

Assay results obtained by qNMR were also confirmed by compar-ing with HPLC techniques. The HPLC analysis was performed withan Extend C18 column (5μm, 250mm×4.6mm i.d., Agilent, USA).The mobile phase consisted of methanol, acetonitrile and waterin the ratio of 41 : 41 : 18 (v/v) with the flow rate of 1.0ml/min.

Table 4. Results for robustness study

Parameters Change ObtainedAVB1a (mg)

Assay(%)

Differencea

(%)

Number of scans 48 6.12 3.76 0.02

64 6.13 3.77 0.01

80 6.12 3.77 0.01

Relaxation delay 40 s 6.09 3.75 0.03

50 s 6.12 3.77 0.01

60 s 6.10 3.75 0.03

Acquired time 4 s 6.14 3.78 0.00

5 s 6.12 3.77 0.01

6 s 6.13 3.77 0.01

Analyte protons 7.04 ppm 6.00 3.69 0.09

8.14 ppm 6.08 3.74 0.04

AVB1a, avermectin B1a.aThe experiment was carried out with AV tablet solution, and the

differences were obtained by comparing with the manufactures

quoted content of AVB1a (3.78%).

wileyonlinelibrary.com/journal/mrc Copyright © 2014 Joh

The detection wavelength was set at 245 nm. Six standard solu-tions of AVB1a (0.2, 0.4, 0.5, 0.6, 0.8 and 1.0mg/ml) were used toprepare calibration curve. The regression equation was yieldedto be Y = 5 × 10�5X� 0.011. The assay of AV tablet was calculatedto be 3.71%. It was testified that there were no significant differ-ences between the assay result obtained by HPLC method andqNMR method (the assay described by manufacturer was 3.78%).

Conclusion

A selective and accurate qNMR method has been established forvalidating the assay of AV tablets by using PMN as internal stan-dard and CDCl3 as the NMR solvent. The results of linearity, preci-sion, accuracy, LOD and LOQ emphasize that 1H NMR can beused as a validated method for quantitative determinations ofAVB1a. There were no significant differences between the assay re-sult obtained by HPLC and qNMR, which also provided evidencesfor qNMR acting as a more rapid and facile choice for the determi-nation of actives in drug formulations. The established qNMRmethod offers an excellent choice for routine quality control andstability analysis of AVB1a both in tablet form and in environment.In addition, the method also offers possibilities for the quantita-tive determination of drugs with low content.

Acknowledgements

We are grateful to the National Natural Science Foundation ofChina (number 21206148) for financial support.

References[1] J. L. Jungnickel, J. W. Forbes. Anal. Chem. 1963, 35, 938–942.[2] D. P. Hollis. Anal. Chem. 1963, 35, 1682–1684.[3] G. Marti, J. Boccard, F. Mehl, B. Debrus, L. Marcourt, P. Merle, E.

Delort, L. Baroux, H. Sommer, S. Rudaz. Food Chem. 2014, 150,235–245. doi:10.1016/j.foodchem.2013.10.103.

[4] S. Shen, J. Yao, Y. Q. Shi. J. Pharmaceut. Biomed. 2014, 89, 118–121.doi:10.1016/j.jpba.2013.10.036.

[5] U. Holzgrabe, I. Wawer, B. Diehl, NMR Spectroscopy in Drug Develop-ment and Analysis, WILEY-VCH Verlag GmbH, Weinheim, 1999.

[6] F. Malz, H. Jancke. J. Pharmaceut. Biomed. 2005, 38, 813–823.doi:10.1016/j.jpba.2005.01.043.

[7] X. R. Liang, L. P. Du, F. Su, H. S. Parekh, W. K. Su. Magn. Reson. Chem.2014, 52, 178–182. doi:10.1002/mrc.4048.

[8] H. Mrozik, P. Eskola, B. O. Linn, A. Lusi, T. L. Shih, M. Tischler, F. S.Waksmunski, M. J. Wyvratt, N. J. Hilton, T. E. Anderson, J. R. Babu,R. A. Dybas, F. A. Preiser, M. H. Fisher. Experientia 1989, 45,315–316. doi:10.1007/BF01951823.

[9] W. L. Shoop, H. Mrozik, M. H. Fisher. Vet. Parasitol. 1995, 59, 139–156.doi:10.1016/0304-4017(94)00743-V.

[10] L. J. Chen, B. H. Sun, J. P. Qu, S. W. Xu, S. Li. Chemosphere 2013, 93,2528–2534. doi:10.1016/j.chemosphere.2013.09.058.

n Wiley & Sons, Ltd. Magn. Reson. Chem. 2014, 52, 480–485

An effective method utilized on AVERMECTIN B1a assessment

[11] W. J. Zhu, M. Li, C. Liu, J. P. Qu, Y. H. Min, S. W. Xu, S. Li. Ecotox. Envi-ron. Safe. 2013, 98, 74–81. doi:10.1016/j.ecoenv.2013.09.021.

[12] H. Khaldoun-Oularbi, C. Richeval, N. Djenas, M. Lhermitte, L.Humbert, A. Baz. Ann. Toxicol. Anal. 2013, 25, 63–70. doi:10.1051/ata/2013039.

[13] C. C. Yang. Curr. Pharm. Biotechno. 2012, 13, 999–1003. doi:10.2174/138920112800399059.

[14] G. Molinari, S. Soloneski, M. L. Larramendy. Cytogenet. Genome Res.2010, 128, 37–45. doi:10.1159/000293923.

[15] N. Macdonald, A. Gledhill. Arch. Toxicol. 2007, 81, 553–563.doi:10.1007/s00204-007-0193-6.

[16] M. L. Reyzer, J. S. Brodbelt. Anal. Chim. Acta 2001, 436, 11–20.doi:10.1016/S0003-2670(01)00893-5.

Magn. Reson. Chem. 2014, 52, 480–485 Copyright © 2014 John

[17] J. Hernandez-Borges, L. M. Ravelo-Perez, E. M. Hernan-dez-Suarez, A.Carnero, M. A. Rodriguez-Delgado. J. Chromatogr. A 2007, 1165,52–57. doi:10.1016/j.chroma.2007.07.033.

[18] L. Giannetti, A. Giorgi, F. Necci, G. Ferretti, F. Buiarelli, B. Neri. Anal.Chim. Acta 2011, 700, 11–15. doi:10.1016/j.aca.2010.12.035.

[19] M. D. Hernando, J. M. Suarez-Barcena, M. J. M. Bueno, J. F. Garcia-Reyes, A. R. Fernandez-Alba. J. Chromatogr. A 2007, 1155, 62–73.doi:10.1016/j.chroma.2007.02.120.

[20] J. He, K. Lu, Y. X. Li. Chin. J. Anal. Lab. 2007, 26, 119–121.[21] H. Gadape, K. Parikh. Anal. Methods 2011, 3, 2341–2347. doi:10.1039/

C1AY05247K.[22] U. Holzgrabe. Prog. Nucl. Magn. Reson. Spectrosc. 2010, 57, 229–240.

doi:10.1016./j.pnmrs.2010.05.001.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/mrc

485