Journal of Pharmaceutical and Biomedical Analysis 

Simultaneous determination of six main components in Bushen Huoxue prescription by HPLC-CAD

Mengjun Xie, Yueting Yu, Ziyu Zhu, Liping Deng, Bo Ren∗, Mei Zhang∗
School of Pharmacy, Chengdu University of Traditional Chinese Medicine, No. 1166 Liutai Avenue, Wenjiang District, Chengdu, Sichuan, 611137, China

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

Article history:

Received 13 November 2020
Received in revised form 13 April 2021 Accepted 16 April 2021
Available online 23 April 2021

Keywords:

HPLC-CAD
Corona charged aerosol detector Bushen Huoxue prescription Content determination
Background: The Bushen Huoxue prescription is a traditional Chinese medicine formula treating diabetic retinopathy, which was developed by our research group. Catalpol, puerarin, salvianolic acid B, ginseno- side Rg1, ginsenoside Rb1 and ginsenoside Rd are six of main effective components, which could be partly representative of this prescription. The corona charged aerosol detector (corona CAD) is one kind of universal detectors equipped with the high performance liquid chromatography (HPLC). The CAD has many advantages for the analysis of complex mixtures, but too few applications in traditional Chinese medicine compounds.
Objective: The aims of this study are to establish a method for the determination of six components in Bushen Huoxue prescription, and to increase the use of the CAD in traditional Chinese medicine compounds.
Methods: HPLC-CAD analysis was performed on an Inertsil ODS-SP (4.6 mm × 250 mm, 5 µm) with a
mobile phase consisting of 0.5 % formic acid solution(A)-acetonitrile(B) at a flow rate of 1 mL/min (0−7 min, 1 % B; 7−12 min, 1 %–12 % B; 12−22 min, 12 %–19 % B; 22-40 min, 19 %–28 % B; 40−43 min,
28 %–33 % B; 43−50 min, 33 % B; 50−65 min, 33 %–42 % B). The column temperature maintained at 30
◦ C, the injection volume was 20 µL, the atomization temperature mode was LOW, the filtration constant
(filter) was 3.6 and data collection rate was 10 Hz. The methodology was examined and the linearity of regression of different functions was compared. Sixteen batches of samples were prepared and their contents were determined.
Results: The six compounds showed a better linearity (R2 > 0.9990) in their concentration ranges when
using the linear function. The average recoveries were 99.18 %–101.30 %. Although the RSD value of puer- arin and ginsenoside Rg1 was slightly out of 3 % during the average recovery investigation, all the other methodological investigations of the six components were within 3 %. The precision, stability and repeata- bility of the method were good. In sixteen batches of Bushen Huoxue prescription samples, the contents of six components were 0.3138 %–0.6042 % for catalpol, 0.8095 %–1.2917 % for puerarin, 0.7416 %–1.1189
% for salvianolic acid B, 0.0231 %–0.0418 % for ginsenoside Rg1, 0.0702 %–0.1724 % for ginsenoside Rb1, 0.0384 %–0.1196 % for ginsenoside Rd.
Conclusion: In this experiment, a method for the determination of six components in Bushen Huoxue prescription based on HPLC-CAD was established with high accuracy, good repeatability and simple operation, and it can provide references for the improvement of quality standard of the Bushen Huoxue prescription. It is reasonable and accessible for the CAD application in the determination of traditional Chinese medicine compound prescriptions.

© 2021 Elsevier B.V. All rights reserved.Abbreviations: HPLC, high performance liquid chromatography; CAD, charged aerosol detector; DAD, diode array detector; RID, refractive index detector; MSD, mass spectrometry detector; FLD, fluorescense detector; ELSD, evaporative light scattering detector; BHP, Bushen Huoxue prescription; TCM, traditional Chinese medicine; PLR, puerariae lobatae radix; SMRR, salviae miltiorrhizae radix et rhizoma; RR, rehmanniae radix; GRR, ginseng radix et rhizoma; LOD, limits of detection; LOQ, limits of quantification.

∗ Corresponding authors.
E-mail addresses: [email protected] (M. Xie), [email protected] (Y. Yu), [email protected] (Z. Zhu), [email protected] (L. Deng), [email protected] (B. Ren), [email protected] (M. Zhang).

0731-7085/© 2021 Elsevier B.V. All rights reserved.

1. Introduction: Traditional Chinese medicine (TCM) is a summary of the Chinese people’s experience in preventing and treating diseases for thou- sands of years. Modern researches showed that it could still make great contributions to the health care, and TCM has the advantages of multi-pathway treatment, less side effects and so on [1]. TCM prescription is an organic combination of many traditional Chinese medicinal materials according to the theory of TCM, which is the most important treatment method of TCM.

The quality control is an important link to ensure the safety and effectiveness of drugs. As a part of testing, the quantitative detection of components has attracted more attention. However, traditional Chinese medicinal materials almost comes from nat- ural products, so TCM prescription has high chemical diversity, and its quality is affected by many factors [2]. These complicated conditions of TCM prescriptions render challenging in detecting. Furthermore, the analytes in TCM formulae can be present in large or very small amounts, so very sensitive and selective methods may be needed for their detection [3].
High performance liquid chromatography (HPLC), as one of the most versatile techniques for separating, has the diversity of the matching detectors, which provides many ideas for the compound analysis of TCM. The common detectors in HPLC methods are the diode array detector (DAD), the refractive index detector (RID), the mass spectrometry detector (MSD), the fluorescence detector (FLD), the evaporative light-scattering detector (ELSD), the corona charged aerosol detector (corona CAD) and so on [3,4]. The RID is not conducive to the gradient elution and high selectivity of the FLD, so it is rarely used in the analysis of TCM prescription. The DAD, one of the branches in the ultraviolet-visible light (UV/vis) detection, is the most common detection for the HPLC, because of its wide field of the applications, the sensitivity, the wide range of linearity and the compatibility with gradient elution. But it comes to its limits when the analyte molecules are lacking a suitable chromophor, also the fluorescence detection for no-fluorophore. Non-chromophoric compounds must go to tedious derivatization procedures or use sensitivity-impairing low-UV-wavelength methods to analyze [5]. Then, people will generally turn to universal detectors, such as the CAD and the ELSD. The electrical aerosol technologies are the base of both two detectors [4,6], however, because the linearity, sensitivity and stability of the CAD are better than those of the ELSD, the selec- tion of the CAD has greater advantages in content determination [7,8].
The increasing inclusion of the CAD methods in the United States
Pharmacopoeia and the European pharmacopoeia, and the new introduction in the general chapter of the Chinese Pharmacopoeia, reflect the positive attitude to the CAD detector. Due to the advan- tages in the detection of complex mixtures, many natural products and their preparations have been tried the CAD as an analytical tool [3,9,10]. At present, the enthusiasm in the use of the CAD in TCM, compound medicine and its preparations is also increasing, such as, Trillium tschonoskii and Trillium kamtschaticum, seeds of Brucea Javanica (L.) Merr., ginseng, Sanhuang Tablet [11–14]. But there are few reports on the application of TCM compounds now.
Bushen Huoxue prescription (BHP) is a compound formula of TCM independently developed by our research group in the early stage to prevent and treat diabetic retinopathy (patent number: 200910308470.3), which is composed of Puerariae Lobatae Radix (PLR), Salviae Miltiorrhizae Radix et Rhizoma (SMRR), Rehman- niae Radix (RR), Ginseng Radix et Rhizoma (GRR). Through previous studies, it was found that puerarin, catalpol, salvianolic acid B and ginsenoside, etc., may be the effective components of this com- pound [15,16]. It is well known that ginsenosides is a kind of compounds lacking an ultraviolet chromophore [13,17]. Therefore, in this study, HPLC-CAD was used for simultaneous quantitative

detection of the content of active components to provide basic sup- port for other relevant studies on BHP, and meanwhile, to increase the application of the CAD in TCM.

2. Materials and methods

2.1. Plant materials

RR is the dried root tuber of Rehmannia glutinosa Libosch., SMRR is the dried root of Salvia miltiorrhiza Bge., GRR is the dried fine branch or fibrous root Panax ginseng C. A. Mey., and PLR is the dried root of Pueraria lobata(Wilid.) Ohwi. (Table 1). They were purchased from the Lotus Pond Chinese Herbal Medicine Wholesale Market (Chengdu, China) and authenticated by Prof. Jin Pei in School of Pharmacy, Chengdu University of TCM, Chengdu, China.

2.2. Apparatus

Analysis was performed on a HPLC system (Dionex Ulti- Mate3000, Thermo Fisher Scientific Inc.) equipped with a quaternary pump, an online degasser, an auto-sampler, a thermo- statically controlled column compartment and a charged aerosol detector (Dionex Corona Veo, Thermo Fisher Scientific Inc.).

2.3. Chemicals and reagents

Methanol (analytical and HPLC grade) and formic acid (HPLC grade) were purchased from the Chron Chemicals Co. LTD (Chengdu, China). Acetonitrile were HPLC grade from the Sigma- Aldrich (Shanghai, China). Deionized water was prepared by using a Milli-Q system (Bedford, MA, USA).
Catalpol (MUST-18103011), puerarin (MUST-12101113) and salvianolic acid B (MUST-16040702) were purchased from the Must Biotechnology Co. LTD (Chengdu, China). Ginsenoside Rg1 (PS0148-0025), ginsenoside Rb1 (S0149-0025) and ginsenoside Rd (PS0343-0025) were purchased from the Push Biotechnology Co. LTD (Chengdu, China). And the purity of six compounds is ≥ 99.98
%, ≥ 99.71 %, ≥ 99.41 %, ≥ 98.0 %, ≥ 98.0 %, ≥ 95.0 %, respectively.

2.4. Preparation of sample solutions and standard solutions

The BHP extract was prepared in a traditional method. According to the weight ratio (1:2:2:2), GRR, RR, SMRR and PLR were respec- tively collected, crushed, decocted with distilled water twice for an hour each time (eight times for the first time and six for the second), filtered and the filtrate was mixed, evaporated and dried until becoming dry cream extraction. Finally, the dry cream was weighted to calculate the extraction rate. The sixteen batches of samples were extracted for further using according to Table 2.
Based on the optimization results and previous experiments [18], the extraction conditions of the sample solution are as fol- lows. Weighing 0.1000 g dry cream powder, adding water to 5 mL, ultrasonic for 40 min, transferring the solution to the EP tube, centrifuging for 10 min, taking the supernatant through the
0.45 µm micro-porous filter membrane, and collecting the filtrate, the sample solution can be obtained.
Then a mixed standard solution containing 2.1376 mg/mL of catalpol, 5.5178 mg/mL of puerarin, 10.0598 mg/mL of salvianolic acid B, 1.8478 mg/mL of ginsenoside Rg1, 3.4344 mg/mL of ginseno- side Rb1 and 3.9615 mg/mL of ginsenoside Rd was prepared. The
stock solution was diluted with water in volumetric flasks to make a serial of concentrations, stored at 4 ◦C for using.

The information of herbsRR Ⓧ2Ⓧ4PLR, Puerariae Lobatae Radix. SMRR, Salviae Miltiorrhizae Radix et Rhizoma. RR, Rehmanniae Radix. GRR Ginseng Radix et Rhizoma.

Table 2
Sixteen batches of samples (g) and their rate of extraction.No Batch PLR SMRR RR GRR Dried extraction Rate of extraction

12 AIⓍ1 1 AIIⓍ2 2 3.993.99 4.004.00 3.994.01 2.001.99 7.327.53 52.36 %53.82 %34 AIIIⓍ3 3AIVⓍ4 1 3.994.00 3.994.00 3.984.00 1.992.00 7.727.99 55.34 %57.07 %5 BIIⓍ2 3 4.00 3.99 3.99 2.017.5453.90 %
6BIIIⓍ4.004.004.002.007.7655.43%7BIVⓍ323.994.003.992.07.6354.58%89BIⓍ13CIIⓍ324.013.984.004.003.993.992.012.006.887.3449.11%52.54%1011CIIIⓍ43CIVⓍ123.99.993.994.003.984.002.002.007.597.7254.37%55.18%1213CIⓍ21DIIⓍ214.013.994.014.003.994.002.002.017.347.6552.39%54.64%141516DIVⓍ4DIⓍ31DIIIⓍ134.004.003.994.003.993.993.973.994.002.012.00.997.617.897.7454.43%56.%55.40 %
Herbs in “Batch” are from Table 1. PLR, Puerariae Lobatae Radix. SMRR, Salviae Miltiorrhizae Radix et Rhizoma. RR, Rehmanniae Radix. GRR Ginseng Radix et Rhizoma.
2.5. Performance of the CAD

HPLC-CAD analysis was performed on an Inertsil ODS-SP (4.6 mm 250 mm, 5 µm, C/N. 5020-02746 S/N. 4DJ92607) with
a mobile phase consisting of 0.5 % formic acid solution(A)- acetonitrile(B) at a flow rate of 1 mL/min (0−7 min, 1 % B; 7−12 min, 1 %–12 % B; 12−22 min, 12 %–19 % B; 22−40 min, 19 %–28 % B;
40−43 min, 28 %–33 % B; 43−50 min, 33 % B; 50−65 min, 33 %–42 %
B). The column temperature maintained at 30 ◦C, the injection vol-
ume was 20 µL, the atomization temperature mode was Low, the filtration constant (filter) was 3.6 and acquisition frequency was 10 Hz.

3. Results

3.1. Optimization of analytic conditions

For better establishing of content determination method, the chromatographic conditions and extraction conditions were opti- mized. Salvianolic acid B in this prescription has a relatively high content, with great efficacy, which is worth determining. However, the poor peak shape caused by a large number of acidic groups of salvianolic acid B is a challenge for this study.
For optimization of chromatographic conditions, we investi- gated the mobile phase, elution gradient, column temperature, flow rate, injection volume, atomization temperature mode, filtration constant, acquisition frequency and chromatographic column. The extraction solvent and the liquid-solid ratio also have been opti-

mized for better extraction (shown in “Appendices”). Based on the above, several optimization experiments are selected for specific explanation.
According to previous studies and references, we investigated the mobile phase types firstly: formic acid solution-methanol, water-acetonitrile, formic acid solution-acetonitrile (, then investigated the concentration of formic acid in mobile phase: 0.5
%, 0.25 % and 0.1 % formic acid solution- acetonitrile . Long et al. [19] found formic acid has no bad effect on the response of the CAD, and low background currents are obtained by using formic acid as mobile phase additives. So it is the buffer solution to regulate the pH. As , there are less peak in the chromatographic pictures of formic acid solution-methanol and water-acetonitrile, and some peak could not be separated. So we chose the formic acid solution-acetonitrile as the mobile phase. More symmetrical shape of the peak and higher column efficiency decided 0.5 % formic acid- acetonitrile as the final, and it avoided the damage to the instrument that we did not increase the concentration.
We weighed 0.1000 g of dried power, added the solution to solute, had an ultrasound and centrifugation of the sample, and filtration were performed to prepare the test products. Keeping all other conditions the same, we investigated the chromatographic behaviors of different kinds of extraction solvents: methanol, 50
% methanol and water. As shown in Table 3 and, after the addition of methanol, many chromatographic peaks could not be separated, resulting in the increase of peak area on the surface. In addition, considering the solvent interference and the traditional preparation method, pure water was chosen as the solvent.

The extraction solvent.
Extraction solvent a b
c d
e f
100 % methanol 2.9199 9.5935 10.3308 0.2527 3.5319 1.3675
50 % methanol 3.3207 13.1463 13.118 0.3626 3.5319 2.2503
Water 2.3784 17.7337 11.4457 0.5273 4.4094 3.2302
100 % methanol, the pure methanol (absolute methanol). 50 % methanol, methanol-water (50:50, v/v).
a Catalpol.
b Puerarin.
c Salvianolic acid B. d Ginsenoside Rg1. e Ginsenoside Rb1. f Ginsenoside Rd.

. 1. Investigation of mobile phase types.

3.2. Specificity and system applicability

20 µL of blank solution (the solvent of samples), standard solu- tion and test solution were taken respectively, and the samples were injected and analyzed according to the above chromato- graphic conditions of “2.5”. The resolutions of six compounds and their own adjacent chromatographic peaks are more than 1.5, the minimum number of theoretical plates in each chromato- graphic peak is more than 15000. Though compared with the chromatograms of test, standard and blank solution, no interfer- ences were detected in the corresponding retention times of the target compounds.

3.3. Linearity

Standard solutions of six compounds diluted in a ratio of 1 : 2 : 5 : 10 : 20 for five concentrations were analyzed . The regression equations were calculated in the form of Y = a X +b, Y= a X2+bX+c, Log(Y)=a Log(X)+b, (X, concentration of standard solution, mg/mL. Y, the corresponding peak area of compounds injected) while a, b and c were the constant terms depending on the analytes

Investigation of concentrations of formic acid in mobile phase. The insert  was magnification drawing of the CAD chromatogram from 29 to 33 min.

and chromatographic conditions. The R2 values showed that linear function and quadratic function are more suitable for our analytes (R2 > 0.9990), compared with log-log function (Table 4). Although the relationship between the signal and the amount of mass present was of non-linear nature in CAD, in theory, we still accepted the lin- ear function instead of the others for the convenience the relatively narrow gradient for six compounds in this study.

3.4. Limit of detection and quantification

The limits of detection (LOD) and limits of quantification (LOQ) under the chromatographic condition in “2.5” were determined at a signal-to-noise (S/N) ratio of 3 and 10 respectively. Results were shown in Table 5.

3.5. System precision

Method precision was evaluated by the analysis of six injections of the sample No. 4 according to the conditions in “2.5”. Record each

The calibration curve (R2 ).

Analyte Linear function Quadratic function Log-log function
a Y = 17.649X +0.1951 (0.9995) Y = -2.0699 X2 + 18.613 X + 0.1365 (0.9995) Log(Y) = 0.8693 Log(X) + 1.1892 (0.9993)
b
Y = 23.811 X +0.6453 (0.9991) Y = -1.5577 X2 + 25.164 X + 0.4921(0.9994) Log(Y) = 0.8342 Log(X)+ 1.3494 (0.9989)
c Y = 31.070 X +1.1178(0.9993) Y = -1.9833 X2 + 33.493 X + 0.7324 (0.9998) Log(Y) = 0.8418 Log(X)+ 1.4889 (0.9988)
d
Y = 51.463 X +0.0358 (0.9995) Y = 20.873 X2 + 48.381 X + 0.0950 (0.9998) Log(Y) = 0.9151 Log(X)+ 1.6164 (0.9989)
e Y = 68.032 X +0.0990 (0.9997) Y = 1.6684 X2 + 67.687 X + 0.1083 (0.9997) Log(Y) = 0.9788 Log(X)+ 1.8226 (0.9996)
f
Y = 63.483 X +0.1406 (0.9993) Y = -11.23 X2 + 66.349 X + 0.0453 (0.9995) Log(Y) = 1.0573 Log(X)+ 1.8738 (0.9969)
X, concentration of standard solution, mg/mL. Y, the corresponding peak area.
a Catalpol.
b Puerarin.
c Salvianolic acid B. d Ginsenoside Rg1. e Ginsenoside Rb1. f Ginsenoside Rd.

Table 5
The regression data, LODs, and LOQs for the six compounds.

0.9991 206∼5517.8 0.1089 0.3224

c 0.9993 289.5∼10059.8 0.7070 2.3333
d 0.9995 35∼1847.8 0.7800 2.4000
e 0.9997 49∼3434.4 0.6778 2.1333
f 0.9993 60.5∼3961.5 0.6889 2.2667
a Catalpol.
b Puerarin.
c Salvianolic acid B. d Ginsenoside Rg1. e Ginsenoside Rb1. f Ginsenoside Rd.

Table 6
The repeatability, recoveries and stability data for the six compounds.

Analyte Precision (n = 6) Recovery (n = 9) Stability Repeatability (n = 6)
RSD (%) Mean (%) RSD (%) RSD (%) RSD (%)
a 2.49 101.30 2.13 2.00 2.94
b
1.82 101.13 3.11 1.40 1.68
c 2.36 101.11 1.67 1.09 2.76
d
2.87 99.59 3.33 2.48 2.86
e 0.95 99.18 1.32 0.96 2.73
f
2.31 100.88 2.00 1.79 1.70
a Catalpol.
b Puerarin.
c Salvianolic acid B.
d Ginsenoside Rg1.
e Ginsenoside Rb1.
f Ginsenoside Rd.

peak area and calculate the RSD (%) value of each component peak area. Results showed there is good precision

3.6. Repeatability

For the repeatability testing, approximately 0.1000 g of the fine powder of sample No. 4 was accurately weighted six times, and six test solutions were prepared. Analysis of six injections according to the conditions in “2.5” was recorded. Each peak area and the RSD (%) value of peak area showed there is qualified repeatability

3.7. Stability

The stability of the test solution from sample No. 4 during anal- yses when stored at indoor temperature was investigated. Record differences in peak areas among five injections of prepared samples after 0, 2.5, 5, 8, 24 h, respectively. And the RSD (%) of six peak areas showed great stability in 24 h

3.8. Accuracy, spike recovery

The accuracy was expressed in terms of compound recovery at three concentration levels (standard: sample = 1.5 : 1, 1 : 1, 0.5 : 1). Weighting 0.0500 g of the fine powder of sample No. 4 accurately, and accurate amounts of mixed standards were added to the pow- der, then the compound was extracted and tested as described in “2.4” and “2.5”. The recoveries of six analytes were shown in  10 and Table 6. The RSD (%) values of a few components are slightly more than 3 %.

3.9. Sample analysis

The samples of 16 batches of compound medicinal materials were prepared according to the method of “2.4”. According to the chromatographic conditions of “2.5”, the samples were injected and the content of six components in medicinal herbs was calculated (Table 7). The contents of 6 components were 0.3138 %–0.6042 % for catalpol, 0.8095 %–1.2917 % for puerarin, 0.7416 %–1.1189 % for

Table 7
Six components contents of 16 batches of BHP.
Batch No a(%) b(%) c(%) d(%) e(%) f(%) AIⓍ1 1 1 0.4367 0.9497 0.9182 0.0234 0.1376 0.0939
AIIⓍ2 2 2 0.5375 1.0151 1.0108 0.0289 0.0899 0.0444
AIIIⓍ3 3 3 0.3457 1.0450 1.0426 0.0315 0.1483 0.0877
AIVⓍ4 1 4 0.4228 0.8095 0.8071 0.0397 0.1702 0.1196
BIIⓍ2 3 5 0.6042 1.1660 0.9839 0.0263 0.1328 0.0883
BIIIⓍ4 1 6 0.5056 1.1836 1.0929 0.0268 0.1272 0.0896
BIVⓍ3 2 7 0.3387 1.2206 0.8060 0.0331 0.0797 0.0474
BIⓍ1 3 8 0.4634 1.1055 0.8505 0.0231 0.1285 0.0749
CIIⓍ3 2 9 0.4495 1.0825 0.9521 0.0338 0.0702 0.0384
CIIIⓍ4 3 10 0.5303 1.2917 0.9081 0.0341 0.1400 0.0984
CIVⓍ1 2 11 0.4619 1.2070 1.0019 0.0347 0.1211 0.0793
CIⓍ2 1 12 0.4465 1.2005 1.1189 0.0259 0.1399 0.0847
DIIⓍ2 1 13 0.3138 1.0154 0.9223 0.0327 0.1257 0.0844
DIVⓍ4 2 14 0.4107 1.1935 0.7416 0.0418 0.0922 0.0528
DIⓍ3 1 15 0.4944 1.2082 0.9445 0.0381 0.1724 0.1035
DIIIⓍ1 3 16 0.3535 1.2328 1.0944 0.0375 0.1336 0.0891
a Catalpol.
b Puerarin.
c Salvianolic acid B. d Ginsenoside Rg1. e Ginsenoside Rb1. f Ginsenoside Rd.

catalpol. b, puerarin. c, salvianolic acid B. d, ginsenoside Rg1. e, ginsenoside Rb1. f, ginsenoside Rd.

. 3. Investigation of extraction solvents.
salvianolic acid B, 0.0231 %–0.0418 % for ginsenoside Rg1, 0.0702
%–0.1724 % for ginsenoside Rb1, 0.0384 %–0.1196 % for ginsenoside Rd. There are large differences between batches, meaning that the source could affect the quality of the BHP.

4. Discussion

In this experiment, on the basis of previous researches and references, through single-factor investigations, chromatographic conditions and extraction conditions were determined. This study

found that the acidity of the mobile phase was a significant factor to the results. The increase of acidity would generally improve the detectable rate and peak shape of components. Vervoort et al. [20] found that formic acid has no significant effect on the response of the CAD. However, as for the salvianolic acid B, because of the mul- tiple acidic groups in its structure , the acidity does change its existence form in the detector, thus to affect the results. This experiment proves that an appropriate acidity is a crucial factor in the determination of chemical compositions, like salvianolic acid B.
When using a CAD detector, the pump should be turned off and let the detector fill with N2 to balance the interior of the CAD before injection and before shutdown to achieve even solvent vaporization and charge transfer effect. In addition, in order to obtain more stable and reliable data, a blank injection can be added before the formal injection.
The establishment of a method is always based on the applicabil- ity of the method, and this applicability should be within a certain range. The CAD is not applicable to the scope of the object being volatile substances, but reasonable for prescriptions of the water decoction. More importantly, the response value is nonlinear over a wide range of concentrations according to the working principle

. 5. Chemical structural formulae of six components. a, catalpol. b, puerarin. c, salvianolic acid B. d, ginsenoside Rg1. e, ginsenoside Rb1. f, ginsenoside Rd.

 6. The chromatographic  of the standard solutions.

7. The chromatographic of the precision(n = 6).

. 8. The chromatographic  of the repeatability(n = 6).

 9. The chromatographic  of the stability (0 h, 2.5 h, 5 h, 8 h, 24 h).

10. The chromatographic  of the recovery (n = 9).
of the CAD. However, when using the CAD to detect the concen- tration of compounds, the analysis range of most compounds is relatively narrow. At the same time, the analyst will also carry out certain screening of function types when linear-regression analysis. According to literature reviews, most of the analysis objects are in a relatively narrow concentration range, so many of them adopt the linear function regression [11,14,19,21]. How- ever, some reports use the quadratic or power function [22–25], including some researches that compare multiple regression func- tions [5,26,27]. Aiming at this phenomenon, this experiment also made a comparison, and found that R2 values were > 0.999 both in the regression of the linear function and the quadratic function, but it could not be achieved after the log-log regression (power function). In addition, we found that the quadratic term coefficient of the quadratic function of ginsenoside Rb1 and ginsenoside Rg1 were > 0, and the remaining four compounds < 0 (Table 4).
There are some relationships between components and response. Firstly, we could know that the concentration is positively correlated with the droplet diameter according to Eq. 1. When dp
> 10 nm, the sensitivity is positively correlated with the droplet diameter (Eq. 2), so a< 0; when dp < 10 nm, the opposite is true (Eq. 3), a > 0. So we speculated that why ginsenoside Rb1 and gin- senoside Rg1 got positive constant in their linearity regression is the lower concentration of ginsenoside Rb1 and ginsenoside Rg1 in this prescription than other test analytes, as Satoshi Furota.et reported in the paper [23].
dp = d0×(c/pp)1⁄3 (1)
S = [(3.01×1011)/pp]×dp−1.89 (2)

S = [(4.4×105)/pp]×dp3.6 (3)
dp(nm): diameter of the dried particle; d0(nm): the initial micro- droplet diameter; c: concentration of analytes; S(fAm3 g−1): sensitivity of detection; pp: density of analytes.
Because the contents of puerarin and salvianolic acid B in this prescription is higher than other components, and the existence of weak ultraviolet absorption substances (ginsenosides), these make it difficult for many detectors to be used in BHP. Compared with the application of UV, ELSD and MS [28,29], the CAD detector has certain advantages. The CAD is universal to most analytes when compared with the UV, and its detector response doesn’t depend on certain functional groups or moieties within the molecules. The peak number detected by the CAD is larger than that detected by the ELSD [6], so a lot of chromatographic information will not be lost. And its cost won’t be as expensive as the MS detector, which will help its popularity at this stage. In addition, the CAD has a stable and consistent response and can detect compounds at high or low concentrations at the same time [4]. All these characteristics make it suitable for the complex mixture analysis of TCM compound.

5. Conclusion

In this experiment, a method for the determination of six com- ponents in BHP based on HPLC-CAD was established with high accuracy, good repeatability and simple operation, and it can pro- vide references for the improvement of quality standard of the BHP. It is reasonable and accessible for the CAD application in the determination of TCM compound prescription, or other complex chemical systems (especially those with non-chromophore).

Ethical approval

This article does not contain any studies with human partici- pants or animals performed by any of the authors.

CRediT authorship contribution statement

Mengjun Xie: Investigation, Methodology, Writing – original draft. Yueting Yu: Visualization. Ziyu Zhu: Methodology. Liping Deng: Investigation. Bo Ren: Supervision. Mei Zhang: Formal anal- ysis, Writing – review & editing, Conceptualization.

Declaration of Competing Interest The authors report no declarations of interest.

Acknowledgements

This work was supported by the National Natural Scientific Foundation of China [grant number: 81774202], the Sichuan Science and Technology Program [grant number: 2017JY0013], the Fund of Scientific Research Innovation Team Construction in Sichuan Provincial University (grant number: 18TD0017), the Xinglin Scholar Research Premotion Project of Chengdu Univer- sity of TCM [grant number: CXTD2018010] and the Open Research Fund of Chengdu University of Traditional Chinese Medicine Key Laboratory of Systematic Research of Distinctive Chinese Medicine Resources in Southwest China [grant number: 2020XSGG017].

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version

References

[1] J.L. Ren, A.H. Zhang, X.J. Wang, Traditional Chinese medicine for COVID-19 treatment, Pharmacol. Res. 155 (2020), 104743.
[2] H.M. Gao, Z.M. Wang, Y.J. Li, Z.Z. Qian, Overview of the quality standard research of traditional Chinese medicine, Front. Med. (Lausanne) 5 (2) (2011) 195–202.
[3] J.L. Wolfender, HPLC in natural product analysis: the detection issue, Planta Med. 75 (7) (2009) 719–934.
[4] Charged aerosol detection for liquid chromatography and related separation techniquesH.A. Azeem, Paul H. Gamache (Eds.), Anal. Bioanal. Chem. 410 (11) (2018) 2663–2664.
[5] K. Filip, G. Grynkiewicz, M. Gruza, K. Jatczak, B. Zagrodzki, Comparison of ultraviolet detection and charged aerosol detection methods for
liquid-chromatographic determination of protoescigenin, Acta Pol. Pharm. 71 (6) (2014) 933–940.
[6] I.A. Haidar Ahmad, A. Blasko, J. Tam, N. Variankaval, H.M. Halsey, R. Hartman,
E.L. Regalado, Revealing the inner workings of the power function algorithm in charged aerosol detection: a simple and effective approach to optimizing power function value for quantitative analysis, J. Chromatogr. A 1603 (2019) 1–7.
[7] H.Y. Eom, S.Y. Park, M.K. Kim, J.H. Suh, H. Yeom, J.W. Min, U. Kim, J. Lee, J.R. Youm, S.B. Han, Comparison between evaporative light scattering detection and charged aerosol detection for the analysis of saikosaponins, J. Chromatogr. A 1217 (26) (2010) 4347–4354.
[8] S. Almeling, D. Ilko, U. Holzgrabe, Charged aerosol detection in pharmaceutical analysis, J. Pharm. Biomed. Anal. 69 (2012) 50–63.
[9] R. Yang, Z.N. Wu, Y.Q. Pu, T. Zhang, B. Wang, Fast and non-derivative method based on high-performance liquid chromatography-charged aerosol detection for the determination of fatty acids from Agastache rugosa (Fisch. et Mey.) O. Ktze. seeds, Nat. Prod. Res. 33 (13) (2019) 1969–1974.

[10] S. Martin-Torres, A.M. Jimenez-Carvelo, A. Gonzalez-Casado, L.
Cuadros-Rodriguez, Differentiation of avocados according to their botanical variety using liquid chromatographic fingerprinting and multivariate classification tree, J. Sci. Food Agric. 99 (11) (2019) 4932–4941.
[11] Y.J. Yang, X.G. Sun, J. Yang, Q. Li, J. Zhang, Y. Zhao, B.P. Ma, B.L. Guo, Determination of three saponins in rhizoma and fibrous root of Trillium tschonoskii and Trillium kamtschaticum, Zhongguo Zhong Yao Za Zhi 42 (6) (2017) 1146–1151.
[12] Z.N. Wu, L. Li, N. Li, T. Zhang, Y.Q. Pu, X.T. Zhang, Y. Zhang, B. Wang, Optimization of ultrasonic-assisted extraction of fatty acids in seeds of Brucea Javanica (L.) Merr. from different sources and simultaneous analysis using high-performance liquid chromatography with charged aerosol detection, Molecules 22 (6) (2017) 931.
[13] S.D. Jia, J. Li, N. Yunusova, J.H. Park, S.W. Kwon, J. Lee, A new application of charged aerosol detection in liquid chromatography for the simultaneous determination of polar and less polar ginsenosides in ginseng products, Phytochem. Anal. 24 (4) (2013) 374–380.
[14] H.Y. Fung, Y. Lang, H.M. Ho, T.L. Wong, D.L. Ma, C.H. Leung, Q.B. Han, Comprehensive quantitative analysis of 32 chemical ingredients of a Chinese patented drug Sanhuang tablet, Molecules 22 (1) (2017) 111.
[15] Y. Li, Preliminary Study on the Chemical Composition in Vivo and Vitro of Bushen Huoxue Prescription and the Mechanism of Treating DR, Chengdu University of Traditional Chinese Medicine, 2018 (in Chinses).
[16] S.Y. Liu, Study on the Basis of the Efficacy of Bushen Huoxue Prescription in the Prevention and Treatment of DR Based on the Multi-dimensional Spectrum-Effect Relationship, Chengdu University of Traditional Chinese Medicine, 2019 (in Chinses).
[17] L.F. Ouyang, Z.L. Wang, J.G. Dai, L. Chen, Y.N. Zhao, Determination of total ginsenosides in ginseng extracts using charged aerosol detection with
post-column compensation of the gradient, Chin. J. Nat. Med. 12 (11) (2014) 857–868.
[18] X.J. Xie, M. Zhang, M.F. Wang, The Invention Relates to a Chinese Traditional Medicine Composition and a Preparation Method Thereof. China, 2021.
[19] Z. Long, Z.M. Guo, I.N. Acworth, X.D. Liu, Y. Jin, X.G. Liu, L. Liu, L. Liang, A
non-derivative method for the quantitative analysis of isosteroidal alkaloids from Fritillaria by high performance liquid chromatography combined with charged aerosol detection, Talanta 151 (2016) 239–244.
[20] N. Vervoort, D. Daemen, G. Torok, Performance evaluation of evaporative light scattering detection and charged aerosol detection in reversed phase liquid chromatography, J. Chromatogr. A 1189 (2008) 92–100.
[21] C. Asthana, G.M. Peterson, M. Shastri, R.P. Patel, Development and validation of a novel high performance liquid chromatography-coupled with corona charged aerosol detector method for quantification of glucosamine in dietary supplements, PLoS One 14 (5) (2019), e0216039.
[22] S.D. Jia, W.J. Lee, W.E. Ji, J.H. Park, S.W. Kwon, J. Lee, Comparison of ultraviolet detection, evaporative light scattering detection and charged aerosol detection methods for liquid-chromatographic determination of anti-diabetic drugs, J. Pharm. Biomed. Anal. 51 (4) (2010) 973–978.
[23] S. Furota, N.O. Ogawa, Y. Takano, T. Yoshimura, N. Ohkouchi, Quantitative analysis of underivatized amino acids in the sub- to several-nanomolar range by ion-pair HPLC using a corona-charged aerosol detector (HPLC-CAD), J Chromatogr B 1095 (2018) 191–197.
[24] C.E. Zhang, L.J. Liang, X.H. Yu, H. Wu, P.F. Tu, Z.J. Ma, K.J. Zhao, Quality assessment of astragali radix from different production areas by simultaneous determination of thirteen major compounds using tandem UV/charged aerosol detector, J. Pharm. Biomed. Anal. 165 (2019) 233–241.
[25] J. Fibigr, D. Satinsky, P. Solich, A UHPLC method for the rapid separation and quantification of phytosterols using tandem UV/Charged aerosol detection – a comparison of both detection techniques, J. Pharm. Biomed. Anal. 140 (2017) 274–280.
[26] S. Granica, J.P. Piwowarski, A.K. Kiss, Determination of C-glucosidic ellagitannins in Lythri herba by ultra-high performance liquid chromatography coupled with charged aerosol detector: method development and validation, Phytochem. Anal. 25 (3) (2014) 201–206.
[27] Y. Wang, Y.X. Liu, H.S. Yue, W.Y. Xu, J.M. Cao, H.Y. Jin, S.C. Ma, Comparison between the charged aerosol detector and evaporative light scattering detector for the analysis of sugar in Zhusheyong Yiqifumai and the study on the accuracy of methods, Zhongguo Zhong Yao Za Zhi 45 (2020) 5511–5517.
[28] Z.Y. Wang, C.Q. Li, M.J. Mu, J.Y. Dong, M.L. Liu, M. Zhang, X.J. Xie, Simultaneous determination of the content of danshensu, puerarin, daidzin and salvianolic acid B in Bushen Huoxue prescription Salvianolic acid B by HPLC wavelength switching technology, Chin. J. Pharm. Anal. 36 (6) (2016) 1020–1026 (in Chinese).
[29] S.Y. Liu, X. Wang, Y.L. Li, Y.X. Pan, Y. Li, X.J. Xie, M. Zhang, HPLC-UV-ELSD fingerprint of different Bushen Huoxue prescriptions with different ratios, Liaoning J. Tradit. Chin. Med. 46 (02) (2019), 362-365+447 (in Chinese).