Simultaneous colorimetric determination of morphine and ibuprofen based on the aggregation of gold nanoparticles using partial least square

Abstract

In this work a new method is presented for simultaneous colorimetric determination of morphine(MOR) and ibuprofen(IBU) based on the aggregation of citrate-capped gold nanoparticles (AuNPs). Citrate-capped gold nanoparticles were aggregated in the presence of morphine and ibuprofen. The difference in kinetics of AuNPs aggregation in the presence of morphine / ibuprofen was used for simultaneous analysis of morphine and ibuprofen. The formation and size of synthesized Au NPs and the aggregated forms were monitored by infra-Red (IR) spectroscopy and transmission electron microscopy (TEM) respectively.. By adding morphine or ibuprofen the absorbance was decreased at 520 nm and increased at 620 nm. The difference in kinetic profiles of aggregation was applied for simultaneous analysis of MOR and IBU using partial least square regression as an efficient multivariate calibration method. The number of PLS latent variables was optimized by leave-one-out cross-validation method using predicted residual error sum of square. The proposed model exhibited a high capability in simultaneous prediction of MOR and IBU concentrations in real samples. Our results showed linear ranges of 1.33–33.29 µg/mL (R²=0.9904) and 0.28–6.9 µg/mL (R²=0.9902) for MOR and IBU respectively with low detection limits of 0.15 and 0.03 µg/mL(S/N=5).

1. Introduction

Analytical methods based on the spectrophotometric measurements, such as UV–Vis, have attracted increasing interest due to their availability, simplicity and ease of operation [1]. Multivariate spectral calibration is becoming a general method which can leads to quantitative spectral analysis, allowing simultaneous determination of several analytes that coexist in samples with minimum number of preparation steps [2].

Among multivariate calibration methods, partial least squares (PLS) regression as a powerful method has been applied because it can model the effect of the interfering components without primary separation [2].

Gold nanoparticles (AuNPs) are particularly attractive because they have surface modifying capability such as bio-labelling and also very useful for biological applications. Some unique characteristics of AuNPs are particularly large surface area, good bio-compatibility, high conductivity and electro catalytic activity [3], [4], [5]. AuNPs are also useful colorimetric probes because of their distance-dependent optical properties and extremely high extinction coefficients in visible region that make it more sensitive than AgNPs [6], [7]. AuNPs simplicities rely on their unique surface plasmon resonance from red to blue, corresponding to their dispersion and aggregation states [8], respectively. Based on this principle, several colorimetric methods have been developed for the detection of biomolecules [9], 2,4,6-trinitrotoluene [10], drugs [9] and melamine [11], [12].

Morphine (MOR), a non-steroidal anti-inflammatory drug (Scheme 1A), is used primarily to treat both acute and chronic severe pain. It is also used for pain due to myocardial infarction and for labor pains [13] and its duration of analgesia is about three to seven hours [13].

Scheme 1.

Structure of (A) morphine and (B) ibuprofen.

Ibuprofen (IBU), is commonly used as an antipyretic drug especially for podiatry (Scheme 1B) [14]. Many procedures have been applied for the quantitative determination of IBU. These procedures include non-aqueous titration [15], polarography [16], colorimetric titration [17], first and second derivative spectrophotometry [18], Infra-red (IR) spectrophotometry [19] and H-Nuclear magnetic resonance (H_NMR) spectroscopy [20]. In addition, several chromatographic procedures are also employed using thin-layer chromatography [21] and gas chromatography [22]. The chromatographic methods are most common direct methods [23] but they are classified as time consuming methods. Also HPLC methods need organic and hazardous solvents.

Various analytical methods have been developed for the determination of morphine. The same as IBU determination, chromatographic method [24], [25] are most commonly methods for MOR detection. Fluorescence [26], enzyme-linked immunosorbent assay [27], immunoassays, such as surface plasmon resonance (SPR) based immunosensors [28] and radioimmunoassays (RIA) [29], molecular imprinting technique [30] amperometric methods [31], chemiluminescence [32] and electrochemical methods [33] are also reported for detection of morphine [34], [35], [36], [37].

In this work, a new method based on aggregation of citrate-capped gold nanoparticles (AuNPs) is presented for simultaneous determination of MOR and IBU. The method is based on the difference in the aggregation rate of citrate-capped gold nanoparticles in the presence of MOR and IBU. In the presence of MOR or IBU the absorbance of the solution containing AuNPs, is decreased in 520 nm and increased in 620 nm which shows the aggregation of AuNPs. Difference in the aggregation kinetics (e.g. absorbance increase versus time at 620 nm) was applied for simultaneous analysis of MOR and IBU using partial least squares (PLS) regression as an efficient multivariate calibration method. This method is selective without any initial sample preparation. The proposed method is a simple, fast, and low cost procedure, in contrast to chromatographic methods. This method does not need any expensive apparatus. Also because of applying the chemometrics methods for analyzing, the proposed method needs no primary time-consuming separation process.

2. Experimental

2.1. Chemicals and materials

All used materials and reagents were of analytical grade, solvents were of spectroscopic grade and double distilled water (DDW) was used throughout the experiments. MOR and IBU pure drugs were obtained from the Department of Food and Drug Administration, Urmia, Iran. All chemicals used in the experiments were of analytical grade and were used without further purification. Tri sodium citrate dehydrate, HCl and NaOH were purchased from Merck (Darmstadt, Germany).

2.2. Apparatus

Absorption spectra were recorded with an Agilent 8453 UV–Visible spectrophotometer with a 1 cm quartz cells. The size, morphology and structure of the synthesized AuNPs were characterized by transmission electron microscopy (TEM, Philips-CMC-300 kV). A Metrohm pH meter model 713 pH-meter was used for pH measurements. A 40 kHz universal ultrasonic cleaner water bath (Elmasonic E60H, German) was used. The calculations were performed in MATLAB (Hyper-cube Inc. Version10) software using PLS mfiles.

2.3. Synthesis of gold nanoparticles

The Au seeds were synthesized according to Ferns method. Briefly, 100 mL of 1 mM aqueous solution of HAuCl₄ was heated to boil with stirring; then 10 mL 1% (w/v) aqueous sodium citrate was added. The color of the mixed solution changed from yellow to wine red in a few minutes, indicating the formation of AuNPs. The boiling and stirring were continued for 15 min [3]. The seed solution was cooled to room temperature and was stored in a dark bottle at 4 C. The solution of the prepared citrate-capped AuNPs was wine red, and had a characteristic localized surface plasmon resonance (LSPR) absorption band of AuNPs at 520 nm (λ_max) with narrow peak. In the solution, monodisperse AuNPs were red and exhibited a relatively narrow surface plasmon absorption band centered at around 520 nm in the UV–Vis spectrum. In contrast, a solution containing aggregated AuNPs was blue corresponding to a characteristic red shift in the surface plasmon resonance to higher wavelength of 620 nm [3]. The results were presented in Fig. 1 (A–B). Also, Fig. 1 (C, D) show the IR spectra of the synthesized Au Nanoparticles, before and after aggregation. The concentration of the synthesized nano particles were estimated as 10⁻⁸ mol L⁻¹ based on the extinction coefficient for 13 nm gold nanoparticles reported in the litratures as 2.8×10⁸. In IR spectra it can be demonstrated the functional groups of PhOH (3000–3600 cm⁻¹), RCOOH (2800–3200 cm⁻¹), OR₂ (1000–1300 cm⁻¹), Allyl-OH (3000–3600 cm⁻¹). In the spectrum of aggregated AuNPs using the studied drugs the NR₃ bands (1000–1350 1/ cm) obviously exist that prove drugs had interaction whit AuNPs. Intensity of IR spectrum of AuNps when drugs coated them or have interaction with them, are decreased and also the functional group of drug are increase in intensity. The absorbance spectra of synthesized AuNPs before and after aggregation were shon in Fig. 1E. Also Fig. 1F represents the changes of absorbance during time for morphine.

Fig. 1.

(A)TEM of synthesized AuNPs, (B) TEM of aggregated AuNPs, (C) IR of aggregated AuNPs, (D) IR of AuNPs, (E) spectrum of synthesized AuNPs and aggregated AuNPs spectrum in optimum condition and (F) aggregation spectra of AuNPs up to 10 min in the presence of morphine 10 µg/mL. Conditions: temperature 25 °C, ionic strength 1 mmol/L and pH 6.

3. Results and discussion

3.1. Optimization of NaCl concentration

Ionic strength has a crucial role in the aggregation of nano particles. It can be attributed to the ability of strong electrolytes to constrict the aroused electrical double-layer from the capping agent around the nano particles. Although the addition of an alectrolyte is necessary for starting the aggregation, it was also found that by increasing the ion strength above a certain limit, the aggregation of nanoparticles occurred even in the absence of analytes [1]. Therefore the effect of ionic strength (electrolyte concentration) should be studied. The results these study were presented in Figs. S1 and 1 mmol L⁻¹ of NaCl was used as optimized value.

3.2. Optimization of pH

Because of the presence of hydroxyl, carboxyl and amine groups in drugs, pH is another critical parameter that should be taken into consideration. Also electrostatic interactions are dominantly responsible for aggregation of AuNPs in the presence of drugs [6]. The results of interaction of studied drugs with AuNPs were presented in Fig. S2. in order to increase the possibility of electrostatic interactions, the best condition is achievable in which the drug molecules are available in the nanoparticle surroundings [38]. Our studied showed that the synthesized AuNps are stable at pH 6 and also the drugs have the most favorable structure for interaction with AuNPs in the pH range of 6–7 (see Fig. S2). Therefore pH 6 was chosen for further studies.

3.3. Optimization of incubation time

Different incubation times were examined to find the optimum value. The results indicated that AuNPs started to aggregate right after mixing with MOR and IBU in optimum conditions and was suppressed and spectral change was detectable in 10 min which considered to be the optimum incubation time. So we selected this range of time as the end time for simultaneous kinetic study (Fig. S3).

3.4. The effect of temperature

The results (Fig. S4) showed that by increasing the temperature the aggregation was also increased. Because the aim of our study was simultaneous kinetic analysis of MOR and IBU using PLS, a temperature should be selected so that a good difference between the kinetic profiles of the analytes is attainable. Based on the results 25 °C was selected for further experiments.

3.5. PLS model development

PLS calibrations for both drugs were constructed by using non-linear iterative partial least squares (NIPALS) algorithm. A training set of 27 standard samples (20 samples as calibration set and 7 samples as prediction set) were taken from different mixtures of MOR and IBU. The correlation between different calibration samples has to be avoided because collinear component in the training set data will tend to cause under-fitting in the PLS models [39].

3.6. Linear range of calibration curves

For the mentioned purposes, under optimum experimental conditions, a typical calibration curve was obtained for the determination of MOR and IBU separately. The calibration curves have a linear range of 1.33–33.29 and 0.28–6.9 µg/mL with y=−0.0127x+1.2291, R²=0.9904 and y=−0.0849x+1.1965, R²=0.9902 (Fig. 2) with the detection limit of 0.15 µg/mL (n=5) and 0.03 µg/mL for MOR and IBU respectively. Table 1 shows a comparison between the results obtained by the present method and those obtained by other methods reported for the determination of these drugs. As can be seen in Table 1, present method had a good detection limit and linear range compared to those obtained from reverse phase-high performance liquid chromatography (RP-HPLC), electrochemical method, liquid chromatography- mass spectrometry (LC-MS-MS), gas chromatography- mass spectrometry (GC-MS) and direct aqueous derivatization methods [40], [41], [42], [43], [44], [45], [46], [47]. It should be highlighted that the major advantages of this method were use of very simple method for simultaneous determination of these drugs.

Fig. 2.

Calibration curve (a) morphine 1.33–33.29 µg/mL and (b) ibuprofen 0.28–6.9 µg/mL. Conditions: temperature 25 °C, ionic strength 1 mmol/L, reaction time 10 min, pH 6: AuNPs, 10 n mol / L.

Table 1.

Linear range and limit of detection of the proposed method in comparison with some previous methods for analysis of morphine and ibuprofen.

Compound	Method	Linear range (μg/mL)	LOD (μg/mL)	Reference
Morphine	RP-HPLC	0.15–2	0.05	[41]
	LC-MS-MS	0.002–2	0.001	[42]
	Direct aqueous derivatization	0.008–5.0	0.002	[43]
	GC-MS	0.02–20	0.003	[44]
	Spectrophotometry	0.025–2	0.002	[45]
Ibuprofen	Derivatives of the ratio spectra method	2–32	0.53	[46]
	HPLC	6.1–200	1.7	[47]
	Spectrophotometry	0.28–6.9(IBU)	0.03(IBU)-	This work
		1.33–33.29 (MOR)	0.15(MOR)

RP-HPLC: Reverse phase-high performance liquid chromatography; LC-MS-MS: Liquid chromatography-tandem mass spectrometry; GC-MS: Gas chromatography-mass spectrometry; HPLC: High performance liquid chromatography

3.7. Simultaneous determination of MOR and IBU

As mentioned above, in order to to select the number of factors in PLS algorithm a cross-validation method leaving out one sample at a time, was employed. Increase in the intensity of the absorbance at 620 nm was directly related to the level of MOR and/or IBU in the sample (Fig. 3a–b). The rate of aggregation of the synthesized AuNPs by citrate were correlated to MOR and IBU level by different rate constants. The calibration and prediction sets were collected by monitoring the absorbance increase at 620 nm and were used in PLS process. For the mentioned set of 20 sample kinetic profiles, PLS1 and PLS2 algorithms were applied and by these calibrations the amount of the samples left out during the calibration process was calculated. Changes in predicted residual error sum of squares (PRESS) in PLS2 calibration as a function of the number of PLS latent variables are given in Fig. 4. As can be seen 3 components were enough to construct the PLS2 model. Nonlinearity in the absorbance–concentration relationship and interaction between the factors could be considered as the other sources of chemical factors.

Fig. 3.

Change in absorbance of AuNPs at 520 and 620 nm versus time (kinetic profiles) by injection of morphine and ibuprofen. Conditions: morphine/ ibuprofen concentration: 5.0 µg/mL, temperature 25 °C, ionic strength 1 mmol/L, pH of 6: AuNPs, 10 n mol / L.

Fig. 4.

Plot of PRESS against the number of factors for simultaneous determination of morphine and ibuprofen.

The predicted values of MOR and IBU levels in the calibration and prediction samples and their corresponding relative prediction errors are listed in Table 2., Table 3., Table 4.. It was observed that the predicted values were very close to the actual amounts and the relative prediction errors were lower than 5.0%. This confirmed the success of PLS regression for accurate prediction of MOR and IBU amounts in samples. It would be beneficial to evaluate whether the results of PLS1 and PLS2 multivariate calibration resulted in a more appropriate model. For this comparison and the validation of the models, some statistical parameters including root mean square error of prediction (RMSEP), root mean square error of cross validation (RMSECV) and root mean square error of calibration (RMSEC) were calculated. The calculated statistical parameters are given in Table 4. One can see from Table 2., Table 3. that the calibration and prediction results of PLS are more consistent than those of univariate calibration.

Table 2.

The level of reference value of IBU and MOR in the prediction set by PLS modeling of kinetic profiles.

Reference value (µg L−1)			Predicted value (µg L−1)
No.	IBU	MOR	PLS1		PLS2
			IBU	MOR	IBU	MOR
1	460	230	418.9	221.5	418.98	221.14
2	530	280	555.06	285.5	555.07	285.49
3	300	160	289.43	156.9	289.46	156.8
4	660	360	652.4	359	652.4	359.01
5	660	370	663.26	367.7	663.25	367.83
6	580	330	594.99	332.4	594.96	332.57
7	600	330	600.3	333.4	600.35	333.54
	R.S.E. (%)		3.55	1.45	3.56	1.49
R.S.E.t (%)						3.19

Table 3.

The level of reference value of IBU and MOR in the calibration set by PLS modeling of kinetic profiles.

Reference value (µg L−1)			Predicted value (µg L−1)
No	IBU	MOR	PLS1		PLS2
			IBU	MOR	IBU	MOR
1	700	400	704.59	365.16	704.62	365
2	520	260	524.96	262.55	524.96	262.54
3	370	190	364.83	190.68	364.83	190.7
4	290	160	293.77	165.03	293.77	165.06
5	320	170	320.16	171.75	320.15	171.77
6	320	170	315.2	173.44	315.2	173.47
7	350	200	346.6	193.8	346	193.82
8	460	230	464.14	232.29	464.14	232.28
9	450	270	453.87	274.37	453.87	274.38
10	740	290	737.16	299.62	737.16	299.52
11	180	80	152.7	79.4	152.64	79.94
12	170	100	152.46	103.04	152.49	102.89
13	170	100	172.58	100.65	172.57	100.67
14	330	220	331.5	218	331.52	217.95
15	350	200	346.44	199.97	346.44	200
16	160	90	162.17	91.43	162.17	91.45
17	240	130	241.6	133.55	241.6	133.56
18	650	370	655.49	376.18	655.49	376.22
19	640	360	639.75	367.63	639.74	367.69
20	200	100	194.34	110.29	194.34	110.34
	R.S.E. (%)		1.91	4.03	1.92	4.04
R.S.E. t (%)							2.55

Table 4.

Statistical parameters of the PLS1 and PLS2 calibration models developed for simultaneously determination of IBU and MOR with kinetic data.

	Model of prediction
Statistical parameters	PLS1 (%)	PLS2 (%)
RMSEP	4.7	4.69
RMSECV	4.1	4.2
RMSEC	5.1	5.2

The relative standard error and the total relative standard error were calculated using following equations and represented in Table 2., Table 3..

$$R.S.E.\left(\%\right)=\left[\frac{{\sum }_{J=1}^N\left(C\mathrm{^}j-\mathit{Cj}\right)2}{{\sum }_{J=1}^N\left(\mathit{Cj}\right)2}\right]1/2\times 100$$

$$R.S.E.t\left(\%\right)=\left[\frac{{\sum }_{J=1}^M{\sum }_{J=1}^N\left(C\mathrm{^}\mathit{ij}-\mathit{Cij}\right)2}{{\sum }_{J=1}^M{\sum }_{J=1}^N\left(\mathit{Cij}\right)2}\right]1/2\times 100$$

3.8. Interference effect

The influences of foreign coexisting substances such as naproxen, ascorbic acid, tramadol, codeine, acetaminophen, saccharides, amino acids and ions were tested. As listed in Table S1, some of the examined coexisting substances had remarkable interference on the assay. From the results, the interference of naproxen, ascorbic acid, Na₂NO₂, tryptophan, tyrosine, glucose, sucrose, fructose and lactose were very weak. Among the tested substances K⁺, Na⁺, NO₃^-, I^-, Cl^-, Mg²⁺, Fe³⁺ and Ca²⁺, codeine and cysteine could be present with relatively higher concentrations without any problems but cefexime, ceftriaxone, NH₂OH, Mn²⁺, Cd²⁺, SO₄^2-, Ca²⁺,Zr²⁺,Co²⁺, Zn²⁺, Ni ²⁺, Al ³⁺, Fe ²⁺ and Cu²⁺ could only be allowed with relatively low concentrations. The allowed concentrations of these interfering substances however, were still rather higher than those of MOR and IBU which indicated that this method had a good selectivity between drugs and other species.

3.9. Real sample analysis

In order to test the applicability of the proposed method, it was applied to determine IBU and MOR in spiked serum and urine samples. The constructed PLS model was applied to estimate the concentration of IBU and MOR in these spiked samples. The obtained results showed good recoveries (99.4–110.7 ℅) (Table S2). The results demonstrated the potential applicability of this method for simultaneous detection of IBU and MOR in real samples.

3.10. Conclusion

The SPR of the AuNPs, synthesized by the reduction of gold ion with citrate, were used as a novel analytical tool for the determination of drugs and species based on the aggregation of these nanoparticles. A direct relationship was found between the aggregation rate recorded in λ_max at about 520 and 620 nm. Multivariate calibration modeling of the kinetic absorbance data by PLS regression produced accurate results and the relative prediction errors were almost lower than 5%. In comparison with available analytical methods for simultaneous determination of IBU and MOR, the proposed method had following advantages: (i) it needed lower amounts of reagents; (ii) the proposed method had a simple, fast, and low cost procedure; and (iii)in contrast to chromatographic methods, this method did not need any expensive apparatus.

Appendix A. Supplementary material

Supplementary material

.