Article pubs.acs.org/ac
Development of a Cyclic System for Chemiluminescence Detection Runkun Zhang, Yufei Hu,* and Gongke Li* School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou 510275, China S Supporting Information *
ABSTRACT: In this paper, we described a new concept of cyclic chemiluminescence (CCL) detection, and a homemade system was designed to realize such detection. The direction of the carrier in the CCL system is in a state of periodical change that can trigger a succession of chemiluminescence (CL) reactions in a single sample injection. Therefore, in contrast to the traditional CL detection, which only records a single signal, CCL allows us to obtain multistage signals. To evaluate the new method, the cataluminescence (CTL) reaction of the volatile organic compounds (VOCs) on a nanosized catalyst was selected as the analytical model. We found that each CCL reaction has a unique exponential decay equation (EDE) to describe the change law of its multistage signals. Further study showed that the initial amount (A) of the EDE is linear with the analyte concentration, while the decay coefficient (k) is a characteristic constant for a given reaction. The formation mechanism of the exponential function and the determinants of the decay coefficient were discussed in detail. As a distinct application, CCL is capable of rapidly discriminating various analytes and even structural isomers.
C
emission,26−28 we wonder what would happen if the direction of the carrier was change periodically to recycle the residual chemical species into the reaction cell repeatedly. Our hypothesis was that a succession of signals would be detected if the residual chemical species at different stages could flow through the reaction cell periodically. In contrast to the single detection, some information related to the characteristics of CL reactions that are unobservable would become observable after cyclic detection, just as the differences in physical fitness among different people can be displayed through intense exercise, such as marathon. To confirm this hypothesis, we designed a new system to realize cyclic chemiluminescence (CCL) detection. As a proofof-principle work, the cataluminescence (CTL) reactions of the volatile organic compounds (VOCs) on nanosized catalysts were studied to demonstrate the new method, because this type of CL reaction only consumes the sample and oxygen from the air,29 which is easy to perform for this exploratory work. We found that CCL detection can obtain multistage signals in a single sample injection, and each CCL reaction has a unique exponential decay equation (EDE) to describe the change law of its multistage signals. The EDE allows facile identification and quantification of a wide range of analytes due to the decay coefficient (k) of the EDE is a characteristic constant for a given reaction, and the initial amount (A) is linear with the analyte concentration. Significantly, the k-values also allow readily resolving structural isomers. Our strategy undoubtedly opens
hemiluminescence (CL) has attracted extensive research interest and has been successfully applied to trace quantities of analytes in various samples.1−3 It has numerous advantages, such as superior sensitivity, wide linear dynamic range, and the lack of background scattering light interference.4 In addition, CL analysis has a wide detectable range of analytes, including metal ions,5 small inorganic or organic compounds,6−9 protein,10,11 as well as DNA.12,13 In recent years, the application of nanomaterials has greatly driven the development of CL analysis. It was found that nanoparticles possess the ability to act as a catalyst,14−16 reductant,17 energy acceptor,18 and chemiluminescence resonance energy transfer platform.1,12,19 The expanding availability of nanoparticles in CL has greatly extended the CL reaction system, because of their high surface areas, good adsorption characteristics, high activity, and high selectivity, which are beneficial to improve the CL quantum yield.20 Since CL is usually produced via rapid chemical reaction, which leads to imprecise measurements, as a result of irreproducible mixing of the samples and reagents, but when coupled with flow injection (FI) analysisnamely, flow injection chemiluminescence (FI-CL)this analytical technology can provide a reproducible, automated, and precise means for rapid detection.21,22 In the FI-CL system, the analyte is injected into a moving, unidirectional carrier, and then mixes with the CL reagent to produce CL emission.23,24 The CL signal must be measured in the reaction cell, while the analyte leaves the reaction cell quickly without complete decomposition, because of the driving force of the carrier. Therefore, only a section of the total intensity can be measured by the single detection.25 Since the residual chemical species of the CL reactions also can react with CL reagent to produce CL © XXXX American Chemical Society
Received: April 5, 2014 Accepted: May 19, 2014
A
dx.doi.org/10.1021/ac5012359 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 1. (A) Schematic diagram of the FI-CL system. (B) Typical result of FI-CL detection in a single sample injection. (C) Schematic diagram of CCL system. (D) Schematic result of CCL detection in a single sample injection.
up new possibilities for CL detection and will find wide applications in related fields.
Method for Data Processing. The data acquisition time for each signal point was set as 0.5 s. The data were recorded by a computer and further processed with Origin 8.0. To confirm the change law of the CCL signal, the peak intensity and the peak time were considered as dependent variable and independent variable, respectively. The mathematical relationship between the two variables was fitted by different mathematical models. Data analysis indicated the CCL signal satisfies the exponential decay law, which has the following original mathematic form:
■
EXPERIMENTAL SECTION Instruments. An ultraweak luminescent analyzer with a photomultiplier tube (PMT) was purchased from the Institute of Biophysics, Academia Sinica (Beijing, China). A peristaltic pump (Model BT100-1J) was purchased from Leader Instruments Co., Ltd. (Baoding, China). A time relay (Model DH48S-2Z) was purchased from Kangtai Electronic Technology Co., Ltd. (Xuzhou, China). Alternating current contactors (Model CJX2-1801) were supported by Huipin Electric Appliance Co., Ltd. (Wenzhou, China). Gas chromatography−mass spectrometry (GC-MS, Model QP2010) was purchased from Shimadzu Corporation (Japan). Sample Preparation. The gas samples were prepared by vaporizing organic liquid samples. In brief, certain microliter levels of liquid samples were injected into the airtight bottles of 650 mL, and then the airtight bottles were placed into a drying oven at a temperature of 100 °C for 5 min to fully vaporize the liquid samples. Finally, 2.5-mL gaseous samples were extracted from the airtight bottles for CL analysis. The detailed information on the organic reagents can be seen in the Supporting Information.
y = A e−x / k + y0
(1)
where y is a dependent variable, x is an independent variable, A stands for the initial amount, and k and y0 are constant parameters. Method for Mechanism Study. In order to demonstrate the principle of the CCL mechanism, the change trends of the concentrations of the residual compounds from the CL reactions of acetone and isobutanol on MgO surface were investigated. The residual compounds were collected at certain times into sampling bags of 1 L. The sampling time was set at 6 min at a rotational speed of 44 rpm. The sampling time was set long enough to ensure all the residual compounds from the CL reactions can be collected into the sampling bags. In order to B
dx.doi.org/10.1021/ac5012359 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 2. CL kinetic curves of four reactions in the FI-CL and CCL systems. Conditions: reaction temperature, 220 °C; detection wavelength, 425 nm; rotational speed: 44 rpm; and the concentrations of acetone, 2-butanone, n-butylether, and isobutanol are 200, 200, 68, and 4.8 ppm, respectively.
and the flow rate into consideration. If the period was set too short, the CL peaks of the adjacent stage CL reactions could not be resolved effectively. If the period was set too long, the sample would leak out of the system and the results in it could not be recycled into the reaction cell again. Here, we set a standard for the determination of the cycle period, which is defined as the time that the analyte is retained in the system without changing the carrier direction. Therefore, the cycle period of a given reaction should be optimized according to its optimum flow rates. However, in order to compare the difference between the k-values of different CL reactions, all of the CL reactions must be performed under the same conditions. In the present work, air was used as carrier, and the cycle period was set at 20 s at a rotational speed of 44 rpm of the peristaltic pump. Development of the CCL detection. The CCL detection can be realized in the above homemade system. We first investigated the CL reactions of acetone, 2-butanone, nbutylether, and isobutanol on nanosized MgO surface both in the CCL and FI-CL systems to demonstrate the merit of the new method. As shown in Figure 2, only a single peak signal on the kinetic curves is recorded by FI-CL. According the CL theory, the CL intensity can be expressed as25,30
ensure the compounds at low concentration also can be detected, a commercial solid-phase microextraction fiber (polydimethylsiloxane/CAR) was used for the extraction and enrichment of the chemical species in the sampling bags. Finally, the samples were analyzed by a GC-MS instrument. The instrument operating parameters of the GC-MS experiments are shown in Table S1 in the Supporting Information.
■
RESULTS AND DISCUSSION Fabrication of the CCL System. The design idea of CCL system is inspired by the direction of the carrier in the FI-CL system being unidirectional (Figure 1A), and the sample only undergoes a single detection process, thus only a single peak signal is measured (Figure 1B). Thus, the key to the design of the CCL system is to realize the periodical change of the direction of the carrier. As shown in Figure 1C, the CCL system consists of a laboratory-made reaction cell (the detailed procedure of the fabrication of the reaction cell can be seen in the Supporting Information), a computerized ultraweak luminescent analyzer with a PMT detector, and the carrier controller. The carrier controller is composed of a peristaltic pump, three of the same miniaturized three-way magnetic valves, a time relay, and two current contactors (AC contactor). The reaction cell is connected to the three-way magnetic valves by polytetrafluoroethylene (PTFE) hose tubes (0.3 mm × 0.2 mm) to form a return circuit. When the three-way magnetic valve is energized, valve ports a and c are connected, while valve port b is closed. When the power switch is off, valve ports a and b are connected, and valve port c is closed. The power sources of these valves are controlled by two asynchronous alternating AC contactors. Thereinto, valves V2 and V4 are controlled by the same AC contractor, and valve V3 is controlled by another AC contractor. By using a time relay to control the working states of the AC contactors, the direction of the carrier can be changed automatically at the indicated cycle period, as indicated by the red and purple arrows in Figure 1C. Thus, the chemical species at different stages can flow through the reaction cell repeatedly to trigger the first stage, second stage, ..., nth stage; the CL reaction proceeds in succession, and the corresponding CL signals can be measured over time (Figure 1D). The cycle period is an important parameter of the system and can be set any value from 0.01 s to 99 h (depends on the adjustable time range of the time relay). However, the determination of cycle period must take the pipe specifications
I=Φ
dc dt
(2)
In this equation, I stands for CL intensity, Φ is the CL quantum yield, and dc/dt is the rate of the consumption of the CL precursor. In this case, the CL peak signal is a dimensionless value that is a function of the concentration of analyte. It is difficult to extract more information from the single signal for analysis, for example, to extract some information for the discrimination of the four analytes. But for the same CL reaction in the CCL system, a succession of peak signals at different stages can be observed on the kinetic curve, as shown in Figure 2. It can be seen that these signals decay exponentially from the maximum to background noise over time. Further study found that the exponential decay equation (EDE) describes the change law of the multistage signals that fits an expression of the following form: In = A e − t / k + I 0
(3)
with C
dx.doi.org/10.1021/ac5012359 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 3. (A) A values and (B) k values of eight types of VOCs on the surfaces of different nanomaterials. Conditions: reaction temperature, 220 °C; detection wavelength, 425 nm; and rotational speed, 44 rpm. VOC concentrations: cyclopentanone, 10 ppm; acraldehyde, 50 ppm; 2-butanone, 200 ppm; acetone, 200 ppm; n-butanal, 24 ppm; n-pentane, 300 ppm; n-butylether, 68 ppm; and ethanol, 120 ppm.
Figure 4. CCL kinetic curves of four analytes at different concentrations on nanosized MgO surface. Conditions: reaction temperature, 220 °C; detection wavelength, 425 nm; and rotational speed, 44 rpm.
t = t P − tmax = P(n − 1)
given CL reaction. In addition, t is the calibration time, tp is the time of each peak on the CL kinetic curve, and tmax is the time of the maximum CL intensity. The purpose of the use of calibration time is to use tmax as a zero time reference. The parameter n stands for the number of cycles (n = 1, 2, 3, 4, ...); it also stands for the number of the reaction stages, due to each
(4)
where In is the CL intensity of reaction stage n, I0 the background noise, A the initial amount (the physical significance of A is the maximum CL intensity), and k a decay coefficient, which is the measurement of the decay speed of the CL signal. As detailed below, k is a unique constant for a D
dx.doi.org/10.1021/ac5012359 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 5. Trends of chromatography peak areas of different compounds versus time. (A) The trends of peak areas of acetone versus time when acetone has initial concentrations of 118 and 684 ppm, respectively. (B) The trends of peak areas of isobutanol and its products versus time when isobutanol at an initial concentration of 38 ppm. (C) The trends of peak areas of isobutanol and its products versus time when isobutanol at an initial concentration of 100 ppm. The trends of peak areas of all compounds satisfy the exponential decay law, indicating that the trends of concentrations also satisfy the exponential decay law. Acetone is under a state of formation and consumption during the oxidation of isobutanol, leading to the observation that the 1/λ value is smaller than that of pure acetone under the same conditions.
cycle, that can trigger a corresponding reaction; P is the cycle period. The exponential decay curves of the CL signal show that the decay speeds of the four analytes on MgO surface differ from each other. For example, the speed of CL signal for acetone (k = 15.5) decays back to background noise faster than that of 2butanone (k = 25.3), although their concentrations are the same. The concentration of n-butylether is less than that of acetone and 2-butanone, but its decay speed (k = 17.2) is between that of the two compounds. Although isobutanol has the lowest concentration, it has the slowest decay speed (k = 35.4). These results indicate that the CCL can provide more information for analysis. We next investigate the CCL reactions of eight types of VOCs on different nanomaterial surfaces to further study the change law of the multistage signals. It was found that the multistage signals of different reactions also satisfy the exponential decay law. The values of A and k of the EDE for different CL reactions are plotted in Figure 3. It can be seen that both A and k are diverse for different CL reactions; that is, the values of A or k are different for a given analyte on different nanomaterials, and the same nanomaterial exhibits different values of A or k upon exposures to different analytes. For example, acraldehyde produced strong signal on SrCO3, but no CL emission on Y2O3 and ZrO2. The k-values of 2-butanone and cyclopentanone on MgO surface are close, but their kvalues are quite different on other nanomaterials. These results fully demonstrate that each CCL reaction has a corresponding equation to describe the change law of its multistage signals (the CL reaction must produce strong enough signals). The reproducibility of the new method was investigated by determining the A and k values of four different CCL reactions five times. The relative standard deviations (RSDs) of both A and k for different reactions were
