- Email: [email protected]
Development of a colorimetric sensor Array for the discrimination of aldehydes
Sensors and Actuators B 196 (2014) 10–17
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Development of a colorimetric sensor Array for the discrimination of aldehydes Junjie Li a , Changjun Hou a,∗ , Danqun Huo a , Mei Yang a , Huan-bao Fa b , Ping Yang c a Key Laboratory of Biorheology Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400044, PR China b College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400044, China c National Engineering Research Center of Solid-State Brewing, Luzhou Laojiao Group Co. Ltd., Luzhou, Sichuan, 646000, PR China
a r t i c l e
i n f o
Article history: Received 21 October 2013 Received in revised form 15 January 2014 Accepted 19 January 2014 Available online 25 January 2014 Keywords: Aldehyde Sensor array Lung cancer Formaldehyde pollution
a b s t r a c t A brand-new colorimetric sensor array was developed based on cross-reactive mechanism to discriminate 9 kinds of aldehydes and 16 volatile organic lung cancer biomarker candidates in low concentration. Sixteen out of twenty-seven dyes were selected to mix with 2,4-dinitrophenylhydrazine (DNPH) in optimized concentration to serve as specific sensing elements, and polyethylene glycol-1000 was chosen from four kinds polyethylene glycols to act as stabilizer. Resultant sensor array shows improved response (about tenfold higher) to aldehydes than former study, and exhibited very good selectivity in low concentration ranging from 40 ppb to 10 ppm with the presence of interfering counterparts. Data analysis was performed by both hierarchical cluster analysis (HCA) and discriminant analysis (DA), which demonstrates the excellent discrimination ability of the sensor to structurally similar aldehydes related to lung cancer. Besides, formaldehyde-spiked air samples were analyzed with developed sensor array, suggesting promising utilization potentiality to monitor such toxic gas. Theoretical detection limit was down to 8.2 ppb with a liner range from 10 ppb to 150 ppm. The sensor array can be further developed for early diagnosis of lung cancer as well as monitoring of domestic and industrial formaldehyde pollution. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Volatile aldehydes are among the most common source of pollution in daily life due to their widespread presence in a plenty of chemical additives, industrial processes and incomplete combustions [1]. Acknowledged as an important components of toxic substances, aldehydes can be widely found in environmental atmospheres [2], industrial materials [3], alcoholic beverage [4], as well as human metabolites [5–7], which testify their significance in environmental, industrial and biological applications. In the area of environmental monitoring, formaldehyde poses a great threat to human health. It can lead to intensive irritant reactions of eyes and mucous membranes, and probable human carcinogen [8,9]. In view of its widespread use, toxicity and volatility, it is therefore highly problematic as an indoor pollutant. Related organizations thus banned it from use in certain applications (such as processing systems, preservatives for liquid-cooling, and antifouling products), and set the maximum allowed concentrations. The safe-exposure standard set by the World Health Organization is 80 ppb averaged over 30 min. While the permissible exposure limit
∗ Corresponding author. Tel.: +86 2365112673; fax: +86 2365102507. E-mail address: [email protected] (C. Hou). 0925-4005/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.snb.2014.01.054
(PEL) set by Occupational Safety and Health Administration (OSHA) is 750 ppb and the immediately dangerous to life or health (IDLH) limit is 20 ppm [10]. Besides, there is an increasing interest concerning the role of volatile aldehydes as biomarker in biomedical field. According to previously studies, there is a significant difference in the composition of volatile organic compounds in human breath and other metabolites. Discrimination of those volatile biomarkers would undoubtedly provide a non-invasive alternative for the diagnosis of lung cancer. Despite controversial debate in accurate VOCs biomarkers, partially due to exogenous interferences and individual difference, it is highly convinced that volatile aldehydes are very likely to be among lung cancer biomarkers [11]. For example, solidphase micro-extraction–gas chromatography–mass spectrometry method had been utilized for the determination of endogenous hexanal and heptanal in urine samples [12]. The results demonstrated that hexanal urinary concentrations in cancer patients were slightly higher than those found in control group ones. Hexanal and heptanal have also been measure in serum samples, which suggested that their concentration levels in the samples of lung cancer patients were sharply higher than those in healthy people [6]. In addition, it was also found that the levels of C3–C9 aldehydes in exhaled breath were increased in the non-small cell lung cancer (NSCLC) patients without slight effect of smoking habits and age
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
[7]. So that straight aldehydes might be better biomarkers associated with NSCLC than other VOC patterns. Since the concentration of volatile aldehydes as toxic pollutant and cancer biomarkers are extremely low, normally several ppb to a hundred ppm, detection of aldehyde at very low concentration level is of great importance for both environmental surveillance and biomedical diagnosis. A plenty of techniques and methods have been devised for the detection and measurement of gaseous aldehyde. Traditionally, aldehydes can be analyzed by standard techniques such as liquid chromatograph [4,13], gas chromatograph [3,12], and electrochemical techniques [1,14]. Based on large instrumentation, those methods do offer very good choices and fine results to analyze this group of compounds. However, despite high sensitivity and good reproducibility, above-mention method all involves unavoidable drawbacks like expensive and cumbersome operations and also time-consuming procedures, which is adverse for real-time and field use. There are also inexpensive techniques including colorimetric or fluorometric methods with the application of pigments, and so forth [2,15,16]. But lack of sensitivity greatly hinders those methods from wide application. Moreover, they also need a relatively long detection time and boring steps to get final results. Thus, there still remains a pressing need for the development of a sensitive, rapid, inexpensive and facile method for aldehyde detection. In the past decades, there is an ever-growing interest in the study on the diagnosis of lung cancer by analysis of VOCs in human breath for that it is non-invasive, easy to handle and time-saving [17]. MS techniques (whether or not coupled to chromatographic equipment) provide a reliable method for the determination of mixtures of VOCs and semi-VOCs in different human metabolites [18]. But high-cost and expert interpretation greatly hinders them from wide application and also prolongs the research period. Therefore, many researchers prefer to techniques involving smaller, less expensive equipment, which is simple to handle and able to be miniaturized, particularly sensors or sensor arrays [19,20]. Outstanding studies include polymer-coated surface acoustic wave sensors [21], coated quartz crystal microbalance sensors [22], gold nanoparticles sensor array [23], and also colorimetric sensor array [24]. Selective of semi-selective electrochemical sensors have been proved to be a good choice for breath analysis [25], and after signal compensation it can largely reduce the influence resulted from humidity fluctuation [26]. Yet there are still challenges to develop such sensor systems to deal with various confounding factors in real-world analysis [27,28]. Since VOCs can be influenced by various internal and external factors such as individual physical conditions (age, respiration rate, nutritional status and so on), smoking circadian rhythm, and composition of environmental air, definition of normal composition and ranges of single VOCs in human breath gas remains a tricky problem [29]. Therefore, sensor systems mentioned before all involve the same limitations, that is, they are focused on the overall composition of the VOCs in the metabolites thus lose the specificity of each VOCs marker, which would definitely restrict the selectivity of the sensor and result in false positive/negative discrimination. Accordingly, by comparing the presence/absence of each experimentally validated VOCs in the LC states relative to the control states, most of possible VOC lung cancer biomarker candidates can be divided into seven compound families: hydrocarbons, alcohols, aldehydes, ketones, esters, nitriles and aromatic compounds [11,30]. A feasible strategy to solve above-mentioned problem is to develop specialized sensors for each biomarker family members and then integrate separated sensors to a sensor system. In view of this thought, we employed colorimetric sensor array as detection technique, and firstly chose aldehydes and then other VOC family as biomarkers to design the sensor, which might lead to breakthrough in this research area.
11
Accordingly, since colorimetric sensor array systems are broadspectrum sensors, overfitting, which is mainly resulted from sensor dots that are widely responsive, of the training data set poses a big problem [27]. In view of this problem, a feasible way is to improve the selectivity of each sensor dot in the array to selected volatile biomarker. Given that most aldehyde can be easily derivatized with many colorimetric reagents, it is possible choose aldehydes as target biomarker to carry out visual colorimetry, and allows easy measurements [15]. A good case in study is an early report of a highly sensitive colorimetric method for fast formaldehyde detection [10]. Using amine-terminated PEGs as specific agent to create a reactive matrix, a simple pH indicator array was applied to detect the change in basicity upon the reaction of a non-volatile primary amine with formaldehyde. It eventually gave a detection limit down to 50 ppb within 10 min. Nevertheless, apart from its great success in formaldehyde detection, the sensor array cannot response to aldehydes with longer carbon chain or molecular weight. Therefore, it can be hardly utilized for longer-carbon-chain aldehydes detection in biomedical application. Herein, we fabricated a more sensitive sensor array using DNPH as aldehyde-specific sensing elements and polyethylene glycol-1000 as stabilizer. The sensor showed very good selectivity and sensitivity towards aldehydes among other VOCs lung cancer biomarker candidates. It can also discriminate nine kinds of C1–C7 aldehydes in low concentration ranging from ppm level to ppb level, using Hierarchical Cluster Analysis (HCA) and Linear Discriminant Analysis (LDA). Besides, formaldehyde-spiked air samples were also analyzed with asprepared sensor array, exhibiting very low detection limit and good linear range. The sensor array would provide an efficient tool for early diagnosis of lung cancer, and monitoring of domestic and industrial formaldehyde pollution. 2. Material and methods 2.1. Chemicals and preparation of stock solutions Four porphyrins were obtained from Frontier Scientific (Logan, UT, USA), and the other indicator dyes (listed in table S1) were supplied by Sigma–Aldrich (St. Louis, MO, USA). Porous hydrophobic membrane used for dye staining was bought from Milli-pore Co. Ltd. (Bedford, MA, USA). 2, 4-dinitrophenylhydrazine (DNPH), polyethylene glycol-400 (PEG-400), polyethylene glycol-800 (PEG800), polyethylene glycol-1000 (PEG-1000), polyethylene glycol4000 (PEG-4000), aldehyde solutions, and other gas solutions were purchased from Jinchun Industry Co., Ltd. (Shanghai, China). All the other chemicals are of analytical pure and used without further purification. Ultra-pure water was generated by a Millipore Direct-Q Water system (Molsheim, France). Accordingly, the reaction between aldehyde and DNPH is catalyzed by acid [31]. At high acid concentrations the carbonyl group of aldehyde is activated by the acid but at lower acid concentrations, or in basic solutions, the carbonyl group becomes less reactive, which hinder aldehyde from dehydration to the hydrazones. Although HClO4 has been frequently used as catalyst to derivatize aldehydes, we use H2 SO4 instead for the sake of stability of resultant solution. Typically, 0.4 g DNPH was added to a mixed solution of 3 mL water and 10 mL ethanol, and then 2 mL concentrated sulfuric acid was added dropwise. The solution was stirred for 10 min and then went through filter paper to obtain a yellow DNPH store solution. For prophyrin solution, 10 mg dye was added into 3 mL DMF solution and stored in dark place before use. Water solutions were used instead for alizarin and nitrazine yellow while all the other dye solutions were prepared using ethanol solution with the same concentration. Gas samples were prepared by a self-made gas distribution device. Briefly, saturated vapor of each aldehyde was produced by dilution with dry and wet nitrogen to
12
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
Fig. 1. Actual difference maps of sensor arrays: (A) selection-oriented sensor array with five parallel dots; (B) detection-oriented sensor array and arrangement of the sensing dots.
achieve the desired concentration, which was testified before-hand by GC-MS, in a gas stream at 50% relative humidity (RH). 2.2. Selections of dye and dye/DNPH ratio In order to realize a high-efficient and less-expensive selection of dyes and dye/DNPH ratio, we first fabricated a selection-oriented sensor array. The sensor array was prepared as following: 50 L dye stock solution and 50 L PEG-400 was mixed with DNPH solution in a 3 mL plastic vial with different dye/DNPH ratio (1:5, 1:4, 1:3, 1:2, and 1:1). The mixture was kept in ultrasonic agitation for ten seconds and then a quartz capillary was used to deliver approximately 0.1 L solution onto the surface of a porous hydrophobic membrane to fabricate a 5 × 5 sensor array (five parallel dots for each ratio, see Fig. 1A). Once printed, the arrays were placed in a 500 mL beaker saturated in nitrogen atmosphere for 30 min and subsequently dried in a 60 ◦ C oven for 24 h after which the oven temperature was reduced to 35 ◦ C and the arrays left for another 24 h. Detailed working principle, measurement of the optical signal, and also data processing have been introduced in our early studies [32,33]. Briefly, we use a complementary metal-oxide semiconductor (CMOS) camera to obtain the color information of the sensor array. When reacting with gas molecule, the color of certain spots would change. By comparing the array images obtained before and after interaction, discrimination of different gases in different concentrations can be easily achieved. Specially, we constructed a gas processing system for the analysis of the VOCs in this study. It consists of a circular gas tank to control gas concentration and humidity, a gas chamber for chemical reaction between the gas and the sensor array, a waste cylinder for the collection of exhaust gases, and a pump to provide driving force for the gas flow. Upon detection, the gas system was first cleaned using N2 before-hand in order to preclude the interference of the air. Selected VOC in given concentration was then prepared by the gas processing systems and reacted with the colorimetric sensor array. After exposure of the array to the as-prepared VOCs, the sensor array images were automatically collected by CMOS camera at given time point. Difference map of each VOC in given concentration subsequently was generated by the image-processing system for further data analysis. Hexanal samples were chosen to testify the relative response and stability of each dye in different dye/DNPH ratio. Specifically, nitrogen gas was applied as control gas and hexanal gas was mixed with ultra-pure nitrogen to a concentration of 5 ppm in a 3 L Tedlar bag. The bag was then connected with the sensor system upon analysis. The gas flow rate were set as 5 mL/s at room temperature (25 ◦ C) with a relative humidity of 50%. The pictures of the sensor array can be taken at set time point until the reaction ends. By comparing the initial image of the array with its final image
after exposure to targeted gas sample, color change profiles were automatically obtained. Detailed working principle of the sensory system and data process can be found in our former studies [33,34]. In order to eliminate the noise and extract useful information, a threshold T is set, which is adjustable in the image analysis software written by our lab. When R + G + B < T, the color is set as R = G = B = 0. In this section, the response threshold was set as 10, 16 kinds of dyes were selected prior to further test. 2.3. Aldehyde analysis method with the sensor array A similar strategy was applied to select the stabilizer PEGs with a sensor array (see Fig. 1B). The sensor arrays of four different PEGs were subjected to above-mentioned preparing procedure and testing method to select the best stabilizer. Optimized stabilizer was then utilized in the fabrication of the final sensor array (detectionoriented sensor arrays). It should be noted that the arrays were stored in a nitrogen-flushed humidity-controllable container with a RH of 50% and room temperature (25 ◦ C) before use. The response threshold was set as 12 in this section and following experiment so as to suppress possible interference. Aldehyde samples and 16 other VOC lung biomarker candidates in concentration of 5 ppm and 20 ppb were prepared using nitrogen gas and testified by the sensor array. Resultant experimental data was subsequently analyzed with both HCA and LDA. As for the detection of formaldehyde-spiked air, gas samples were prepared by air instead of nitrogen gas. 3. Results and discussion 3.1. Detection principle and fabrication of the sensor array The selection of dye and dye/DNPH ratio determines the sensitivity and specificity of the sensor array. As key sensory components, responsive dyes are crucial for the performance of a colorimetric sensor array. In previous studies, porphyrins, pH indicators, metal salts and metal oxides have been used to fabricate colorimetric sensor array for a variety of gaseous analytes [35–38]. The selectivity of the sensor array depends on two aspects: specialized response of dyes to different analytes and cross-responsive interaction between analytes and the sensor array, in other words, unique combination of the responses of all the dyes to target analyte. Generally, the more versatile a sensor system is, the less selective the performance for that versatility usually means more interference for a selected target. Therefore, it is in urgent need to improve the specificity of each sensor dot in the sensor array so as to improve the overall sensor performance. In view of this problem, we proposed the combination of dye and DNPH as sensing element based on the specific reaction of aldehyde with DNPH. As shown in
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
13
Fig. 2. Schematic illustration of the procedure to develop the colorimetric sensor array.
Fig. 2, we hypothesize the detection principle as following: Aldehyde reacts with DNPH resulting in the production of H2 O; H2 SO4 serves as catalyst in the reaction and then protonized by the product H2 O, which lead to the pH variation in the sensor dot. Selected dyes are sensitive to the pH change and give a color change in turn; since different kinds and amount of aldehyde have different affinity to DNPH and produce different volume of H2 O, the color change given by the dyes will be unique from each other. Measurement of color change of the sensor array thus result in the discrimination of different aldehydes and other possible interfering counterparts. Herein, we applied a high-efficient and less-expensive way to realize the selecting procedure relying on selection-oriented sensor arrays. In order to achieve the highest response as possible, we first developed a sensor array to optimize the selection of dyes and dye/DNPH ratios. As illustrated in Fig. 1A, the selection-oriented sensor array has five parallel dots for each dye/DNPH combination so that we can measure the respective response of different dye/DNPH ratios at the same time with five controls. It will largely reduce the cost to develop the sensor array and get a relatively more reliable result. Likely, we use squared Euclidean Distance change (R2 , G2 and B2 of each sensor dot) to measure the response [39]. Using hexanal as target and nitrogen gas as control, relative response and stability of twenty-seven frequently-reported dyes with five dye/DNPH ratios have been measured. Dyes that are more responsive and with better stability to hexanal upon control were subsequently selected as final sensor dots, the first 16 out of 135 combination and detailed information are summarized in Table S1. The micro morphology of the substrate PVDF film before and after dye-printing have also been observed by scanning electron microscope (Figure S4). It was shown that the PVDF film shows very uniform morphology and porosity before dye-printing. And when the dye solution was printed into the film, no significant change of the morphology or porosity was observed, indicating that the DNPH, dye and also PEGs molecules homogeneously diffused into the porous PVDF matrices without agglomeration or impairing the film. Besides the response and specificity to target analytes, stability is another important factor to determine the performance of a sensor system. PEG-400 had been used in the fabrication of colorimetric sensor array to improve the stability previously [10]. Accordingly, the benefits using PEG-400 can be concluded as two aspects, that is, protecting dye from external environment and insulating the interference of hydrophilic compounds like water vapor. In the present study, we investigated the performance of
four different PEGs aiming to choose a better stabilizer for the sensor array, and the results were summarized in Fig. S1. The squared Euclidean Distance change represents the response of each PEGs to hexanal gas while the relative deviation indicates the stability. As is shown in the figure, PEG-1000 exhibited the best response and relative lower deviation, which is indicative of good stabilizing ability. Therefore, PEG-1000 was finally chosen as stabilizer for follow-up fabrication of detection-oriented sensor array. 3.2. Discrimination of structurally similar aldehydes over other VOC lung cancer biomarker candidates When the detection-oriented sensor array was developed, discrimination of different aldehydes and other possible VOC markers was subsequently studied. Apart from the aldehyde family, most VOCs either have no carbonyl group or cannot react with DNPH to produce H2 O. The former VOCs would not interact with the sensor dots while the latter could not lead to enough pH change thus will not result in enough response that can be detected by the sensor array. Firstly, time-dependent response of the sensor array to different aldehydes was studied to establish equilibration. The total squared Euclidean Distance for each aldehyde was analyzed to track the response of as-prepared array. As can be seen in Fig. 3, squared Euclidean Distance was increasing for each aldehyde until it reached saturation. Although the reaction between the sensor array and most aldehydes saturated within 2 min, a response time of 4 min was chosen as optimized reaction time for subsequent experiments so as to ensure obtained squared Euclidean Distance in the final equilibration. Then we assessed the selectivity between aldehydes and 16 other VOC biomarker candidates, the difference maps are shown in Fig. 4 and Fig. S3. The results suggested that discriminating ability of developed sensor array was so good that one can easily tell aldehydes from other counterparts. Either in the concentration of 10 ppm or 40 ppb, there are no mis-recognition observed even by naked-eyes. Meanwhile, identification of structurally similar aldehydes remains a big challenge because they are closely alike in structures and exhibit very similar properties. Based on cross-reactive mechanism, the recognition of a target is resulted from the sum and organic combination of specific response of each sensing element, thus providing an efficient alternative to discriminate analytes with high similarity. Pattern recognition has been proved to be a powerful tool for cross-reactive analysis due to its ability to discriminate complicate mixtures. Herein, we applied the same methodology to
14
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
Fig. 3. Time-dependent response of different aldehydes.
realize discrimination between structurally similar aldehydes. To demonstrate the applicability of as-prepared colorimetric sensor array for aldehyde discrimination, nine kinds of aldehydes in the concentration of 10 ppm and 40 ppb were tested. Upon exposure to the aldehyde sample, the array underwent specific reactions, resulting in well-defined color changes in final difference maps. Color change profiles with clear difference were shown in Fig. 4. The color difference maps showed the distinctive patterns for all nine aldehydes both in the concentration of 10 ppm and 40 ppb. When the aldehyde samples were in higher concentrations, the sensor array shows obviously more responsive dots. It demonstrated that as-prepared have the potential to be developed for further analysis of VOCs in human breath for lung cancer diagnosis. Notably, it also can be seen from Fig. 3 that formaldehyde exhibited the largest response than other eight aldehydes, which possibly due to its higher affinity to DNPH. In view of the change of squared Euclidean Distance upon exposure to formaldehyde in the same concentration range, sensor array developed here exhibited
response ten times higher than former study [10]. Basically, for C1 to C8 linear-chain aldehydes, resultant response decreases with the carbon chain length increases. 3.3. HCA and LDA analysis of resultant data In order to further examine the sensory performance of asprepared sensor array, statistical methodology was utilized to analyze resultant experimental data with the application of HCA and LDA, which are widely used in related studies. For both analysis, the variables were define as relative change in each color vector (R2 or G2 or B2 divided by total squared Euclidean Distance), and totally 48 (4 × 4 × 3) parameters were used to conduct statistical analysis. In each case, the data was detected repeatedly for 5 times with 5 parallel samples. The HCA dendrograms for concentrations in 10 ppm and 40 ppb was separately shown in Fig. S3. HCA is a multivariate statistical analysis method which is comprised of agglomerative and divi-
Fig. 4. Difference maps of different aldehydes after interaction with the sensor array in the concentration of 40 ppb (A) and 10 ppm (B): a, formaldehyde; b, acetaldehyde; c, propanal; d, n-butanal; e, pentanal; f, hexanal; g, heptanal; h, acetal; i, isobutyraldehyde.
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
15
Fig. 5. Linear discriminant analysis results of nine kinds of aldehydes with five parallel samples in the concentration of 40 ppb (A) and 10 ppm (B).
sive methods that find clusters of observations within a data set. It could be seen from the figures that all 9 aldehyde gas samples were accurately identified against each other in two concentration levels. And nearly no misclassification was observed upon examination of 45 gas samples. Remarkably, for both concentration levels, five formaldehyde samples were classified into a single cluster with nearly no error. It indicated that the sensor array have very good stability for formaldehyde and can be further developed as gas sensor for such toxic gas pollution. It also showed that hexanal samples in low concentrations seemed to be more easily clustered than in high concentration. Since concentration discrepancy of hexanal between normal people and lung cancer patients was in low levels, it would benefit the diagnosis of lung cancer. As a multivariate statistical analysis method, HCA is comprised of agglomerative and divisive methods, and can be utilized to screen cluster of observations within certain data set. The results obviously suggested that the sensor array have acceptable selectivity to classify different aldehydes in low concentration. Besides, the HCA analysis is also indicative of further application of the sensor system when a database of difference maps of different aldehydes was established,
one can easily recognize an unknown aldehyde and even its concentration level just by checking the library entries. Discriminant analysis results provided another evidence to prove the good selectivity of as-prepared sensor array. Dissimilar to HCA, LDA is a supervised method for separating classes of objects and for assigning new components to appropriate classes. The discriminants are linear combinations of the measured sensor responses. Discriminant functions are calculated to maximize the Mahalanobis distance between classes relative to the variation within classes [40]. It is able to find a linear combination of features which characterizes or separates two or more classes of objects or events. Based on this mechanism, 9 kinds of aldehydes in concentration of 10 ppm and 40 ppb were classified into 9 groups (Fig. 5). In both concentration levels, all five parallel samples gather together with no misclassification. Different aldehyde groups were well separated with no clear overlap. Notably, the acetal group in high concentration was significantly farther away for other eight aldehydes than in low concentration. As acetal has completely different structure with the other ones, it is reasonable to conclude that higher concentration is beneficial for the discrimination
Fig. 6. Response of the sensor array upon exposure to formaldehyde in different concentration. Inserted figure denotes the linear fitting of concentration-dependent response in concentrations ranging from 40 ppb to 10 ppm.
16
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
of structurally dissimilar inferences. The result of LDA analysis demonstrated an extraordinary high level of classification of structurally similar aldehydes based on a simple colorimetric technique.
3.4. Application of the sensor array to monitor formaldehyde The feasibility to apply as-prepared sensor array for formaldehyde pollution monitoring was subsequently investigated. Specifically, we prepared formaldehyde-spiked air samples in concentration ranging from ppt level to ppm level, and then analyzed the samples with the sensor array. As shown in Fig. 6, total squared Euclidean distance of changes in RGB showed positive correlation with the formaldehyde concentration of gas samples, which base a semi-quantitative analysis simply on the total response of the array. Besides, linear regression analysis indicated that logarithmic formaldehyde concentration has good linear relationship with the total squared Euclidean distance in the range from 40 ppb to 10 ppm (see inserted figure in Fig. 6). Theoretical detection limit was down to 8.2 ppb and the response concentration was even as lower as ppt level. Compared to previous study, the sensor array showed less response time and it would not be difficult to improve the response limit just by increasing the exposure time of gas simple to the sensor array. Moreover, due to the specific interaction between DNPH and aldehyde, the sensor array could be hardly influenced by most gaseous interference. We analyzed a plenty of possible interfering gases yet only SO2 and NH3 in high concentrations (>250 ppm) led to significant interference in the detection. Yet the interference can be excluded simply by using basic and/or acidic absorbent before detection, or just adding basic and/or acidic controls in the sensor array. Therefore, it is promising to develop such sensor array to monitor formaldehyde pollution of low concentration in both domestic and industrial application.
4. Conclusion As a proof of concept, colorimetric sensor array system offers a feasible method to develop sensors on the basis of cross-reactive mechanism. Despite their wide application for a variety of analytes, it is crucial to select unique sensory components for specific targets so as to improve the sensor performance. In the present study, we presented the development of a brand-new colorimetric sensor array to discriminate nine kinds of structurally similar aldehydes in low concentration. When optimizing the fabrication procedure of the sensor array, we applied a selection-oriented sensor to realize a high-efficient and less-expensive selection of dyes, dye/DNPH ratio and stabilizers. Sixteen out of twenty-seven dyes in optimized concentrations were first selected to mix with DNPH to serve as aldehyde-specific sensing elements, and then polyethylene glycol-1000 was chosen from four kinds of polyethylene glycols to improve the stability. Compared with former study, our sensor array shows about tenfold higher response to aldehydes, and exhibited very good sensor performance in low concentration ranging from ppm level to ppb level, with nearly no interference of sixteen interfering VOC lung cancer biomarker candidates. Hierarchical cluster analysis and principal component analysis were further applied to analyze resultant experimental data, which demonstrates the excellent discrimination ability to structurally similar aldehydes. Besides, formaldehyde-spiked air samples were analyzed using the sensor array. Results suggested that it had promising utilization potentiality to monitor this kind of toxic gas. Our present work concentrated on further development of the sensor array for early diagnosis of lung cancer as well as monitoring of formaldehyde pollution for both domestic and industrial application.
Acknowledgements This work was supported by National Natural Science Foundation of China (81271930, 81171414 and 31171684), Key Technologies R&D Program of China (2012BAI19B03), Chongqing Natural Science Foundation (2010BB5226), Doctoral Fund of Ministry of Education (20090191110030) of China and Sharing fund of Chongqing University’s large equipment. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2014.01.054. References [1] D. Calestani, R. Mosca, M. Zanichelli, M. Villani, A. Zappettini, Aldehyde detection by ZnO tetrapod-based gas sensors, J. Mater. Chem. 1553 (2011) 15532–15536. [2] K. Kawamura, K. Kerman, M. Fujihara, N. Nagatani, T. Hashiba, E. Tamiya, Development of a novel hand-held formaldehyde gas sensor for the rapid detection of sick building syndrome, Sensors Actuators B: Chem. 2 (2005) 495–501. [3] M.E. Davis, A.P. Blicharz, J.E. Hart, F. Laden, E. Garshick, T.J. Smith, Occupational exposure to volatile organic compounds and aldehydes in the U.S. trucking industry, Environ. Sci. Technol. 20 (2007) 7152–7158. [4] J. You, T. Yan, H. Zhao, Z. Sun, L. Xia, Y. Suo, Y. Li, A sensitive fluorescence reagent for the determination of aldehydes from alcoholic beverage using highperformance liquid chromatography with fluorescence detection and mass spectrometric identification, Anal. Chim. Acta 1 (2009) 95–104. [5] B.W.M. Moeskops, M.M.L. Steeghs, K. van Swam, S.M. Cristescu, P.T.J. Scheepers, F.J.M. Harren, Real-time trace gas sensing of ethylene propanal and acetaldehyde from human skin in vivo, Physiol. Meas. 11 (2006) 1187. [6] H. Xu, S. Wang, A novel sorptive extraction method based on polydimethylsiloxane frit for determination of lung cancer biomarkers in human serum, Anal. Chim. Acta 0 (2012) 61–66. [7] D. Poli, M. Goldoni, M. Corradi, O. Acampa, P. Carbognani, E. Internullo, A. Casalini, A. Mutti, Determination of aldehydes in exhaled breath of patients with lung cancer by means of on-fiber-derivatisation SPME-GC/MS, J. Chromatogr. B 27SI (2010) 2643–2651. [8] J.A. Swenberg, W.D. Kerns, R.I. Mitchell, E.J. Gralla, K.L. Pavkov, Induction of squamous cell carcinomas of the rat nasal cavity by inhalation exposure to formaldehyde vapor, Cancer Res. 9 (1980) 3398–3402. [9] T.M. Monticello, J.A. Swenberg, E.A. Gross, J.R. Leininger, J.S. Kimbell, S. Seilkop, T.B. Starr, J.E. Gibson, K.T. Morgan, Correlation of regional and nonlinear formaldehyde-induced nasal cancer with proliferating populations of cells, Cancer Res. 5 (1996) 1012–1022. [10] L. Feng, C.J. Musto, K.S. Suslick, A simple and highly sensitive colorimetric detection method for gaseous formaldehyde, J. Am. Chem. Soc. 12 (2010) 4046–4047. [11] M. Hakim, Y.Y. Broza, O. Barash, N. Peled, M. Phillips, A. Amann, H. Haick, Volatile organic compounds of lung cancer and possible biochemical pathways, Chem. Rev. 11 (2012) 5949–5966. [12] R. Guadagni, N. Miraglia, A. Simonelli, A. Silvestre, M. Lamberti, D. Feola, A. Acampora, N. Sannolo, Solid-phase microextraction–gas chromatography– mass spectrometry method validation for the determination of endogenous substances: urinary hexanal and heptanal as lung tumor biomarkers, Anal. Chim. Acta 1 (2011) 29–36. [13] B. Mann, M.L. Grayeski, New chemiluminescent derivatizing agent for the analysis of aldehydes and ketones by high-performance liquid chromatography with peroxyoxalate chemiluminescence, J. Chromatogr. A 0 (1987) 149–158. [14] N. Dossi, S. Susmel, R. Toniolo, A. Pizzariello, G. Bontempelli, Simultaneous determination of derivatized light aldehydes by microchip electrophoresis with electrochemical detection, J. Chromatogr. A 1 (2) (2008) 169–174. [15] Y. Suzuki, N. Nakano, K. Suzuki, Portable sick house syndrome gas monitoring system based on novel colorimetric reagents for the highly selective and sensitive detection of formaldehyde, Environ. Sci. Technol. 24 (2003) 5695–5700. [16] W. Potter, U. Karst, Identification of chemical interferences in aldehyde and ketone determination using dual-wavelength detection, Anal. Chem. 19 (1996) 3354–3358. [17] M. Phillips, R.N. Cataneo, A.R.C. Cummin, A.J. Gagliardi, K. Gleeson, J. Greenberg, R.A. Maxfield, W.N. Rom, Detection of lung cancer with volatile markers in the breath*, CHEST J. 6 (2003) 2115–2123. [18] B. Buszewski, J. Rudnicka, T. Ligor, M. Walczak, T. Jezierski, A. Amann, Analytical and unconventional methods of cancer detection using odor, TrAC Trends Anal. Chem. 0 (2012) 1–12. [19] M.D.L. de Castro, M.A. Fernández-Peralbo, Analytical methods based on exhaled breath for early detection of lung cancer, TrAC Trends Anal. Chem. (2012) 13–20. [20] P.J. Mazzone, Exhaled breath volatile organic compound biomarkers in lung cancer, J. Breath Res. 2 (2012) 27106. [21] X. Chen, F. Xu, Y. Wang, Y. Pan, D. Lu, P. Wang, K. Ying, E. Chen, W. Zhang, A study of the volatile organic compounds exhaled by lung cancer cells in vitro for breath diagnosis, Cancer 4 (2007) 835–844.
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17 [22] C. Di Natale, A. Macagnano, E. Martinelli, R. Paolesse, G. D’Arcangelo, C. Roscioni, A. Finazzi-Agrò, A. D’Amico, Lung cancer identification by the analysis of breath by means of an array of non-selective gas sensors, Biosensors Bioelectron. 10 (2003) 1209–1218. [23] G. Peng, U. Tisch, O. Adams, M. Hakim, N. Shehada, Y.Y. Broza, S. Billan, R. AbdahBortnyak, A. Kuten, H. Haick, Diagnosing lung cancer in exhaled breath using gold nanoparticles, Nature Nanotechnol. 10 (2009) 669–673. [24] P.J. Mazzone, X. Wang, Y. Xu, T. Mekhail, M.C. Beukemann, J. Na, J.W. Kemling, K.S. Suslick, M. Sasidhar, Exhaled breath analysis with a colorimetric sensor array for the identification and characterization of lung cancer, J. Thoracic Oncol. 1 (2012). [25] Y. Zilberman, U. Tisch, G. Shuster, W. Pisula, X. Feng, K. Müllen, H. Haick, Carbon nanotube/hexa-peri-hexabenzocoronene bilayers for discrimination between nonpolar volatile organic compounds of cancer and humid atmospheres, Adv. Mater. 38 (2010) 4317–4320. [26] G. Konvalina, H. Haick, Effect of humidity on nanoparticle-based chemiresistors: a comparison between synthetic and real-world samples, ACS Appl. Mater. Interf. 1 (2012) 317–325. [27] G. Konvalina, H. Haick, Sensors for breath testing: from nanomaterials to comprehensive disease detection, Accounts Chem. Res. 0 (2013), in press. [28] Y.Y. Broza, H. Haick, Nanomaterial-based sensors for detection of disease by volatile organic compounds, Nanomedicine 5 (2013) 785–806. [29] A. Wehinger, A. Schmid, S. Mechtcheriakov, M. Ledochowski, C. Grabmer, G.A. Gastl, A. Amann, Lung cancer detection by proton transfer reaction massspectrometric analysis of human breath gas, Int. J. Mass Spectrom. 1 (2007) 49–59. [30] H. Haick, Y.Y. Broza, P. Mochalski, V. Ruzsanyi, A. Amann, Assessment, origin, and implementation of breath volatile cancer markers, Chem. Soc. Rev. (2014). [31] T.L. Wang, H.W. Tong, X.Y. Yan, L.Q. Sheng, J. Yang, S.M. Liu, Determination of volatile carbonyl compounds in cigarette smoke by LC-DAD, Chromatographia 11-12 (2005) 631–636. [32] C. Hou, J. Lei, D. Huo, K. Song, J. Li, X. Luo, M. Yang, H. Fa, Discrimination of lung cancer related volatile organic compounds with a colorimetric sensor array, Anal. Lett. 13 (2013) 2048–2059. [33] C. Hou, J. Li, D. Huo, X. Luo, J. Dong, M. Yang, X. Shi, A portable embedded toxic gas detection device based on a cross-responsive sensor array, Sensors Actuators B: Chem. 1 (2012) 244–250.
17
[34] C. Hou, J. Dong, G. Zhang, Y. Lei, M. Yang, Y. Zhang, Z. Liu, S. Zhang, D. Huo, Colorimetric artificial tongue for protein identification, Biosensors Bioelectron. 10 (2011) 3981–3986. [35] H. Lin, M. Jang, K.S. Suslick, Preoxidation for colorimetric sensor array detection of VOCs, J. Am. Chem. Soc. 42 (2011) 16786–16789. [36] L. Feng, C.J. Musto, J.W. Kemling, S.H. Lim, W. Zhong, K.S. Suslick, Colorimetric sensor array for determination and identification of toxic industrial chemicals, Anal. Chem. 22 (2010) 9433–9440. [37] L. Feng, C.J. Musto, J.W. Kemling, S.H. Lim, K.S. Suslick, A colorimetric sensor array for identification of toxic gases below permissible exposure limits, Chem. Commun. (2010) 2037–2039. [38] M.C. Janzen, J.B. Ponder, D.P. Bailey, C.K. Ingison, K.S. Suslick, Colorimetric sensor arrays for volatile organic compounds, Anal. Chem. 11 (2006) 3591–3600. [39] N.A. Rakow, K.S. Suslick, A colorimetric sensor array for odour visualization, Nature 6797 (2000) 710–713. [40] A. Schiller, B. Vilozny, R.A. Wessling, B. Singaram, Recognition of phospho sugars and nucleotides with an array of boronic acid appended bipyridinium salts, Anal. Chim. Acta 2 (2008) 203–211.
Biography
Professor Chanjun Hou completed his PhD in biomedical engineering in 2004 at Chongqing University. After a period of research at the University of Illinois at UrbanaChampaign as a visiting scholar, he returned to China and got tenure at the college of bioengineering in Chongqing University. In 2010, he became a full professor in biomedical engineering. His scientific interests are the design and construction of biochemical sensing systems based on porphyrins and other sensitive materials.
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Development of a colorimetric sensor Array for the discrimination of aldehydes Junjie Li a , Changjun Hou a,∗ , Danqun Huo a , Mei Yang a , Huan-bao Fa b , Ping Yang c a Key Laboratory of Biorheology Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400044, PR China b College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400044, China c National Engineering Research Center of Solid-State Brewing, Luzhou Laojiao Group Co. Ltd., Luzhou, Sichuan, 646000, PR China
a r t i c l e
i n f o
Article history: Received 21 October 2013 Received in revised form 15 January 2014 Accepted 19 January 2014 Available online 25 January 2014 Keywords: Aldehyde Sensor array Lung cancer Formaldehyde pollution
a b s t r a c t A brand-new colorimetric sensor array was developed based on cross-reactive mechanism to discriminate 9 kinds of aldehydes and 16 volatile organic lung cancer biomarker candidates in low concentration. Sixteen out of twenty-seven dyes were selected to mix with 2,4-dinitrophenylhydrazine (DNPH) in optimized concentration to serve as specific sensing elements, and polyethylene glycol-1000 was chosen from four kinds polyethylene glycols to act as stabilizer. Resultant sensor array shows improved response (about tenfold higher) to aldehydes than former study, and exhibited very good selectivity in low concentration ranging from 40 ppb to 10 ppm with the presence of interfering counterparts. Data analysis was performed by both hierarchical cluster analysis (HCA) and discriminant analysis (DA), which demonstrates the excellent discrimination ability of the sensor to structurally similar aldehydes related to lung cancer. Besides, formaldehyde-spiked air samples were analyzed with developed sensor array, suggesting promising utilization potentiality to monitor such toxic gas. Theoretical detection limit was down to 8.2 ppb with a liner range from 10 ppb to 150 ppm. The sensor array can be further developed for early diagnosis of lung cancer as well as monitoring of domestic and industrial formaldehyde pollution. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Volatile aldehydes are among the most common source of pollution in daily life due to their widespread presence in a plenty of chemical additives, industrial processes and incomplete combustions [1]. Acknowledged as an important components of toxic substances, aldehydes can be widely found in environmental atmospheres [2], industrial materials [3], alcoholic beverage [4], as well as human metabolites [5–7], which testify their significance in environmental, industrial and biological applications. In the area of environmental monitoring, formaldehyde poses a great threat to human health. It can lead to intensive irritant reactions of eyes and mucous membranes, and probable human carcinogen [8,9]. In view of its widespread use, toxicity and volatility, it is therefore highly problematic as an indoor pollutant. Related organizations thus banned it from use in certain applications (such as processing systems, preservatives for liquid-cooling, and antifouling products), and set the maximum allowed concentrations. The safe-exposure standard set by the World Health Organization is 80 ppb averaged over 30 min. While the permissible exposure limit
∗ Corresponding author. Tel.: +86 2365112673; fax: +86 2365102507. E-mail address: [email protected] (C. Hou). 0925-4005/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.snb.2014.01.054
(PEL) set by Occupational Safety and Health Administration (OSHA) is 750 ppb and the immediately dangerous to life or health (IDLH) limit is 20 ppm [10]. Besides, there is an increasing interest concerning the role of volatile aldehydes as biomarker in biomedical field. According to previously studies, there is a significant difference in the composition of volatile organic compounds in human breath and other metabolites. Discrimination of those volatile biomarkers would undoubtedly provide a non-invasive alternative for the diagnosis of lung cancer. Despite controversial debate in accurate VOCs biomarkers, partially due to exogenous interferences and individual difference, it is highly convinced that volatile aldehydes are very likely to be among lung cancer biomarkers [11]. For example, solidphase micro-extraction–gas chromatography–mass spectrometry method had been utilized for the determination of endogenous hexanal and heptanal in urine samples [12]. The results demonstrated that hexanal urinary concentrations in cancer patients were slightly higher than those found in control group ones. Hexanal and heptanal have also been measure in serum samples, which suggested that their concentration levels in the samples of lung cancer patients were sharply higher than those in healthy people [6]. In addition, it was also found that the levels of C3–C9 aldehydes in exhaled breath were increased in the non-small cell lung cancer (NSCLC) patients without slight effect of smoking habits and age
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
[7]. So that straight aldehydes might be better biomarkers associated with NSCLC than other VOC patterns. Since the concentration of volatile aldehydes as toxic pollutant and cancer biomarkers are extremely low, normally several ppb to a hundred ppm, detection of aldehyde at very low concentration level is of great importance for both environmental surveillance and biomedical diagnosis. A plenty of techniques and methods have been devised for the detection and measurement of gaseous aldehyde. Traditionally, aldehydes can be analyzed by standard techniques such as liquid chromatograph [4,13], gas chromatograph [3,12], and electrochemical techniques [1,14]. Based on large instrumentation, those methods do offer very good choices and fine results to analyze this group of compounds. However, despite high sensitivity and good reproducibility, above-mention method all involves unavoidable drawbacks like expensive and cumbersome operations and also time-consuming procedures, which is adverse for real-time and field use. There are also inexpensive techniques including colorimetric or fluorometric methods with the application of pigments, and so forth [2,15,16]. But lack of sensitivity greatly hinders those methods from wide application. Moreover, they also need a relatively long detection time and boring steps to get final results. Thus, there still remains a pressing need for the development of a sensitive, rapid, inexpensive and facile method for aldehyde detection. In the past decades, there is an ever-growing interest in the study on the diagnosis of lung cancer by analysis of VOCs in human breath for that it is non-invasive, easy to handle and time-saving [17]. MS techniques (whether or not coupled to chromatographic equipment) provide a reliable method for the determination of mixtures of VOCs and semi-VOCs in different human metabolites [18]. But high-cost and expert interpretation greatly hinders them from wide application and also prolongs the research period. Therefore, many researchers prefer to techniques involving smaller, less expensive equipment, which is simple to handle and able to be miniaturized, particularly sensors or sensor arrays [19,20]. Outstanding studies include polymer-coated surface acoustic wave sensors [21], coated quartz crystal microbalance sensors [22], gold nanoparticles sensor array [23], and also colorimetric sensor array [24]. Selective of semi-selective electrochemical sensors have been proved to be a good choice for breath analysis [25], and after signal compensation it can largely reduce the influence resulted from humidity fluctuation [26]. Yet there are still challenges to develop such sensor systems to deal with various confounding factors in real-world analysis [27,28]. Since VOCs can be influenced by various internal and external factors such as individual physical conditions (age, respiration rate, nutritional status and so on), smoking circadian rhythm, and composition of environmental air, definition of normal composition and ranges of single VOCs in human breath gas remains a tricky problem [29]. Therefore, sensor systems mentioned before all involve the same limitations, that is, they are focused on the overall composition of the VOCs in the metabolites thus lose the specificity of each VOCs marker, which would definitely restrict the selectivity of the sensor and result in false positive/negative discrimination. Accordingly, by comparing the presence/absence of each experimentally validated VOCs in the LC states relative to the control states, most of possible VOC lung cancer biomarker candidates can be divided into seven compound families: hydrocarbons, alcohols, aldehydes, ketones, esters, nitriles and aromatic compounds [11,30]. A feasible strategy to solve above-mentioned problem is to develop specialized sensors for each biomarker family members and then integrate separated sensors to a sensor system. In view of this thought, we employed colorimetric sensor array as detection technique, and firstly chose aldehydes and then other VOC family as biomarkers to design the sensor, which might lead to breakthrough in this research area.
11
Accordingly, since colorimetric sensor array systems are broadspectrum sensors, overfitting, which is mainly resulted from sensor dots that are widely responsive, of the training data set poses a big problem [27]. In view of this problem, a feasible way is to improve the selectivity of each sensor dot in the array to selected volatile biomarker. Given that most aldehyde can be easily derivatized with many colorimetric reagents, it is possible choose aldehydes as target biomarker to carry out visual colorimetry, and allows easy measurements [15]. A good case in study is an early report of a highly sensitive colorimetric method for fast formaldehyde detection [10]. Using amine-terminated PEGs as specific agent to create a reactive matrix, a simple pH indicator array was applied to detect the change in basicity upon the reaction of a non-volatile primary amine with formaldehyde. It eventually gave a detection limit down to 50 ppb within 10 min. Nevertheless, apart from its great success in formaldehyde detection, the sensor array cannot response to aldehydes with longer carbon chain or molecular weight. Therefore, it can be hardly utilized for longer-carbon-chain aldehydes detection in biomedical application. Herein, we fabricated a more sensitive sensor array using DNPH as aldehyde-specific sensing elements and polyethylene glycol-1000 as stabilizer. The sensor showed very good selectivity and sensitivity towards aldehydes among other VOCs lung cancer biomarker candidates. It can also discriminate nine kinds of C1–C7 aldehydes in low concentration ranging from ppm level to ppb level, using Hierarchical Cluster Analysis (HCA) and Linear Discriminant Analysis (LDA). Besides, formaldehyde-spiked air samples were also analyzed with asprepared sensor array, exhibiting very low detection limit and good linear range. The sensor array would provide an efficient tool for early diagnosis of lung cancer, and monitoring of domestic and industrial formaldehyde pollution. 2. Material and methods 2.1. Chemicals and preparation of stock solutions Four porphyrins were obtained from Frontier Scientific (Logan, UT, USA), and the other indicator dyes (listed in table S1) were supplied by Sigma–Aldrich (St. Louis, MO, USA). Porous hydrophobic membrane used for dye staining was bought from Milli-pore Co. Ltd. (Bedford, MA, USA). 2, 4-dinitrophenylhydrazine (DNPH), polyethylene glycol-400 (PEG-400), polyethylene glycol-800 (PEG800), polyethylene glycol-1000 (PEG-1000), polyethylene glycol4000 (PEG-4000), aldehyde solutions, and other gas solutions were purchased from Jinchun Industry Co., Ltd. (Shanghai, China). All the other chemicals are of analytical pure and used without further purification. Ultra-pure water was generated by a Millipore Direct-Q Water system (Molsheim, France). Accordingly, the reaction between aldehyde and DNPH is catalyzed by acid [31]. At high acid concentrations the carbonyl group of aldehyde is activated by the acid but at lower acid concentrations, or in basic solutions, the carbonyl group becomes less reactive, which hinder aldehyde from dehydration to the hydrazones. Although HClO4 has been frequently used as catalyst to derivatize aldehydes, we use H2 SO4 instead for the sake of stability of resultant solution. Typically, 0.4 g DNPH was added to a mixed solution of 3 mL water and 10 mL ethanol, and then 2 mL concentrated sulfuric acid was added dropwise. The solution was stirred for 10 min and then went through filter paper to obtain a yellow DNPH store solution. For prophyrin solution, 10 mg dye was added into 3 mL DMF solution and stored in dark place before use. Water solutions were used instead for alizarin and nitrazine yellow while all the other dye solutions were prepared using ethanol solution with the same concentration. Gas samples were prepared by a self-made gas distribution device. Briefly, saturated vapor of each aldehyde was produced by dilution with dry and wet nitrogen to
12
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
Fig. 1. Actual difference maps of sensor arrays: (A) selection-oriented sensor array with five parallel dots; (B) detection-oriented sensor array and arrangement of the sensing dots.
achieve the desired concentration, which was testified before-hand by GC-MS, in a gas stream at 50% relative humidity (RH). 2.2. Selections of dye and dye/DNPH ratio In order to realize a high-efficient and less-expensive selection of dyes and dye/DNPH ratio, we first fabricated a selection-oriented sensor array. The sensor array was prepared as following: 50 L dye stock solution and 50 L PEG-400 was mixed with DNPH solution in a 3 mL plastic vial with different dye/DNPH ratio (1:5, 1:4, 1:3, 1:2, and 1:1). The mixture was kept in ultrasonic agitation for ten seconds and then a quartz capillary was used to deliver approximately 0.1 L solution onto the surface of a porous hydrophobic membrane to fabricate a 5 × 5 sensor array (five parallel dots for each ratio, see Fig. 1A). Once printed, the arrays were placed in a 500 mL beaker saturated in nitrogen atmosphere for 30 min and subsequently dried in a 60 ◦ C oven for 24 h after which the oven temperature was reduced to 35 ◦ C and the arrays left for another 24 h. Detailed working principle, measurement of the optical signal, and also data processing have been introduced in our early studies [32,33]. Briefly, we use a complementary metal-oxide semiconductor (CMOS) camera to obtain the color information of the sensor array. When reacting with gas molecule, the color of certain spots would change. By comparing the array images obtained before and after interaction, discrimination of different gases in different concentrations can be easily achieved. Specially, we constructed a gas processing system for the analysis of the VOCs in this study. It consists of a circular gas tank to control gas concentration and humidity, a gas chamber for chemical reaction between the gas and the sensor array, a waste cylinder for the collection of exhaust gases, and a pump to provide driving force for the gas flow. Upon detection, the gas system was first cleaned using N2 before-hand in order to preclude the interference of the air. Selected VOC in given concentration was then prepared by the gas processing systems and reacted with the colorimetric sensor array. After exposure of the array to the as-prepared VOCs, the sensor array images were automatically collected by CMOS camera at given time point. Difference map of each VOC in given concentration subsequently was generated by the image-processing system for further data analysis. Hexanal samples were chosen to testify the relative response and stability of each dye in different dye/DNPH ratio. Specifically, nitrogen gas was applied as control gas and hexanal gas was mixed with ultra-pure nitrogen to a concentration of 5 ppm in a 3 L Tedlar bag. The bag was then connected with the sensor system upon analysis. The gas flow rate were set as 5 mL/s at room temperature (25 ◦ C) with a relative humidity of 50%. The pictures of the sensor array can be taken at set time point until the reaction ends. By comparing the initial image of the array with its final image
after exposure to targeted gas sample, color change profiles were automatically obtained. Detailed working principle of the sensory system and data process can be found in our former studies [33,34]. In order to eliminate the noise and extract useful information, a threshold T is set, which is adjustable in the image analysis software written by our lab. When R + G + B < T, the color is set as R = G = B = 0. In this section, the response threshold was set as 10, 16 kinds of dyes were selected prior to further test. 2.3. Aldehyde analysis method with the sensor array A similar strategy was applied to select the stabilizer PEGs with a sensor array (see Fig. 1B). The sensor arrays of four different PEGs were subjected to above-mentioned preparing procedure and testing method to select the best stabilizer. Optimized stabilizer was then utilized in the fabrication of the final sensor array (detectionoriented sensor arrays). It should be noted that the arrays were stored in a nitrogen-flushed humidity-controllable container with a RH of 50% and room temperature (25 ◦ C) before use. The response threshold was set as 12 in this section and following experiment so as to suppress possible interference. Aldehyde samples and 16 other VOC lung biomarker candidates in concentration of 5 ppm and 20 ppb were prepared using nitrogen gas and testified by the sensor array. Resultant experimental data was subsequently analyzed with both HCA and LDA. As for the detection of formaldehyde-spiked air, gas samples were prepared by air instead of nitrogen gas. 3. Results and discussion 3.1. Detection principle and fabrication of the sensor array The selection of dye and dye/DNPH ratio determines the sensitivity and specificity of the sensor array. As key sensory components, responsive dyes are crucial for the performance of a colorimetric sensor array. In previous studies, porphyrins, pH indicators, metal salts and metal oxides have been used to fabricate colorimetric sensor array for a variety of gaseous analytes [35–38]. The selectivity of the sensor array depends on two aspects: specialized response of dyes to different analytes and cross-responsive interaction between analytes and the sensor array, in other words, unique combination of the responses of all the dyes to target analyte. Generally, the more versatile a sensor system is, the less selective the performance for that versatility usually means more interference for a selected target. Therefore, it is in urgent need to improve the specificity of each sensor dot in the sensor array so as to improve the overall sensor performance. In view of this problem, we proposed the combination of dye and DNPH as sensing element based on the specific reaction of aldehyde with DNPH. As shown in
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
13
Fig. 2. Schematic illustration of the procedure to develop the colorimetric sensor array.
Fig. 2, we hypothesize the detection principle as following: Aldehyde reacts with DNPH resulting in the production of H2 O; H2 SO4 serves as catalyst in the reaction and then protonized by the product H2 O, which lead to the pH variation in the sensor dot. Selected dyes are sensitive to the pH change and give a color change in turn; since different kinds and amount of aldehyde have different affinity to DNPH and produce different volume of H2 O, the color change given by the dyes will be unique from each other. Measurement of color change of the sensor array thus result in the discrimination of different aldehydes and other possible interfering counterparts. Herein, we applied a high-efficient and less-expensive way to realize the selecting procedure relying on selection-oriented sensor arrays. In order to achieve the highest response as possible, we first developed a sensor array to optimize the selection of dyes and dye/DNPH ratios. As illustrated in Fig. 1A, the selection-oriented sensor array has five parallel dots for each dye/DNPH combination so that we can measure the respective response of different dye/DNPH ratios at the same time with five controls. It will largely reduce the cost to develop the sensor array and get a relatively more reliable result. Likely, we use squared Euclidean Distance change (R2 , G2 and B2 of each sensor dot) to measure the response [39]. Using hexanal as target and nitrogen gas as control, relative response and stability of twenty-seven frequently-reported dyes with five dye/DNPH ratios have been measured. Dyes that are more responsive and with better stability to hexanal upon control were subsequently selected as final sensor dots, the first 16 out of 135 combination and detailed information are summarized in Table S1. The micro morphology of the substrate PVDF film before and after dye-printing have also been observed by scanning electron microscope (Figure S4). It was shown that the PVDF film shows very uniform morphology and porosity before dye-printing. And when the dye solution was printed into the film, no significant change of the morphology or porosity was observed, indicating that the DNPH, dye and also PEGs molecules homogeneously diffused into the porous PVDF matrices without agglomeration or impairing the film. Besides the response and specificity to target analytes, stability is another important factor to determine the performance of a sensor system. PEG-400 had been used in the fabrication of colorimetric sensor array to improve the stability previously [10]. Accordingly, the benefits using PEG-400 can be concluded as two aspects, that is, protecting dye from external environment and insulating the interference of hydrophilic compounds like water vapor. In the present study, we investigated the performance of
four different PEGs aiming to choose a better stabilizer for the sensor array, and the results were summarized in Fig. S1. The squared Euclidean Distance change represents the response of each PEGs to hexanal gas while the relative deviation indicates the stability. As is shown in the figure, PEG-1000 exhibited the best response and relative lower deviation, which is indicative of good stabilizing ability. Therefore, PEG-1000 was finally chosen as stabilizer for follow-up fabrication of detection-oriented sensor array. 3.2. Discrimination of structurally similar aldehydes over other VOC lung cancer biomarker candidates When the detection-oriented sensor array was developed, discrimination of different aldehydes and other possible VOC markers was subsequently studied. Apart from the aldehyde family, most VOCs either have no carbonyl group or cannot react with DNPH to produce H2 O. The former VOCs would not interact with the sensor dots while the latter could not lead to enough pH change thus will not result in enough response that can be detected by the sensor array. Firstly, time-dependent response of the sensor array to different aldehydes was studied to establish equilibration. The total squared Euclidean Distance for each aldehyde was analyzed to track the response of as-prepared array. As can be seen in Fig. 3, squared Euclidean Distance was increasing for each aldehyde until it reached saturation. Although the reaction between the sensor array and most aldehydes saturated within 2 min, a response time of 4 min was chosen as optimized reaction time for subsequent experiments so as to ensure obtained squared Euclidean Distance in the final equilibration. Then we assessed the selectivity between aldehydes and 16 other VOC biomarker candidates, the difference maps are shown in Fig. 4 and Fig. S3. The results suggested that discriminating ability of developed sensor array was so good that one can easily tell aldehydes from other counterparts. Either in the concentration of 10 ppm or 40 ppb, there are no mis-recognition observed even by naked-eyes. Meanwhile, identification of structurally similar aldehydes remains a big challenge because they are closely alike in structures and exhibit very similar properties. Based on cross-reactive mechanism, the recognition of a target is resulted from the sum and organic combination of specific response of each sensing element, thus providing an efficient alternative to discriminate analytes with high similarity. Pattern recognition has been proved to be a powerful tool for cross-reactive analysis due to its ability to discriminate complicate mixtures. Herein, we applied the same methodology to
14
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
Fig. 3. Time-dependent response of different aldehydes.
realize discrimination between structurally similar aldehydes. To demonstrate the applicability of as-prepared colorimetric sensor array for aldehyde discrimination, nine kinds of aldehydes in the concentration of 10 ppm and 40 ppb were tested. Upon exposure to the aldehyde sample, the array underwent specific reactions, resulting in well-defined color changes in final difference maps. Color change profiles with clear difference were shown in Fig. 4. The color difference maps showed the distinctive patterns for all nine aldehydes both in the concentration of 10 ppm and 40 ppb. When the aldehyde samples were in higher concentrations, the sensor array shows obviously more responsive dots. It demonstrated that as-prepared have the potential to be developed for further analysis of VOCs in human breath for lung cancer diagnosis. Notably, it also can be seen from Fig. 3 that formaldehyde exhibited the largest response than other eight aldehydes, which possibly due to its higher affinity to DNPH. In view of the change of squared Euclidean Distance upon exposure to formaldehyde in the same concentration range, sensor array developed here exhibited
response ten times higher than former study [10]. Basically, for C1 to C8 linear-chain aldehydes, resultant response decreases with the carbon chain length increases. 3.3. HCA and LDA analysis of resultant data In order to further examine the sensory performance of asprepared sensor array, statistical methodology was utilized to analyze resultant experimental data with the application of HCA and LDA, which are widely used in related studies. For both analysis, the variables were define as relative change in each color vector (R2 or G2 or B2 divided by total squared Euclidean Distance), and totally 48 (4 × 4 × 3) parameters were used to conduct statistical analysis. In each case, the data was detected repeatedly for 5 times with 5 parallel samples. The HCA dendrograms for concentrations in 10 ppm and 40 ppb was separately shown in Fig. S3. HCA is a multivariate statistical analysis method which is comprised of agglomerative and divi-
Fig. 4. Difference maps of different aldehydes after interaction with the sensor array in the concentration of 40 ppb (A) and 10 ppm (B): a, formaldehyde; b, acetaldehyde; c, propanal; d, n-butanal; e, pentanal; f, hexanal; g, heptanal; h, acetal; i, isobutyraldehyde.
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
15
Fig. 5. Linear discriminant analysis results of nine kinds of aldehydes with five parallel samples in the concentration of 40 ppb (A) and 10 ppm (B).
sive methods that find clusters of observations within a data set. It could be seen from the figures that all 9 aldehyde gas samples were accurately identified against each other in two concentration levels. And nearly no misclassification was observed upon examination of 45 gas samples. Remarkably, for both concentration levels, five formaldehyde samples were classified into a single cluster with nearly no error. It indicated that the sensor array have very good stability for formaldehyde and can be further developed as gas sensor for such toxic gas pollution. It also showed that hexanal samples in low concentrations seemed to be more easily clustered than in high concentration. Since concentration discrepancy of hexanal between normal people and lung cancer patients was in low levels, it would benefit the diagnosis of lung cancer. As a multivariate statistical analysis method, HCA is comprised of agglomerative and divisive methods, and can be utilized to screen cluster of observations within certain data set. The results obviously suggested that the sensor array have acceptable selectivity to classify different aldehydes in low concentration. Besides, the HCA analysis is also indicative of further application of the sensor system when a database of difference maps of different aldehydes was established,
one can easily recognize an unknown aldehyde and even its concentration level just by checking the library entries. Discriminant analysis results provided another evidence to prove the good selectivity of as-prepared sensor array. Dissimilar to HCA, LDA is a supervised method for separating classes of objects and for assigning new components to appropriate classes. The discriminants are linear combinations of the measured sensor responses. Discriminant functions are calculated to maximize the Mahalanobis distance between classes relative to the variation within classes [40]. It is able to find a linear combination of features which characterizes or separates two or more classes of objects or events. Based on this mechanism, 9 kinds of aldehydes in concentration of 10 ppm and 40 ppb were classified into 9 groups (Fig. 5). In both concentration levels, all five parallel samples gather together with no misclassification. Different aldehyde groups were well separated with no clear overlap. Notably, the acetal group in high concentration was significantly farther away for other eight aldehydes than in low concentration. As acetal has completely different structure with the other ones, it is reasonable to conclude that higher concentration is beneficial for the discrimination
Fig. 6. Response of the sensor array upon exposure to formaldehyde in different concentration. Inserted figure denotes the linear fitting of concentration-dependent response in concentrations ranging from 40 ppb to 10 ppm.
16
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17
of structurally dissimilar inferences. The result of LDA analysis demonstrated an extraordinary high level of classification of structurally similar aldehydes based on a simple colorimetric technique.
3.4. Application of the sensor array to monitor formaldehyde The feasibility to apply as-prepared sensor array for formaldehyde pollution monitoring was subsequently investigated. Specifically, we prepared formaldehyde-spiked air samples in concentration ranging from ppt level to ppm level, and then analyzed the samples with the sensor array. As shown in Fig. 6, total squared Euclidean distance of changes in RGB showed positive correlation with the formaldehyde concentration of gas samples, which base a semi-quantitative analysis simply on the total response of the array. Besides, linear regression analysis indicated that logarithmic formaldehyde concentration has good linear relationship with the total squared Euclidean distance in the range from 40 ppb to 10 ppm (see inserted figure in Fig. 6). Theoretical detection limit was down to 8.2 ppb and the response concentration was even as lower as ppt level. Compared to previous study, the sensor array showed less response time and it would not be difficult to improve the response limit just by increasing the exposure time of gas simple to the sensor array. Moreover, due to the specific interaction between DNPH and aldehyde, the sensor array could be hardly influenced by most gaseous interference. We analyzed a plenty of possible interfering gases yet only SO2 and NH3 in high concentrations (>250 ppm) led to significant interference in the detection. Yet the interference can be excluded simply by using basic and/or acidic absorbent before detection, or just adding basic and/or acidic controls in the sensor array. Therefore, it is promising to develop such sensor array to monitor formaldehyde pollution of low concentration in both domestic and industrial application.
4. Conclusion As a proof of concept, colorimetric sensor array system offers a feasible method to develop sensors on the basis of cross-reactive mechanism. Despite their wide application for a variety of analytes, it is crucial to select unique sensory components for specific targets so as to improve the sensor performance. In the present study, we presented the development of a brand-new colorimetric sensor array to discriminate nine kinds of structurally similar aldehydes in low concentration. When optimizing the fabrication procedure of the sensor array, we applied a selection-oriented sensor to realize a high-efficient and less-expensive selection of dyes, dye/DNPH ratio and stabilizers. Sixteen out of twenty-seven dyes in optimized concentrations were first selected to mix with DNPH to serve as aldehyde-specific sensing elements, and then polyethylene glycol-1000 was chosen from four kinds of polyethylene glycols to improve the stability. Compared with former study, our sensor array shows about tenfold higher response to aldehydes, and exhibited very good sensor performance in low concentration ranging from ppm level to ppb level, with nearly no interference of sixteen interfering VOC lung cancer biomarker candidates. Hierarchical cluster analysis and principal component analysis were further applied to analyze resultant experimental data, which demonstrates the excellent discrimination ability to structurally similar aldehydes. Besides, formaldehyde-spiked air samples were analyzed using the sensor array. Results suggested that it had promising utilization potentiality to monitor this kind of toxic gas. Our present work concentrated on further development of the sensor array for early diagnosis of lung cancer as well as monitoring of formaldehyde pollution for both domestic and industrial application.
Acknowledgements This work was supported by National Natural Science Foundation of China (81271930, 81171414 and 31171684), Key Technologies R&D Program of China (2012BAI19B03), Chongqing Natural Science Foundation (2010BB5226), Doctoral Fund of Ministry of Education (20090191110030) of China and Sharing fund of Chongqing University’s large equipment. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2014.01.054. References [1] D. Calestani, R. Mosca, M. Zanichelli, M. Villani, A. Zappettini, Aldehyde detection by ZnO tetrapod-based gas sensors, J. Mater. Chem. 1553 (2011) 15532–15536. [2] K. Kawamura, K. Kerman, M. Fujihara, N. Nagatani, T. Hashiba, E. Tamiya, Development of a novel hand-held formaldehyde gas sensor for the rapid detection of sick building syndrome, Sensors Actuators B: Chem. 2 (2005) 495–501. [3] M.E. Davis, A.P. Blicharz, J.E. Hart, F. Laden, E. Garshick, T.J. Smith, Occupational exposure to volatile organic compounds and aldehydes in the U.S. trucking industry, Environ. Sci. Technol. 20 (2007) 7152–7158. [4] J. You, T. Yan, H. Zhao, Z. Sun, L. Xia, Y. Suo, Y. Li, A sensitive fluorescence reagent for the determination of aldehydes from alcoholic beverage using highperformance liquid chromatography with fluorescence detection and mass spectrometric identification, Anal. Chim. Acta 1 (2009) 95–104. [5] B.W.M. Moeskops, M.M.L. Steeghs, K. van Swam, S.M. Cristescu, P.T.J. Scheepers, F.J.M. Harren, Real-time trace gas sensing of ethylene propanal and acetaldehyde from human skin in vivo, Physiol. Meas. 11 (2006) 1187. [6] H. Xu, S. Wang, A novel sorptive extraction method based on polydimethylsiloxane frit for determination of lung cancer biomarkers in human serum, Anal. Chim. Acta 0 (2012) 61–66. [7] D. Poli, M. Goldoni, M. Corradi, O. Acampa, P. Carbognani, E. Internullo, A. Casalini, A. Mutti, Determination of aldehydes in exhaled breath of patients with lung cancer by means of on-fiber-derivatisation SPME-GC/MS, J. Chromatogr. B 27SI (2010) 2643–2651. [8] J.A. Swenberg, W.D. Kerns, R.I. Mitchell, E.J. Gralla, K.L. Pavkov, Induction of squamous cell carcinomas of the rat nasal cavity by inhalation exposure to formaldehyde vapor, Cancer Res. 9 (1980) 3398–3402. [9] T.M. Monticello, J.A. Swenberg, E.A. Gross, J.R. Leininger, J.S. Kimbell, S. Seilkop, T.B. Starr, J.E. Gibson, K.T. Morgan, Correlation of regional and nonlinear formaldehyde-induced nasal cancer with proliferating populations of cells, Cancer Res. 5 (1996) 1012–1022. [10] L. Feng, C.J. Musto, K.S. Suslick, A simple and highly sensitive colorimetric detection method for gaseous formaldehyde, J. Am. Chem. Soc. 12 (2010) 4046–4047. [11] M. Hakim, Y.Y. Broza, O. Barash, N. Peled, M. Phillips, A. Amann, H. Haick, Volatile organic compounds of lung cancer and possible biochemical pathways, Chem. Rev. 11 (2012) 5949–5966. [12] R. Guadagni, N. Miraglia, A. Simonelli, A. Silvestre, M. Lamberti, D. Feola, A. Acampora, N. Sannolo, Solid-phase microextraction–gas chromatography– mass spectrometry method validation for the determination of endogenous substances: urinary hexanal and heptanal as lung tumor biomarkers, Anal. Chim. Acta 1 (2011) 29–36. [13] B. Mann, M.L. Grayeski, New chemiluminescent derivatizing agent for the analysis of aldehydes and ketones by high-performance liquid chromatography with peroxyoxalate chemiluminescence, J. Chromatogr. A 0 (1987) 149–158. [14] N. Dossi, S. Susmel, R. Toniolo, A. Pizzariello, G. Bontempelli, Simultaneous determination of derivatized light aldehydes by microchip electrophoresis with electrochemical detection, J. Chromatogr. A 1 (2) (2008) 169–174. [15] Y. Suzuki, N. Nakano, K. Suzuki, Portable sick house syndrome gas monitoring system based on novel colorimetric reagents for the highly selective and sensitive detection of formaldehyde, Environ. Sci. Technol. 24 (2003) 5695–5700. [16] W. Potter, U. Karst, Identification of chemical interferences in aldehyde and ketone determination using dual-wavelength detection, Anal. Chem. 19 (1996) 3354–3358. [17] M. Phillips, R.N. Cataneo, A.R.C. Cummin, A.J. Gagliardi, K. Gleeson, J. Greenberg, R.A. Maxfield, W.N. Rom, Detection of lung cancer with volatile markers in the breath*, CHEST J. 6 (2003) 2115–2123. [18] B. Buszewski, J. Rudnicka, T. Ligor, M. Walczak, T. Jezierski, A. Amann, Analytical and unconventional methods of cancer detection using odor, TrAC Trends Anal. Chem. 0 (2012) 1–12. [19] M.D.L. de Castro, M.A. Fernández-Peralbo, Analytical methods based on exhaled breath for early detection of lung cancer, TrAC Trends Anal. Chem. (2012) 13–20. [20] P.J. Mazzone, Exhaled breath volatile organic compound biomarkers in lung cancer, J. Breath Res. 2 (2012) 27106. [21] X. Chen, F. Xu, Y. Wang, Y. Pan, D. Lu, P. Wang, K. Ying, E. Chen, W. Zhang, A study of the volatile organic compounds exhaled by lung cancer cells in vitro for breath diagnosis, Cancer 4 (2007) 835–844.
J. Li et al. / Sensors and Actuators B 196 (2014) 10–17 [22] C. Di Natale, A. Macagnano, E. Martinelli, R. Paolesse, G. D’Arcangelo, C. Roscioni, A. Finazzi-Agrò, A. D’Amico, Lung cancer identification by the analysis of breath by means of an array of non-selective gas sensors, Biosensors Bioelectron. 10 (2003) 1209–1218. [23] G. Peng, U. Tisch, O. Adams, M. Hakim, N. Shehada, Y.Y. Broza, S. Billan, R. AbdahBortnyak, A. Kuten, H. Haick, Diagnosing lung cancer in exhaled breath using gold nanoparticles, Nature Nanotechnol. 10 (2009) 669–673. [24] P.J. Mazzone, X. Wang, Y. Xu, T. Mekhail, M.C. Beukemann, J. Na, J.W. Kemling, K.S. Suslick, M. Sasidhar, Exhaled breath analysis with a colorimetric sensor array for the identification and characterization of lung cancer, J. Thoracic Oncol. 1 (2012). [25] Y. Zilberman, U. Tisch, G. Shuster, W. Pisula, X. Feng, K. Müllen, H. Haick, Carbon nanotube/hexa-peri-hexabenzocoronene bilayers for discrimination between nonpolar volatile organic compounds of cancer and humid atmospheres, Adv. Mater. 38 (2010) 4317–4320. [26] G. Konvalina, H. Haick, Effect of humidity on nanoparticle-based chemiresistors: a comparison between synthetic and real-world samples, ACS Appl. Mater. Interf. 1 (2012) 317–325. [27] G. Konvalina, H. Haick, Sensors for breath testing: from nanomaterials to comprehensive disease detection, Accounts Chem. Res. 0 (2013), in press. [28] Y.Y. Broza, H. Haick, Nanomaterial-based sensors for detection of disease by volatile organic compounds, Nanomedicine 5 (2013) 785–806. [29] A. Wehinger, A. Schmid, S. Mechtcheriakov, M. Ledochowski, C. Grabmer, G.A. Gastl, A. Amann, Lung cancer detection by proton transfer reaction massspectrometric analysis of human breath gas, Int. J. Mass Spectrom. 1 (2007) 49–59. [30] H. Haick, Y.Y. Broza, P. Mochalski, V. Ruzsanyi, A. Amann, Assessment, origin, and implementation of breath volatile cancer markers, Chem. Soc. Rev. (2014). [31] T.L. Wang, H.W. Tong, X.Y. Yan, L.Q. Sheng, J. Yang, S.M. Liu, Determination of volatile carbonyl compounds in cigarette smoke by LC-DAD, Chromatographia 11-12 (2005) 631–636. [32] C. Hou, J. Lei, D. Huo, K. Song, J. Li, X. Luo, M. Yang, H. Fa, Discrimination of lung cancer related volatile organic compounds with a colorimetric sensor array, Anal. Lett. 13 (2013) 2048–2059. [33] C. Hou, J. Li, D. Huo, X. Luo, J. Dong, M. Yang, X. Shi, A portable embedded toxic gas detection device based on a cross-responsive sensor array, Sensors Actuators B: Chem. 1 (2012) 244–250.
17
[34] C. Hou, J. Dong, G. Zhang, Y. Lei, M. Yang, Y. Zhang, Z. Liu, S. Zhang, D. Huo, Colorimetric artificial tongue for protein identification, Biosensors Bioelectron. 10 (2011) 3981–3986. [35] H. Lin, M. Jang, K.S. Suslick, Preoxidation for colorimetric sensor array detection of VOCs, J. Am. Chem. Soc. 42 (2011) 16786–16789. [36] L. Feng, C.J. Musto, J.W. Kemling, S.H. Lim, W. Zhong, K.S. Suslick, Colorimetric sensor array for determination and identification of toxic industrial chemicals, Anal. Chem. 22 (2010) 9433–9440. [37] L. Feng, C.J. Musto, J.W. Kemling, S.H. Lim, K.S. Suslick, A colorimetric sensor array for identification of toxic gases below permissible exposure limits, Chem. Commun. (2010) 2037–2039. [38] M.C. Janzen, J.B. Ponder, D.P. Bailey, C.K. Ingison, K.S. Suslick, Colorimetric sensor arrays for volatile organic compounds, Anal. Chem. 11 (2006) 3591–3600. [39] N.A. Rakow, K.S. Suslick, A colorimetric sensor array for odour visualization, Nature 6797 (2000) 710–713. [40] A. Schiller, B. Vilozny, R.A. Wessling, B. Singaram, Recognition of phospho sugars and nucleotides with an array of boronic acid appended bipyridinium salts, Anal. Chim. Acta 2 (2008) 203–211.
Biography
Professor Chanjun Hou completed his PhD in biomedical engineering in 2004 at Chongqing University. After a period of research at the University of Illinois at UrbanaChampaign as a visiting scholar, he returned to China and got tenure at the college of bioengineering in Chongqing University. In 2010, he became a full professor in biomedical engineering. His scientific interests are the design and construction of biochemical sensing systems based on porphyrins and other sensitive materials.