Article Cite This: ACS Sens. XXXX, XXX, XXX-XXX
pubs.acs.org/acssensors
Organic Gas Sensor with an Improved Lifetime for Detecting Breath Ammonia in Hemodialysis Patients Ming-Yen Chuang,† Chang-Chiang Chen,*,‡,§ Hsiao-Wen Zan,*,† Hsin-Fei Meng,*,∥ and Chia-Jung Lu⊥ †
Department of Photonics, College of Electrical and Computer Engineering, §Department of Biological Science and Technology, and Institute of Physics, National Chiao Tung University, 1001, Ta Hsueh Rd., 300 Hsinchu, Taiwan ‡ Department of Internal Medicine, Division of Nephrology, National Taiwan University Hospital Hsin-Chu Branch, 25, Ln. 442, Sec. 1, Jingguo Rd., 300 Hsinchu, Taiwan ⊥ Department of Chemistry, National Taiwan Normal University, 162, Heping East Rd., Section 1, 106 Taipei, Taiwan ∥
S Supporting Information *
ABSTRACT: In this work, a TFB (poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(4,4′-(N-(4-s-butylphenyl)diphenylamine)]) sensor with a cylindrical nanopore structure exhibits a high sensitivity to ammonia in ppb-regime. The lifetime and sensitivity of the TFB sensor were studied and compared to those of P3HT (poly(3-hexylthiophene)), NPB (N,N′-di(1-naphthyl)-N,N′-diphenyl-(1,1′biphenyl)-4,4′-diamine), and TAPC (4,4′-cyclohexylidenebis[N,N-bis(4-methylphenyl) benzenamine]) sensors with the same cylindrical nanopore structures. The TFB sensor outstands the others in sensitivity and lifetime and it shows a sensing response (current variation ratio) of 13% to 100 ppb ammonia after 64 days of storage in air. A repeated sensing periods testing and a long-term measurement have also been demonstrated for the test of robustness. The performance of the TFB sensor is stable in both tests, which reveals that the TFB sensor can be utilized in our targeting clinical trials. In the last part of this work, we study the change of ammonia concentration in the breath of hemodialysis (HD) patients before and after dialysis. An obvious drop of breath ammonia concentration can be observed after dialysis. The reduction of breath ammonia is also correlated with the reduction of blood urea nitrogen (BUN). A correlation coefficient of 0.82 is achieved. The result implies that TFB sensor may be used as a real-time and low cost breath ammonia sensor for the daily tracking of hemodialysis patients. KEYWORDS: gas sensor, ammonia, organic, lifetime, breath, hemodialysis
N
technology for the detection of breath gases has been well established.27−29 By using a selected ion flow tube technique, a correlation coefficient of 0.57 between the logarithms of ammonia concentration and BUN levels from 26 continuous ambulatory peritoneal dialysis (CAPD) patients has been reported.17 Similar results have also been realized from studies on hemodialysis (HD) patients.18−24 Measurements from ion mobility and cavity ring-down spectroscopy have been reported. Both spectroscopic methods demonstrate that breath ammonia directly correlates the BUN levels to correlation coefficients of ∼0.8.21 More recently, a 0.85 correlation coefficient was also found by making use of a colorimetric tube.24 In the past few years, numerous studies have reported the use of organic sensors to detect breath ammonia.22,30−36 A device based on a polyaniline nanoparticle sensor for measuring
umerous research studies focusing on the use of breath analysis to detect a variety of diseases have been reported in the literature.1−10 For example, nitric oxide in breath is regarded as a biomarker for asthma,11−13 while acetone is considered to be a biomarker for diabetes.14−16 Moreover, ammonia in breath is related to kidney diseases.17−24 Patients suffering from end-stage renal disease (ESRD) rely on dialysis treatment to remove waste and toxins from their bodies. In general, the adequacy of dialysis is determined by urea clearance.25 Related parameters, such as urea dynamics (Kt/ V) and urea reduction rate (URR), are based on the calculation of pre- and post-dialysis blood urea nitrogen (BUN) levels. Moreover, the simultaneous reduction of breath ammonia concentration and BUN levels has been observed during dialysis,17−24 implying that breath ammonia detection may serve as a real-time in situ monitor to determine dialysis efficacy. An adequate dialysis dose is supposed to lead to better clinical outcomes,26 while inadequate treatment could result in secondary organ failure and rapid death.18 To date, widespread © XXXX American Chemical Society
Received: August 10, 2017 Accepted: November 10, 2017 Published: November 10, 2017 A
DOI: 10.1021/acssensors.7b00564 ACS Sens. XXXX, XXX, XXX−XXX
Article
ACS Sensors breath ammonia in HD patients has also been developed.22 Organic-based sensors are promising because of their low production cost. These sensors also have advantages on the integration of a miniaturized sensing system. However, most organic sensors suffer from lifetime issues and degrade under air ambience. Moreover, the quantities of water vapor strongly affect the current in organic sensors. Hence, their short lifetime restricts the utilization of organic sensors to single-use disposable sensors. In our previous work,37 we reported a PQT-12 sensor that exhibits a good storage lifetime (25 days). A continuous sensing operation in air, however, is yet to be achieved. As a result, we aim to develop a sensor whose sensitivity can be retained for longer periods of time, even after continuous operation. In this work, we have tested several organic materials as sensing layers. These include TFB, P3HT, NPB, and TAPC. The sensor structure is based on numerous cylindrical nanopores.38 Every sensor exhibited high sensitivity toward ammonia in ppb-regime (part per billion by volume) concentrations. However, the TFB sensors exhibited far longer lifetimes than the other sensors. Therefore, we focused on TFB sensors and studied their robustness. The TFB sensors exhibited similar sensing responses after frequent operation and could last for over 60 days. Finally, we utilized the TFB sensors to study the variation in breath ammonia before and after dialysis. Forty HD patients were recruited for the clinical trial. The results indicated that the breath ammonia concentration was significantly reduced after dialysis. A correlation coefficient of 0.82 between the natural logarithm of ammonia concentration and natural logarithm of BUN levels was recorded; this was in good agreement with other reported results.17−24 Interestingly, among these studies, a higher correlation coefficient was observed for individual patients,18,22 revealing that a more precise result can be achieved on a case by case basis. Over the past few years, with advances in technology and the popularization of medical resources, the demands on personal health care have also increased. Thus, the development of a medical database for individuals should provide more accurate information and be cost-effective. For example, real-time monitoring during dialysis can afford better control of dose adjustment as this varies for each individual; this prevents waste of medical resources. So far, the low-cost and long-lifetime organic sensors employed in this study are exhibiting highly correlated results from 40 HD patients. We believe that these results can be further improved by focusing on tracing ammonia concentrations in individuals, and thus, we are developing a personal database as the next step of our research.
■
Figure 1. Sensor fabrication processes: (a) spin-coating a PVP layer and a thin P3HT layer on an ITO patterned glass substrate. (b) Coating PS nanospheres on the P3HT layer. (c) Depositing the Al top electrode. (d) Removing PS nanospheres. (e) Oxygen plasma etching. (f) Spin-coating different sensing materials to conclude sensor fabrication. (g) SEM image of the TFB sensor. The scale bar represents 500 nm. immediately after dipping into IPA. These nanospheres served as a shadow mask during subsequent deposition of the Al electrode. They were then removed with 3M Scotch Tape to form a porous top electrode. The thickness of the Al electrode was 40 nm and the diameter of the holes was ∼200 nm. Next, to form a cylindrical pore structure, the PVP in the areas not covered by the electrode was etched by oxygen plasma. The process steps are reported in detail in our previous work.38 In Figure 1f, four types of materials were bladecoated onto this structure. These include TFB (0.7 wt % in toluene), P3HT (1.5 wt % in chlorobenzene), NPB (1.3 wt % in chloroform), and TAPC (1.5 wt % in chloroform). The TFB sensor was annealed at 180 °C for 40 min, while the P3HT sensor was annealed at 200 °C for 10 min. Both the NPB and TAPC sensors were annealed at 80 °C for 10 min. All devices were stored under atmosphere after fabrication. Figure 1g illustrates the scanning electron microscope (SEM) image of the TFB sensor (SEM images of the other sensors are displayed in Figure S-1. The TFB covers the surface of the cylindrical nanopores with a thickness of ∼30 nm. The cylindrical pore structure is feasible for a variety of sensing materials. Moreover, it increases the surface to volume ratio to realize ppb-regime sensitivity.38 Sensing System. As illustrated in Figure 2, the sensing system comprises a desiccation cylinder, sensing chamber, rotameter, pump,
Figure 2. Picture of the sensing system. The components include a desiccation cylinder, a chamber, a rotameter, and a pump. The voltage is applied by a portable Keysight U2722A USB Modular Source Measure Unit.
EXPERIMENTAL SECTION
Fabrication of the Ammonia Sensor. Ammonia sensors with cylindrical nanopore structures were fabricated via the colloid lithography method. The process steps are presented in Figure 1. First, a cross-linkable poly(4-vinylphenol) (PVP) layer (thickness ∼250 nm) was spin-coated onto an indium tin oxide (ITO) patterned glass substrate and annealed at 200 °C for 1 h. A thin P3HT layer was then spin-coated on it as a surface modification layer to adsorb the polystyrene (PS) nanospheres.39,40 To coating the polystyrene nanospheres, the substrate was submerged into a dilute ethanol solution of negatively charged polystyrene (PS) spheres (Fluka). Optimized sphere densities (5−8 #/μm2) were obtained by using the concentration of the PS spheres as 0.24 wt % with 40 s soaking time. The wet substrate was then dipped into boiling IPA for 10 s, in which the bubble flow removed excess PS spheres not absorbed on the sensing layer. Finally, the substrate was blown dried by nitrogen
and electric signal measurement instrument (not shown in the picture). During measurement, the steady flow was controlled by the rotameter and pump in the terminal part of the system. The gases were first pumped through the desiccation cylinder to reduce water vapor and limit the humidity to a certain level. This was necessary because all our organic sensors were sensitive to changes in humidity. Particularly, during breath ammonia testing, the humidity in human breath is normally >90% RH (relative humidity), while the indoor ambient air humidity ranges from 40% to 70% RH. Thus, if the breath gases contact the sensors without first passing through a desiccation cylinder, the difference in humidity would lead to a change in current, thereby affecting the ammonia concentration measurements. In this B
DOI: 10.1021/acssensors.7b00564 ACS Sens. XXXX, XXX, XXX−XXX
Article
ACS Sensors work, NaOH (sodium hydroxide), which can maintain the humidity at ∼10% RH, was chosen as desiccant. After passing through sodium hydroxide, the dried gases then flowed into the sensing chamber. The sensor was mounted in the chamber with an electric signal feed passing through. In the clinical trial, the TFB sensor was chosen to detect breath ammonia. Selectivity to ammonia was examined by testing several main components in human breath,41,42 including acetone, carbon dioxide, isoprene, nitric oxide, and ethanol, as shown in Figure S-2. The sensor exhibited no response to 5% carbon dioxide and 2 ppm isoprene, and exhibited little response to 1 ppm acetone, 1 ppm nitric oxide, and 78 ppm ethanol. (Acetone, nitric oxide, ethanol, and isoprene in breath are normally
