Gas analyzer for continuous monitoring of trace level methanethiol by microchannel collection and fluorescence detection

Analytica Chimica Acta 841 (2014) 1–9

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Gas analyzer for continuous monitoring of trace level methanethiol by microchannel collection and fluorescence detection Kei Toda *, Haruka Kuwahara, Hidetaka Kajiwara, Kazutoshi Hirota, Shin-Ichi Ohira Department of Chemistry, Kumamoto University, 2-39-1 Kurokami, Kumamoto 860-8555, Japan

H I G H L I G H T S

G R A P H I C A L A B S T R A C T


An inexpensive and small system is developed for analysis of CH3SH in ambient air.
CH3SH is collected by a microchannel scrubber and reacted with DBD-F.
DBD-F reacts with CH3SH to produce fluorescence for fluorometric analysis.
With this method, CH3SH can be detected at the odor threshold.
CH3SH emitted from pulping and a piggery could be monitored continuously.

A R T I C L E I N F O

A B S T R A C T

Article history: Received 15 March 2014 Received in revised form 9 June 2014 Accepted 12 June 2014 Available online 13 June 2014

The highly odorous compound methanethiol, CH3SH, is commonly produced in biodegradation of biomass and industrial processes, and is classed as 2000 times more odorous than NH3. However, there is no simple analytical method for detecting low parts-per-billion in volume ratio (ppbv) levels of CH3SH. In this study, a micro gas analysis system (mGAS) was developed for continuous or near real time measurement of CH3SH at ppbv levels. In addition to a commercial fluorescence detector, a miniature high sensitivity fluorescence detector was developed using a novel micro-photomultiplier tube device. CH3SH was collected by absorption into an alkaline solution in a honeycomb-patterned microchannel scrubber and then mixed with the fluorescent reagent, 4-(N,N-dimethylaminosulfonyl)-7-fluoro-2,1,3benzoxadiazole (DBD-F). Gaseous CH3SH was measured without serious interference from other sulfur compounds or amines. The limits of detection were 0.2 ppbv with the commercial detector and 0.3 ppbv with the miniature detector. CH3SH produced from a pulping process was monitored with the mGAS system and the data agreed well with those obtained by collection with a silica gel tube followed by thermal desorption–gas chromatography–mass spectrometry. The portable system with the miniature fluorescence detector was used to monitor CH3SH levels in near-real time in a stockyard and it was shown that the major odor component, CH3SH, presented and its concentration varied dynamically with time. ã 2014 Elsevier B.V. All rights reserved.

Keywords: Methyl mercaptan Odor analysis Micro gas analysis system (mGAS) 4-(N,N-dimethylaminosulfonyl)-7-fluoro2,1,3-benzoxadiazole (DBD-F) Pulping process Odor from pig production

1. Introduction

* Corresponding author. Tel.: +81 96 342 3389. E-mail address: [email protected] (K. Toda). http://dx.doi.org/10.1016/j.aca.2014.06.019 0003-2670/ ã 2014 Elsevier B.V. All rights reserved.

Reduced sulfur compounds are highly odorous and are emitted naturally and anthropogenically from many sources. Among these compounds, methanethiol (CH3SH, common name methyl mercaptan) is the most odorous and is naturally emitted with

2

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

biological activity. Emission of methanethiol is high from anthropogenic sources such as landfill facilities [1], pit latrines [2] and pig farming [3]. Chemical production of methanethiol is also a problem in industries such as pulp and paper production [4]. The odor threshold for CH3SH, 1.6 ppbv (parts-per-billion in volume ratio), is four orders of magnitude smaller than that for NH3 (52 ppmv) [5]. Consequently, CH3SH detection at low concentrations is required for odor control. However, measurement of CH3SH is very difficult [6] and there are no reports of analytical methods for continuous CH3SH detection at levels around the odor threshold. Pandey and Kim [7] and Ras et al. [8] reviewed the available methods for determination of reduced sulfur compounds. The most common method for trace level determination of CH3SH involves collection into a stainless steel canister or a Tedlar bag in the field and subsequent laboratory analysis. In the laboratory, the gas is preconcentrated at liquid argon temperature, thermally desorbed to introduce into a gas chromatograph (GC) and detected by a flame photometric detector. Loss of CH3SH through reaction and adsorption during transport and storage is of a concern in the conventional method. Deshmukh et al. determined the odor intensity in methyl mercaptan equivalents using an electric nose comprising an array of commercial gas sensors and a data processor [9]. The electric nose gave real time data but could only detect CH3SH at concentrations down to the ppmv levels. Soft ionization mass spectrometry is an attractive real time method for CH3SH detection [3] but it is extremely costly and too large for mobile monitoring. Kudryavtsev et al. demonstrated CH3SH analysis with an original portable MS coupled with a preconcentration device [10]. The same inline preconcentration system was previously developed by our group for a field instrument called a single column trapping/ separation–ozone-induced chemiluminescence measurement (SCTS-CL) [11,12]. The SCTS-CL was applied to monitoring of CH3SH emitted from a pulp plant on the Lake Baikal shore [13] and indoor air monitoring at a pig farm [14]. Automated field monitoring on a ship and a car with the SCTS-CL instrument was successful, with sequential analyses conducted at 15 min intervals. We recently developed a micro gas analysis system (mGAS) to monitor atmospheric H2S and SO2 [15], HCHO in forests and city streets [16], and NO in exhaled breath [17]. The mGAS can also be used to measure unstable dissolved volatile compounds in natural water at trace levels (nanomoles per liter order) by coupling with a vapor generator [18]. A honeycomb-patterned microchannel scrubber, a key part of the mGAS, enables collection of gaseous CH3SH, H2S and SO2 into an alkaline solution. The fluorescent reagent, 4-(N,N-dimethylaminosulfonyl)-7fluoro-2,1,3-benzoxadiazole (DBD-F) [19], is commercially available for derivatization of amino [20] and thiol compounds [21] for their HPLC analysis. We tested this reagent for gaseous CH3SH analysis and found that gaseous CH3SH could be analyzed without interference from coexisting gases. DBD-F is a non-fluorescent

benzoxadiazole compound, and its fluoro functional group is exchanged with SCH3 of CH3SH to produce a highly fluorescent derivative, DBD-SCH3 (Scheme 1). In this work, the system and absorbing/reagent solutions were optimized for CH3SH analysis. Unfortunately, the miniature fluorescence (FL) detector previously developed for the mGAS [15] was not sufficiently sensitive to detect DBD-SCH3 FL. Therefore, a high performance miniature FL detector was developed using a novel micro-photomultiplier tube (mPMT) for field monitoring. The performance of the method was evaluated by comparison of the obtained data with those from thermal–desorption gas chromatograph–mass spectrometry (TD–GC–MS). The system was used to measure CH3SH emissions from a laboratory-scale pulping process and in the field beside a piggery enclosure. 2. Experimental 2.1. Chemicals DBD-F was obtained from Tokyo Kasei (Tokyo, Japan). A 1 mmol L1 DBD-F stock solution was prepared by dissolving DBD-F in acetonitrile (ACN) and was stored in a refrigerator. A working reagent solution was prepared by diluting the DBD-F stock solution 100 times with purified water. The final composition of the working reagent solution was 10 mmol L1 DBD-F in an aqueous 1% ACN solution. The absorbing solution for collection of CH3SH was 0.1 mol L1 NaOH. The following chemicals were used to investigate interference in the aqueous phase: sodium methanethiol (CH3SNa) and anhydrous sodium sulfide from Sigma–Aldrich (St. Louis, MO), dimethyl sulfide (DMS) from Wako Pure Chemical Industries (Osaka, Japan), and 1-propanethiol from Kanto Chemical (Tokyo, Japan). Each of these compounds was dissolved in 0.3 mol L1 NaOH. For amine compounds, ammonium chloride from Wako Pure Chemical Industries and L-cysteine and histamine from Nacalai Tesuque (Kyoto, Japan) were dissolved in purified water. For gas tests, the following standard gases were used: CH3SH, H2S, DMS, SO2, CS2, and NH3. The gas concentration in each cylinder was 100 ppmv with the balance as nitrogen gas. Ethanethiol and trimethylamine gases were prepared using permeation tubes P-72-H (0.90 mg min1 at 20  C) and P-180-H (13 mg min1 at 20  C) (Gastec, Kanagawa, Japan), respectively. 1Propanethiol gas was prepared by placing liquid 1-propanethiol in a diffusion tube (D-01, 1.6 mm i.d.  60 mm, Gastec) to obtain a vaporization rate of 0.83 mg min1. All gases were diluted with Peltier-cooler dried air purified with columns packed with silica gel and soda lime/activated carbon. The gas concentration was adjusted by controlling the flow rates of the cylinder gas and the purified air using mass flow controllers (Models 3600 and 3666A, Kofloc, Kyoto, Japan). Low concentrations (ppbv) of gases were prepared by diluting twice with the purified air.

Scheme 1. Reaction of CH3SH with DBD-F.

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

3

Fig. 1. Measurement system for gaseous CH3SH. mRP: micro-ring pump, HMCS: honeycomb-patterned microchannel scrubber, FD: commercial flow-through FL detector, mFD: micro FL detector, AC: activated carbon column, 3SV: three-way solenoid valve, AFC: airflow controller, AP: airpump.

2.2. Measurement system and procedure The measurement system developed for gaseous CH3SH is shown in Fig. 1. An air sample was aspirated with a miniature airpump (CM-15-12, Enomoto Micro Pump, Tokyo, Japan) and the flow rate was controlled with a flow meter equipped with a needle valve (RK200-V, Kofloc). A three-way solenoid valve (3SV) switched the sampling, and when the valve was on, air was introduced through an activated charcoal column to remove CH3SH to obtain a baseline reading. An absorbing solution, typically 0.1 mol L1 NaOH, was introduced in a honeycombpatterned microchannel scrubber [15] at 120 mL min1 with the help of two micro-ring-pumps (RP-Q1-N-P20A-DC3V, Aquatech, Osaka, Japan) equipped with Norprene tubing (1.2 mm  2.5 mm). In the microchannel scrubber, acidic gases in the air channel diffused through a gas permeable membrane and into the alkaline absorbing solution flowing in the microchannel. The eluent from the scrubber was mixed with a stream of reagent solution (10 mmol L1 DBD-F) and thiol compounds reacted with DBD-F in the reactor. The reactor consisted of a 0.5 mm i.d. Teflon tube coiled around a porcelain resistor (47 V, w7 mm  50 mm) powered (12 V) at 70  C with a temperature controller (E5CNRTC, Omron, Kyoto, Japan) and a thermocouple. Generation of the fluorescent product was monitored with a FL detector equipped with a flow-through-cell (FP-2020 plus, Jasco, Tokyo, Japan) at lEx 390 nm and lEm 526 nm (gain 1000, attenuation 1, bandwidth of emission light 40 nm). The signal of the FL detector was recorded every second by a data logger (GL-200, Graphtec, Yokohama, Japan). All the reagent and waste bottles were 50 mL plastic sample tubes. The absorbing solution tube was hung 20 cm above the microchannel scrubber in order to make the solution supply smooth with the gravity. 2.3. Measurement of CH3SH produced during the pulping process A pulping process [22] was conducted on a laboratory-scale. CH3SH emitted from the reactor was measured using the developed method and trapping with silica gel column followed by TD–GC–MS, and the data from the two methods were compared. In a stainless steel pressure cooker, 12 g of NaOH, 11 g of Na2S 9H2O, 18 g of Na2CO3 and 24 g Na2SO4 were dissolved in 1 L of water. Wood chips (70 g) were added, and the pressure cooker

pan was tightly sealed with a lid. The pressure cooker was then placed on a hot plate, and vapor coming out from the vent of the cooker was exhausted together with room air through a plastic chimney as in Fig. 2. Air in the chimney was aspirated by the mGAS at 200 mL min1 and an extra airpump equipped with a mass flow controller at 100 mL min1. Because removal of water vapor was necessary for silica gel trapping, water vapor was condensed using two ice baths in series and the remaining water vapor was removed by a calcium chloride column. Then the gas was introduced to the mGAS and simultaneously to glass column (4 mm i.d.  89 mm) packed with 400 mg of Davison silica gel (Supelco, St. Louis, MO). Both ends of the silica gel column were plugged with silanized quartz wool. The mGAS analysis was performed in continuous monitoring mode or 5 min cyclic mode, and the silica gel column for GC analysis was exchanged every 30 min. The gas collection tubes were settled in a thermal desorption instrument (TurboMatrix 650ATD, PerkinElmer Japan, Yokohama, Japan) and heated at 150  C for 3 min to desorb CH3SH. The desorbed gas was trapped in a secondary focusing column packed with Tenax TA at 20  C and again desorbed at 150  C to introduce to the GC–MS via a transfer line (200  C). The analytes were introduced to the MS detector via a separation column (InertCap 1MS, 0.32 mm i.d.  60 m, 0.4 mm, GL Sciences, Tokyo). The column temperature was initially held at 35  C for 4 min, increased to 50  C at 5  C min1, to 200  C at 40  C min1, and held at 200  C for 4 min. CH3SH was monitored at m/z 48 in selected ion monitoring mode, and had a retention time of 4.1 min. 2.4. Miniature FL detector and instrumentation for field use A miniature FL detector (mFD) was developed for field analysis instrumentation. All devices were placed in an aluminum box (w 35 cm  d 15.5 cm  h 17.5 cm) for field use. The schematic structure of the mFD is shown in Fig. 3. The mFD consisted of a black micro FL flow cell (m-cell), a pair of light emitting diodes (LEDs) and a micro PMT (mPMT). The m-cell (LCFL50-B-1, 5.7 mm  5.7 mm  12 mm, GL Sciences, Tokyo, Japan) had a rectangular flow channel (1.0 mm  1.0 mm) where the two LEDs (lmax 405 nm, OSSV5111A, Optosupply, Fo Tan, Hong Kong), which were powered at 15 mA, were placed facing the optical windows of m-cell. The remaining optical window was placed on the photo detective area (1 mm  3 mm) of the mPMT

4

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

Fig. 2. Process monitoring for laboratory-scale pulping process using mGAS and a silica gel trap followed by TD–GC–MS analysis. Craft pulping was performed in a pressure cooker and steam from the vent was exhausted to a laboratory duct. Sample gas was aspirated to the mGAS and silica gel trap tube (SGT) at flow rates of 200 and 100 mL min1, respectively. Water vapor in the sample gas was removed by two water traps cooled with ice water and a small column packed with CaCl2. MFC: mass flow controller, P: airpump, F: line filter.

(H12400-00-01, Hamamatsu Photonics, Hamamatsu, Japan) via two sheets of sharp-cut optical filter (SC460, Fujifilm, Tokyo, Japan). A high voltage of 666 V was applied to the mPMT by a DC–DC converter (OPTON-1NC-15, Matsusada, Japan). The multiplied photo current was converted to voltage by an operational amplifier circuit with 10 MV feedback resistor, and the signal via a voltage follower was recorded with a data logger (midi Logger 200A, Graphtec, Japan) at 1 Hz. The current-to-voltage converter and voltage follower circuits were prepared with a dual operational amplifier chip including junction gate field-effect transistor inputs (TL082, Texas Instruments, Dallas, TX).

Consequently, the signal response increased as the NaOH concentration increased up to 0.1 mol L1, and then gradually decreased at NaOH concentrations above 0.1 mol L1 (Fig. 5). It should be noted that the final NaOH concentration at the detector was half than that of the absorbing solution. From these results, 0.1 mol L1 NaOH was selected for the optimum CH3SH absorbing solution. We previously developed miniature FL detectors for H2S [15,23,24] and NO [17] measurements using a transparent Teflon

3. Results and discussion 3.1. Characteristics of DBD-F for gaseous CH3SH analysis Fluorescence of DBD-F derivatives is usually measured in ACN [19–21], but an aqueous alkaline solution was required in this work to collect CH3SH in the wet scrubber. Accordingly, we explored the FL characteristics in various water/ACN mixtures and investigated the effect of pH. Spectra of 10 mmol L1 DBD-SCH3 obtained with different ratios of water/ACN are shown in Fig. 4a. The FL intensity decreased and both the excitation and emission maxima shifted to longer wavelengths as the water content increased. While the excitation and emission maxima were at 385 nm and 501 nm, respectively, in the original spectra, the excitation and emission maxima were at 391 nm and 518 nm, respectively, in aqueous 1% ACN. The maximum FL intensity was four times lower in the aqueous medium (1% ACN) than in pure ACN. Furthermore, the FL intensity decreased as the NaOH concentration increased (Fig. 4b). However, the product DBD-SCH3 still fluoresced even in 0.25 mol L1 NaOH, and FL detection of DBD-F/thiol products could be performed in the aqueous alkaline medium. The absorbing solution had to be alkaline to collect and retain CH3SH. The mGAS response for “gaseous” CH3SH was examined with various concentrations of NaOH for the absorbing solution (Fig. 5). No response was obtained without NaOH because it was difficult to collect the acidic gas by absorption into water. The acid dissociation constant, Ka, of CH3SH is 1013.6, which is six orders of magnitude lower than that of H2S (Ka 107.0). Strongly alkaline solutions provide good CH3SH collection, but high concentrations of NaOH reduced the FL intensity as shown in Fig. 4b.

Fig. 3. Structure of the miniature FL detector and a photograph of the black plastic box for the cell and white box for the amplifier circuit. The potentiometer on the white box adjusts the high voltage for the mPMT.

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

(a)

Fluorescence intensity

1600

Ex

(b)

500

Em

5

ACN 100%

Em

Ex 400

1200

0.038 M

80% 50%

800

30% 10%

300

0.11 M

200

0.2 M

1%

400

NaOH 0.009 M

0.25 M 100

0 300

400

500

600

0 300

400

500

600

Wavelength (nm)

500

(c)

Ex

Filter transfer efficiency

80

400

Cathode radiation 300 sensitivity (mA/W)

60 40

Em

LED radiation

200

20

100

Fluorescence intensity

Sensitivity (mA/W), transfer efficiency (%), LED

100

Sensitivity with filter (mA/W)

0 300

0 400 500 Wavelength (nm)

600

Fig. 4. Fluorescence spectra of 10 mmol L1 DBD-SCH3 in water/ACN solutions of different ratios with 0.009 mol L1 NaOH (a), and in different NaOH concentrations with 1% ACN (b). Optical characteristics of LED, mPMT (cathode radiation sensitivity in mA W1) and filter are shown in panel (c) together with the FL spectrum of DBD-F/CH3SH product in 0.009 mol L1 NaOH and 1% ACN aqueous medium.

tubing cell equipped with a photodiode or a PMT module, respectively. Both analyses used fluorescein-framed reagents, namely fluorescein mercuric acetate for H2S and diaminofluorescein-2 for NO, and sufficient FL intensity was obtained with the simple flow FL detector. By contrast, the FL of DBD-SCH3 was 1/100th that of the fluorescein reagents, and much more effective detection was required for DBD-SCH3. Fig. 4c shows spectra of DBD-SCH3 and the LED emission, cathode radiation sensitivity of the mPMT in mA W1, and filter transfer efficiency. The LED emission overlapped well with the excitation spectrum of DBDSCH3, and the actual mPMT sensitivity compensated with the filter efficiency matched well with the DBD-SCH3 emission spectrum. Accordingly, the developed mFD with the mPMT should perform well for detection of DBD-SCH3 FL.

3.2. Optimization of reactor temperature The reaction of CH3SH with DBD-F is relatively slow, and acceleration was required for continuous CH3SH analysis. Therefore, the effect of the reactor temperature was examined (Fig. 6). The reactor was a Teflon tube coiled around a small porcelain resistor (47 V) powered by 12 V DC. The temperature was monitored with a small wired thermocouple, and data fed back to the power supply control to maintain constant temperature. No fluctuation was observed in the temperature. While the peak height for 2 and 5 ppbv CH3SH increased with the reactor temperature up to 70  C, bubbles formed in the reactor and interrupted the FL signal above 70  C. Therefore, 70  C was chosen as the optimal reactor temperature.

6

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

(a) with FD for HPLC

5 ppbv

150

CH3SH 8 ppbv

200 mV

5

100 2

2 ppbv

1

0

50

20 min

0

0.0

0.1

0.2

0.3

0.4

NaOH (M)

(b)

Fig. 5. Effect of absorbing solution alkalinity on the responses to gaseous CH3SH obtained in the mGAS operation.

15 min

The developed mGAS can monitor CH3SH either continuously or cyclically every 5 min. When the gas concentration changed from 0 to 5 ppbv, the signal started to increase at a lag time of 0.58 min and reached 90% of the final signal at 2.0 min. Data were obtained every second with a response time of 2.0 min in a continuous mode, which was conducted without using three-way solenoid valve (3SV). In cyclic mode, the baseline and signal were obtained alternately by switching the 3SV, and this allowed data collection for low concentrations. A 90% response was obtained 1.4 min after the sample introduction began, and the sample introduction time was set as 2 min for the final protocol in cyclic mode. After sample introduction, the CH3SH removal column was used for purification for 3 min to return the signal to the baseline. Therefore, data were obtained every 5 min in cyclic mode. Performance of the system was tested with the commercial FL detector. Responses to 0–3, 5 and 8 ppbv CH3SH are shown in Fig. 7a. Responses were obtained three times for each concentration by switching the three-way solenoid valve to introduce the test gas and CH3SH free gas alternately for 2 and 3 min, respectively. Good repeatability was obtained as shown in Fig. 7a, and the calibration curve (see Fig. S1 in the Supplementary data) showed good linearity with a measurable range up to

250 200

5 ppbv

150

100

µFD output (V)

3.3. Performance of the gas analysis system

0.4

40 ppbv

20 ppbv

with HPLC FD

Peak height (mV)

3

10 ppbv

0.6

0.4

0.2

5 ppbv

FD output (V)

Peak height by µGAS (mV)

200

0.0 0.2

with µFD 0.0

Fig. 7. Response signals of the mGAS for the CH3SH test gas obtained with a commercial FL detector (a) and upper in (b), and the mFD comprising the LEDs and mPMT bottom in (b).

100 ppbv with R2 = 0.999 (Fig. S2). The limits of detection and quantification were 0.2 (S/N = 3) and 0.6 ppbv (S/N = 10), respectively. Repeatability was tested by repeated alternate introduction of CH3SH and the purified air as described above, and the results were 3.2% for 2 ppbv CH3SH and 1.4% for 5 ppbv CH3SH (n = 20, 5 min cycle, see Fig. S3 for raw signals). Next, the mGAS was tested with the mFD containing the LEDs and mPMT to compare with the commercial HPLC FL detector which was connected after the mFD (note: larger dispersion of sample zone at HPLC FL than the case of Fig. 7a). The mPMT current was amplified with an operational amplifier. Even with the simple detector, good response curves were obtained as shown in Fig. 7b, and this was comparable to the response signal obtained with the commercial FL detector, for which the gain was set as 1000 (highest sensitivity mode). The signal-to-noise ratio was a little worse than that with the commercial FL detector. The limit of detection was 0.3 ppbv with the mFD, and the performance was sufficient for CH3SH detection at the odor threshold level. This is the first portable instrument that is capable of measuring CH3SH odors, and this is the first report of application of a mPMT to FL detection. 3.4. Interferences

2 ppbv

50

0

0

20

40 60 o Temperature ( C)

80

100

Fig. 6. Effect of the reactor temperature on signals for 2 and 5 ppbv CH3SH.

Because DBD-F had been developed for derivatization of aminoand thiol-compounds for HPLC analyses [19–21], gaseous amines and other reduced sulfur gases may interfere with the CH3SH analysis. Fluorescence spectra were examined to see if 10 mmol L1 DBD-F reacted with excess sulfur compounds (CH3SNa, 1-propanethiol, Na2S, DMS, L-cysteine) and amines (NH3, trimethyl amine (TMA), histamine) after 10 min at 70  C in 50% aqueous ACN. The

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

7

Table 1 Interference from other sulfur compounds and amines.

Methyl mercaptan

Aqueous with excess of the compoundsa

Gas

Relative FL intensityb

Concentration (ppbv)

Response (mV)

Relative response

1.00

2 7 12

49.2  3.0 239  7.1 406  13

1.00

50 60 80 120

41.4  3.0 46.9  1.0 57.9  3.6 92.4  3.2

0.022

35 70 115 10  103 10  103 10  103 10  103

1.36  0.050 2.68  0.045 4.33  0.067 NDd ND ND ND

0.0011

10  103 1.3  103

ND ND

Ethyl mercaptan

Propyl mercaptan 0.042 Hydrogen sulfide Dimethyl sulfide Sulfur dioxide Carbon disulfide L-Cysteine Ammonia Trimethyl amine Histamine a b c d

ND ND

ND ND ND 0.0009c

Spectra in 50% acetonitrile (10 mM DBD-F was all reacted with 300 mM of the compounds). FL intensities were obtained at Ex 388 nm and Em 512 nm. FL intensity for histamine was indicated as the value obtained at 512 nm as same as the other compounds, but maximum FL was at Ex 421 nm and Em 544 nm. ND: not detected.

results are summarized in Table 1. Among the amino compounds, NH3 and TMA did not produce FL. Only very weak FL was observed with histamine, which gave a FL intensity of less than 1/1000th that of DBD-SCH3, and the excitation and emission maximums were at longer wavelengths (421 nm and 554 nm, respectively). Because of its weak FL and non-volatility, histamine is not of concern as an interference in gaseous CH3SH analysis. Volatile NH3 and TMA did not show any interference in the batch-wise liquid phase tests. Among the reduced sulfur compounds, neither H2S nor DMS produced fluorescent products. The 1-propanethiol/DBD-F product produced FL but the intensity was only 1/20th that of DBDSCH3. The reaction with longer chain thiols likely needs a longer reaction time. Ethanethiol could not be obtained as liquid reagent, and was tested as gaseous material together with 1-propanethiol. The DBD-F based mGAS responded to gaseous ethanethiol as well as methanethiol. However, the intensity for ethanethiol was only 1/40th that of methanethiol (Fig. S4). Gaseous 1-propanethiol gave a lower response at 800 times less than that of CH3SH. These results indicate the DBD-F based mGAS can analyze gaseous CH3SH selectively.

because the emission increased as the reactor dried out. A similar situation was observed around 6 h in another run (Fig. 8b), and the emission decreased between 7 and 8 h because of completion of the treatment. With TD–GC–MS, a 30 min sampling time was required to obtain sufficient sensitivity. By comparison, the mGAS gave CH3SH data in near real time with much better time resolution. Data over 30 min from the mGAS were averaged and

3.5. Comparison of analytical data for the pulping process We previously investigated odorous sulfur gases around a paper and pulp plant on the shore of Lake Baikal [13], and found that CH3SH was the major odorous compound present. In the present study, we performed a similar process on a laboratory-scale and emitted CH3SH was monitored by the two methods, mGAS and TD– GC–MS after gas sampling with a silica gel column. The data obtained by the two methods are shown in Fig. 8. In the pulp and paper industry, wood chips in a pressure cooker are treated with an alkaline solution containing sodium sulfide to decompose lignin, and the lignin reacts with sulfide to produce CH3SH, CH3SCH3 and CH3SSCH3 [22]. The mGAS was operated in the cyclic mode with baseline readings and signals from the gas obtained alternately at 5 min intervals (Fig. 8a) or in continuous mode (Fig. 8b). While there was no emission of CH3SH initially, CH3SH emission gradually increased as the temperature was increased. After 8 h, the CH3SH concentration increased to be over 30 ppbv, probably

Fig. 8. Analyzed data of CH3SH emitted from the mini-pulp plant chamber measured by mGAS and TD–GC–MS. The mGAS was operated in cyclic mode (a) and continuous mode (b). The temperature indicated is the setting for the hot plate under the chamber (a), and measured reactor temperature (b).

8

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9

4. Conclusions A portable analytical system was successfully developed for detection of CH3SH for the first time. The proposed mGAS can analyze CH3SH continuously or in near real time even at concentrations around the odor threshold level. A miniature FL detector was developed using mPMT, and good sensitivity was achieved with the mobile apparatus. Acknowledgments This work was supported by Grants-in-Aid for Basic Research (C) (Grant No. 21550087) and Grants-in-Aid for Basic Research (B) (Grant No. 25288068) from the Japan Society for the Promotion of Science (JSPS). 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.aca.2014.06.019.

Fig. 9. Field monitoring data obtained at a pig farm.

References compared with data from the TD–GC–MS. The pulping process was run four times to compare the data from the two methods (see Fig. S5 for woodchip reactor picture and comparison data), and good agreement was obtained as shown in Eq. (1). CH3 SH by mGAS ðppbvÞ ¼ 1:008 CH3 SH by TD  GC  MS ðppbvÞ þ0:672

n ¼ 53; R2 ¼ 0:9633

(1)

The above correlation shows that data from the mGAS are reliable. With the mGAS, the gas concentration could be monitored on site without any treatment. 3.6. Field monitoring application Field monitoring was conducted at a pig farm. The mGAS device equipped with the miniature FL detector in an aluminum box was placed just behind the farm building. Simultaneous monitoring of H2S and NH3 was also conducted with previously developed devices [15,25]. These gas concentrations changed dynamically with time. NH3 was dominant among the three odor components detected, and its concentration varied between 0.1 and 1 ppmv. The level of NH3 was higher when it was more humid at midnight and in the early morning. The second most abundant gas was H2S, and its concentration reached 100 ppbv. In the measurement period of 23 h, the average concentrations were H2S 18.8 ppbv, CH3SH 0.783 ppbv, and NH3 417 ppbv. Although NH3 was the most abundant, NH3 was not a large contributor to the odor. According to an earlier study, the odor thresholds are H2S = 8.1 ppbv, CH3SH = 1.6 ppbv, and NH3 = 5200 ppbv [5]. The odor intensity values, which were obtained by dividing the concentration by each odor threshold, were H2S = 2.32, CH3SH = 0.473 and NH3 = 0.0780 on average over the 23 h measurement period. Time variations for the odor intensities are shown in Fig. 9. The H2S odor was the strongest as shown in Fig. 9, but CH3SH was also a major odor with an odor intensity value approximately 1/5th that of H2S. It should be noted that while the sensitivity of the human nose to H2S is initially very high it rapidly becomes insensitive. This is why people sometimes die unwittingly from H2S inhalation. By comparison, smell sensitivity to heavier thiol compounds remains for a longer time. Therefore, the odor perceived with CH3SH was thought to be stronger than the calculated odor intensity suggested. The CH3SH odor was the strong among the compounds analyzed, and it is important to monitor CH3SH at farms and biomass treatment facilities.

[1] K.-H. Kim, Emissions of reduced sulfur compounds (RSCs) as a landfill gas (LFG): a comparative study of young and old landfill facilities, Atmos. Environ. 40 (2006) 6567–6578. [2] J. Lin, J. Aoll, Y. Niclass, M.I. Velasco, L. Wünsche, J. Pika, C. Starkenmann, Qualitative and quantitative analysis of volatile constituents from latrines, Environ. Sci. Technol. 47 (2013) 7876–7882. [3] A. Feilberg, D. Liu, A.P.S. Adamsen, M.J. Hansen, K.E.N. Jonassen, Odorant emissions from intensive pig production measured by online proton-transferreaction mass spectrometry, Environ. Sci. Technol. 44 (2010) 5894–5900. [4] R. Pal, K.-H. Kim, E.-C. Jeon, S.-K. Song, Z.-H. Shon, S.-Y. Park, K.-H. Lee, S.-J. Hwang, J.-M. Oh, Y.-S. Koo, Reduced sulfur compounds in ambient air surrounding an industrial region in Korea, Environ. Monit. Access. 148 (2009) 109–125. [5] J.E. Amoore, E. Hautala, Odor as an aid to chemical safety: odor thresholds compared with threshold limit valuesand volatilities for 214 industrial chemicals in air and water dilution, J. Appl. Toxicol. 3 (1983) 272–290. [6] W. Wardencki, Review problems with the determination of environmental sulphur compounds by gas chromatography, J. Chromatogr. A 793 (1998) 1–19. [7] S.K. Pandey, K.-H. Kim, A review of methods for the determination of reduced sulfur compounds (RSCs) in air, Environ. Sci. Technol. 43 (2009) 3020–3029. [8] M.R. Ras, F. Borrull, R.M. Marcé, Sampling and preconcentration techniques for determination of volatile organic compounds in air samples, Trends Anal. Chem. 28 (2009) 347–361. [9] S. Deshmukh, A. Jana, N. Bhattacharyya, R. Bandyopadhyay, R.A. Pandey, Quantitative determination of pulp and paper industry emissions and associated odor intensity in methyl mercaptan equivalent using electric nose, Atmos. Environ. 82 (2014) 401–409. [10] A.S. Kudryavtsev, A.L. Makas, M.L. Troshkov, M. A. Grachev, S.P. Pod'yachev, Onsite determination of trace concentrations of methyl mercaptan and dimethyl sulfide in air using a mobile mass spectrometer with atmospheric pressure chemical ionization, combined with a fast gas chromatographic system, Talanta 123 (2014) 140–145. [11] M.A.K. Azad, S. Ohira, K. Toda, Single column trapping/separation and chemiluminescence detection for on-site measurement of methyl mercaptan and dimethyl sulfide, Anal. Chem. 78 (2006) 6252–6259. [12] T. Nagahata, H. Kajiwara, S. Ohira, K. Toda, Simple field device for measurement of dimethyl sulfide and dimethylsulfoniopropionate in natural waters, based on vapor generation and chemiluminescence detection, Anal. Chem. 85 (2013) 4461–4467. [13] K. Toda, T. Obata, V.A. Obolkin, V.L. Potemkin, K. Hirota, M. Takeuchi, S. Arita, T. M. Khodzher, M.A. Grachev, Atmospheric methanethiol emitted from a pulp and paper plant on the shore of Lake Baikal, Atmos. Environ. 44 (2010) 2427– 2433. [14] M.J. Hansen, K. Toda, T. Obata, A.P. Adamsen, A. Feilberg, Evaluation of single column trapping/separation and chemiluminescence detection for measurement of methanethiol and dimethyl sulfide from pig production, J. Anal. Methods Chem. (2012) 489239, doi:http://dx.doi.org/10.1155/2012/489239. [15] S. Ohira, K. Toda, Micro gas analysis system for measurement of atmospheric hydrogen sulfide and sulfur dioxide, Lab Chip 5 (2005) 1374–1379. [16] K. Toda, W. Tokunaga, T. Gushiken, K. Hirota, T. Nose, D. Suda, J. Nagai, S. Ohira, Mobile monitoring along a street canyon and stationary forest air monitoring of formaldehyde by means of micro-gas analysis system, J. Environ. Monit. 14 (2012) 1462–1472. [17] K. Toda, T. Koga, J. Kosuge, M. Kashiwagi, H. Oguchi, T. Arimoto, Micro gas analyzer measurement of nitric oxide in breath by direct wet scrubbing and fluorescence detection, Anal. Chem. 81 (2009) 7031–7037.

K. Toda et al. / Analytica Chimica Acta 841 (2014) 1–9 [18] K. Toda, H. Kuwahara, S. Ohira, On-site measurement of trace-level sulfide in natural waters by vapor generation and microchannel collection, Environ. Sci. Technol. 45 (2011) 5622–5628. [19] S.T. Uchiyama, N. Okiyama, N. Furushima, K. Imai, Fluorogenic and fluorescent labeling reagents with a benzofurazan skeleton, Biomed. Chromatogr. 15 (2001) 295–318. [20] H. Kawanishi, T. Toyo-oka, K. Ito, M. Maeda, T. Hamada, T. Fukushima, M. Kato, S. Inagaki, Rapid determination of histamine and its metabolites in micehair by ultra-performance liquid chromatography with time-of-flight mass spectrometry, J. Chromatogr. A 1132 (2006) 148–156. [21] J. Toyooka, H. Jinno, Determination of rat hepatocellular gluthathione by reversed phase liquid chromatography with fluorescence detection and cytotoxicity evaluation of environmental pollutants based on the concentration change, Biomed. Chromatogr. 15 (2001) 240–247.

9

[22] J.Y. Zhu, X.-S. Chai, X.J. Pan, Q. Luo, J. Li, Quantification and reduction of organic sulfur compound formation in a commercial wood pulping process, Environ. Sci. Technol. 36 (2002) 2269–2272. [23] K. Toda, P.K. Dasgupta, J. Li, G.A. Tarver, G.M. Zarus, Fluorometric field instrument for continuous measurement of atmospheric hydrogen sulfide, Anal. Chem. 73 (2001) 5716–5724. [24] K. Toda, S. Ohira, T. Tanaka, T. Nishimura, P.K. Dasgupta, Field instrument for simultaneous large dynamic range measurement of atmospheric hydrogen sulfide, methanethiol, and sulfur dioxide, Environ. Sci. Technol. 38 (2004) 1529–1536. [25] S. Ohira, M. Heima, T. Yamasaki, T. Tanaka, T. Koga, K. Toda, Flow-based ammonia gas analyzer with an open channel scrubber for indoor environments, Talanta 116 (2013) 527–534.