- Email: [email protected]
Performance evaluation of oxygen adsorbents using negative corona discharge–ion mobility spectrometry
Accepted Manuscript Performance Evaluation of Oxygen Adsorbents Using Negative Corona Discharge– Ion Mobility Spectrometry Kamal Azadkish, Mohammad T. Jafari, Hassan S. Ghaziaskar PII:
S0003-2670(16)31413-1
DOI:
10.1016/j.aca.2016.12.001
Reference:
ACA 234934
To appear in:
Analytica Chimica Acta
Received Date: 11 September 2016 Revised Date:
26 November 2016
Accepted Date: 1 December 2016
Please cite this article as: K. Azadkish, M.T. Jafari, H.S. Ghaziaskar, Performance Evaluation of Oxygen Adsorbents Using Negative Corona Discharge–Ion Mobility Spectrometry, Analytica Chimica Acta (2017), doi: 10.1016/j.aca.2016.12.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Graphical Abstract
ACCEPTED MANUSCRIPT
Performance Evaluation of Oxygen Adsorbents Using Negative Corona Discharge–Ion Mobility Spectrometry Kamal Azadkish, Mohammad T. Jafari∗, Hassan S. Ghaziaskar
RI PT
Department of Chemistry, Isfahan University of Technology, Isfahan 84156–83111, Iran
Abstract
SC
Trace amounts of oxygen was determined using negative corona discharge as an ionization source for ion mobility spectrometry. A point-in-cylinder geometry with novel design was used -
M AN U
to establish the corona discharge without interferences of negative ions such as NOX . The desirable background spectrum shows only electrons peak, providing the instrument capable of trace analysis of oxygen in gaseous samples. The limit of detection and linear dynamic range with high coefficient of determination (r2=0.9997), were obtained for oxygen as 8.5 and 2814204 ppm, respectively. The relative standard deviations of the method for intraday and interday were obtained 4 and 11 %, respectively. The satisfactory results revealed the ability of
TE D
the negative corona discharge ion mobility spectrometry for investigating the performance of synthesized oxygen adsorbents in nitrogen streams. Two oxygen scavengers of MnO and Cu powder were prepared and the optimum temperature of the reactor containing MnO and Cu powder were obtained as 180 and 230 °C, respectively. Due to higher lifetime of copper powder,
EP
it was selected as the oxygen scavenger and some parameters such as: the type of adsorbent support, the size of adsorbent particles, and the amount of copper were studied for preparation of
AC C
more efficient oxygen adsorbent.
Keywords: Oxygen Analysis; Negative corona discharge; Ion mobility spectrometry; Oxygen adsorbents ∗
Corresponding author: Tel.: +98-31-3391-2351, Fax: +98-31-3391-2350
E-mail: [email protected].
1
ACCEPTED MANUSCRIPT
1. Introduction The determination of oxygen (O2) in different gases such as helium, hydrogen, nitrogen, argon, carbon dioxide, mixed gases, process gases, and ambient air is very important because oxygen content is a major factor influencing gas or fuel quality, internal corrosion, and user safety. In
RI PT
addition, the purification of inert gases is necessary for different applications such as metallurgical processes and chemical reactions that need to have a neutral environment, the nuclear and electronic fields, and laboratory and device applications [1]. For example, oxygen is the most detrimental contaminant for columns of gas chromatography (GC). Due to irreversible
SC
oxidation of stationary phases by oxygen, the destruction of chromatographic columns or reduction of their lifetime is occurred [2]. In the GC with electron-capture detector, the presence
M AN U
of oxygen impurity in the carrier gas, even at trace level, leads to reduce the sensitivity [3]. Some physical and chemical processes are used for purification and filtration of contaminants from gaseous streams. The physical processes include adsorption, refrigeration and low temperature distillation, and the chemical processes contain the gas-solid reactions and using the metallic getters and catalytic processes such as catalytic combination of oxygen and hydrogen [1]. Yan et al. have been studied oxygen catalytic removal from biomass-derived syngas by the Pt, Pd, and
TE D
Cu catalysts [4]. Han et al. applied catalytic oxidation of CuO–CeO2/Al2O3 catalyst for oxygen removal from syngas [5]. On the other hand, utilizing the adsorption/absorption techniques are suitable methods for O2-removal from inert gas streams. In the adsorption/absorption techniques, membrane processes, oxygen scavengers (e.g. MnO, Cu, and FeO) and oxygen acceptors (e.g.
EP
CuCl2, SrO, and Cr2O3) compounds are usually used for oxygen removal from a gas stream under special conditions [6-9].
AC C
So far, some common types of oxygen analyzers such as zirconium oxide analyzers, paramagnetic devices, electrochemical analyzers, and GC with thermal conductivity detector (GC-TCD) have been used for analysis of O2. However, these methods suffer from some drawbacks and limitations, so they cannot be kindly used for routine and real-time analysis of O2. Zirconium oxide analyzers need to operate at temperatures above 810 K. Therefore, the gas stream requires being at an elevated temperature or the instrument needs a sample heater. Therefore, it is usually used for measuring the excess oxygen in burners. Paramagnetic oxygen analyzers are very precise, but, because of its moving parts, they are very delicate. The positionand vibration-sensitive devices need to compensate and temperature control. In electrochemical 2
ACCEPTED MANUSCRIPT
oxygen analyzers, since the oxygen concentration is proportional to generated electrical current by reaction of oxygen as oxidizing agent and electrolyte, other oxidizing gases can react with electrolyte, resulting in poisoning the cell and give false readings [10,11]. Finally, in addition to long run time, the low sensitivity is problematic in the oxygen analysis by GC-TCD method.
RI PT
Ion mobility spectrometry (IMS) is an analytical technique based on measurement of the electrophoretic mobility of ions (affected by ion’s mass, charge, size and shape) through a neutral gas [12]. In this technique, different ions can be formed during an ionization process and then separated inside a drift tube under a constant electric field at ambient pressure. The
SC
application of IMS was traditionally limited to analysis of vapors or gaseous samples mainly of chemical warfare agents, drugs of abuse, and explosives [13]. Contemporary IMS research has
M AN U
expanded enormously to the analysis of gaseous, liquid, and solid samples in many fields including biology, medicine, environmental studies, forensics, pharmaceuticals, and food research. The popularity of IMS as an analytical technique is due to its excellent figures-of-merit including low limits of detection (pg to µg), fast separation times (µs to ms), low cost for handheld or standalone devices, and high throughput (seconds per sample) [13]. Ionization source is one of the key parts in IMS instrument so that a variety of ionization sources have been
TE D
applied for this purpose. Two popular ionization sources of
63
Ni and corona discharge have
mostly been used for analytes in gaseous phase [13, 14]. In addition to non-radioactivity, the corona discharge dedicates higher sensitivity (~ 10 times) related to
63
Ni [15]. However,
application of corona discharge in negative mode has not been expanded due to serious problems
EP
originated form NOx formation in the source. NOx molecules have the highest electron affinity (~ 3.9 eV) among chemical compounds and thus they essentially reduce the sensitivity of IMS
AC C
for other analytes [16]. In addition, analysis of some analytes such as O2, SOx, COx, or halogen gases has not been possible in the presence of NOx species due to overlapping of their signals [17]. In our previous work [18], a novel design of point-in-cylinder geometry was used to -
establish the negative corona discharge without interferences of negative ions such as NOx , for analysis of Ethion as an organophosphorus pesticide. The background ion mobility spectrum in pure nitrogen containing only electrons peak was obtained, promising the possibility of O2 analysis by the instrument. In this work, for the first time the capability of negative corona discharge ion mobility spectrometry (NCD–IMS) for oxygen measurement in nitrogen gas stream has been evaluated. 3
ACCEPTED MANUSCRIPT
The background ion mobility spectra of instrument in two gases of nitrogen and air were obtained and lack of NOx species production was demonstrated. The proposed method for O2 analysis was applied to evaluate the performance of some synthesized oxygen adsorbents.
RI PT
2. Experimental 2.1. Materials
Materials such as silicon dioxide, alumina, and manganese (II) chloride monohydrate were
SC
obtained from Merck Co., Germany. Sodium carbonate, bentonite, copper (II) nitrate trihydrate, were purchased from Sigma-Aldrich Co., USA. To prepare all solutions through entire work four-time distilled water was used. The commercial grade 2,4,6-trinitrotoluene (TNT) was
M AN U
provided from Chemical Industries, Zarrinshahr, Isfahan, Iran and used as the internal standard in IMS.
2.2. Instrumentation
The ion mobility spectrometer used for this research was designed and constructed at Isfahan University of Technology and described previously in detail [18, 19]. The schematic diagram of
TE D
the IMS cell used in this work is shown in Fig. S1, in the Supporting Information. In brief, the IMS instrument included a cell containing the corona discharge needle, two high voltage power supplies, a pulse generator, an analog to digital converter, and a computer. The IMS cell consisted from 15 aluminum rings (8 mm thickness, i.d. 40 and 25 mm for drift and reaction
EP
regions, respectively) separated by PTFE insulators with 1 mm thickness. The length of drift and reaction regions was 11 and 5 cm, respectively and separated from each other by Bradbury-
AC C
Nielsen shutter grid. The detail of corona discharge ionization source without formation of -
negative ions of NOX was presented in an Iranian patent [20]. In this geometry, a sharp needle (0.1 mm diameter) was located inside an aluminum ring (i.d. 8 mm) which used as a cylindrical counter electrode. To remove water vapor or other contaminants, the gas was filtered using a 4Åmolecular sieve (Merck) in a 40 cm length × 25 mm i.d. cylindrical trap, before entering into the IMS cell. Table 1 summarizes all the operation conditions of NCD–IMS under which the mobility spectra were taken.
4
ACCEPTED MANUSCRIPT
2.3. Preparation of different concentrations of oxygen in nitrogen In order to draw the calibration curve, the exponential dilution flask method was used for preparation of different concentrations of oxygen. The usefulness of this method for calibration of gas chromatographic detectors and preparation of dilute gas samples has been discussed
RI PT
elsewhere [21]. In addition, this method has been successfully used for introducing the gas samples into the IMS [22, 23]. Generally, in this research purified air (20% O2 and 80% N2) was trapped in a flask with known volume (1225 mL) and continuously flushed out with the exact flow rate of pure nitrogen (540 mL min-1), as carrier gas. To detect the signal variation related to
SC
changing of oxygen concentration, flask outlet was directly connected to the IMS cell. In this method, the outlet oxygen concentration (C) after time (t) is given by Eq. (1).
M AN U
C = C0 exp (−Qt/V)
(1)
Where C0 is the initial concentration of the oxygen that was equivalent to 20% or 200000 ppm in air, Q is the gas flow rate and V is the flask volume.
2.4. Preparation of oxygen scavenger compounds
TE D
To investigate the performance of oxygen adsorbents for removal of oxygen from nitrogen gas, suitable oxygen scavengers such as MnO and Cu powder were prepared.
2.4.1. MnO Preparation
EP
To prepare the MnO adsorbent according to Eq. (2), appropriate amounts of MnCl2 (H2O) and Na2CO3 with proper stoichiometric ratios were individually dissolved in minimum amount of
AC C
water. The solution of Na2CO3 was slowly added to the solution of MnCl2 to form MnCO3. The precipitate was separated using Buchner funnel under vacuum, and kept in 50 ℃ for 10 h to dryness. The dried MnCO3 was transferred into a tubular reactor (316 stainless steel tube, 4 mm i.d. and 50 mm length) and heated at 400 ℃ for 4 h while passing pure argon gas (99.9995%). In this state, MnCO3 decomposed to MnO and CO2, according to equation (3) [24]. MnCl2 (aq) + Na2CO3 (aq) MnCO3 (s) Δ Ar
MnCO3 (s) + 2 NaCl (aq)
MnO (s) + CO2 (g)
(2) (3)
5
ACCEPTED MANUSCRIPT
2.4.2. Cu powder Preparation For preparation of Cu powder, the copper oxide was prepared according to equation (4) using calcination of copper (II) nitrate at temperature of 500 ℃ for 5 h [25]. Afterward, pure hydrogen
RI PT
gas was passed through the CuO powder with flow rate of 10-20 mL min-1 at temperature of 200 ℃. At this step, Cu powder would be prepared according to Eq. (5) [3]. 2 Cu(NO3)2 → 2 CuO + 4 NO2 + O2
(5)
SC
CuO + H2 → Cu + H2O
(4)
2.5. Cu adsorbents on different supports
To prevent agglomeration of copper powder and to increase the surface area and consequently its
M AN U
performance at lower temperatures, copper adsorbents was dispersed on different supports such as silicon dioxide, bentonite, and alumina, via the impregnation method.
2.5.1. Preparation of Cu adsorbent supported on silicon dioxide
For the preparation of Cu adsorbent supported on silicon dioxide with 40 w% Cu, 45.62 g
TE D
Cu(NO3)2.3H2O was dissolved in water and 18 g silicon dioxide was added to the solution. The solution was stirred at room temperature overnight. After impregnation, water was removed by vacuum drying at 80℃ in a rotary evaporator. The obtained powder was kept in 120 ℃ to dryness. The dried powder was calcinated in air at 500 ℃ for 5h. Finally, the adsorbent was
EP
prepared after passing pure hydrogen gas through the calcinated powder at flow rate of 10-20 mL
AC C
min-1 and temperature of 200 ℃.
2.5.2. Preparation of Cu adsorbent supported on bentonite and alumina Preparation of Cu adsorbent supported on bentonite and alumina (40 w%, Cu content) was accomplished by addition of 18.0 and 19.8 g of alumina and bentonite, to Cu(NO3)2.3H2O solutions containing 12 g Cu. Other steps were carried out similar to section 2.5.1. It is notable that the bentonite used in this work was sodium-type and its loss on ignition was 10%. So we ought to use 19.8 g to achieve 18 g bentonite after calcination.
3. Results and discussion 6
ACCEPTED MANUSCRIPT
3.1. Ion mobility spectra The background ion mobility spectrum of NCD-IMS with pure nitrogen (99.999% and used as drift and carrier gas) is shown in Fig. 1-A. For more clarification of signal to noise ratio, the background amplified 200 times which is shown as inset of the figure. The signal at drift time 6.3
RI PT
ms is related to the trace amount of oxygen impurity in the nitrogen gas. The spectrum shows a very large peak at zero drift time corresponding to electrons. As this figure obviously shows, no -
significant negative ions (e.g. NOx ) were formed during the negative corona discharge by using -
the novel design. Lack of NOx production has some major benefits, which are very important in
SC
both qualitative and quantitative analysis by NCD-IMS. Firstly, high electron affinity of these ions can prevent sufficient ionization of other substances even halogenated compounds [16],
M AN U
reducing the sensitivity of the device, extremely. Secondly, it is impossible to analyze NOX, themselves or SOx, COx, O2 and other chemical compounds with the same drift time as background ions [17]. Finally, NOX ions can form coordination complexes with interest analytes in some cases, resulting in complex spectra [26]. Consequently, very high signal to noise ratio was expected for the oxygen determination when NOx production was inhibited during the process of negative corona discharge.
TE D
Additionally, Fig. 1-B depicts the ion mobility spectrum of background in purified air (as drift/carrier gas in IMS) compared with that obtained after injection of 25 µL purified air, when pure nitrogen was used as drift/carrier gas. As this figure shows, when air was injected into the IMS with N2 drift gas, only one ion peak was originated by ionization in negative corona
EP
discharge. In order to calculate the reduced mobility value (Ko) of this negative ion, 2,4,6trinitrotoluene (TNT) with the Ko=1.55 cm2V-1s-1 [27], was used as standard compound. The Ko
AC C
value was calculated to be 2.59 cm2V-1s-1 which is well compared with 2.61 cm2V-1s-1, reported -
by Ewing and Waltman [17] for O2 ions. In order to show more compelling reasons for this claim, NOx gases, produced by pouring the copper powder in nitric acid solution, were injected into the IMS. Fig. 2 shows the ion mobility spectra obtained after injection of NOx gases, in two cases of drift/carrier gases of pure nitrogen (upper) and purified air (below). These spectra are compared with those of their corresponded backgrounds. To assign the originated ions in these spectra, the reduced mobility values were calculated and tabulated in Table 2. The reduced -
-
-
mobility values reported previously by Ewing and Waltman [17] for NO2 , NO3 , and O2 ions,
7
ACCEPTED MANUSCRIPT
are also shown in this Table. By comparison of these results, it can be concluded that the one ion -
-
peaks produced by air ionization is only O2 and none NOx ions. When we used pure nitrogen -
instead of air, the single ion peak of O2 was also disappeared in the mobility background and
RI PT
only the electron signal would be observed in the spectrum. These favorable results reveal the capability of NCD-IMS with pure nitrogen for qualitative and quantitative analysis of oxygen.
3.2. Analytical performance
As mentioned above, the exponential dilution flask method was used to prepare different oxygen
SC
concentration. The IMS signal originated from O2 was followed at the different times of air dilution process and the obtained results are shown in Fig. 3-A. The analyte concentrations were
M AN U
calculated using the equation (1) at different times and the linear dynamic range was prepared by plotting the IMS signals versus their corresponded O2 amount. The calibration curve (Fig. 3-B) shows an acceptable coefficient of determination (r2=0.9997) for oxygen in the linear dynamic range of 28-14204 ppm. The inset figure was also prepared in order to see better signals of instrument for low concentrations of oxygen (< 850 ppm). The limit of detection defined by signal–to–noise ratio of 3 was calculated to be 8.5 ppm. The relative standard deviations (RSD)
TE D
were achieved, 4 and 11% for one day (n=3) and over a period of three days (n=3), respectively. The favorable analytical results obtained for the analysis of oxygen in this method encourage us to investigate the performance of some synthesized oxygen sorbents, described in following sections.
EP
For more evaluation of the method, the relative errors of the method were calculated for some preapred gaseous solutions entered into the IMS cell. To that end, the solutions with
AC C
various oxygen concentrations of 1105, 5054, and 8171 ppm were prepared according to equation (5).
(5)
Ca = 20 ×
In the equation Ca is the prepared oxygen concentration in percent, Q2 and Q1 are the air and nitrogen gases flow rate, respectively. The IMS-signals were calculated and compared with those obtained by calibration curve. The relative error values were obtained 8.1, 3.3, and 2.2 %, for concentrations of 1105, 5054, and 8171 ppm, respectively. These satisfactory results insure the qualitative analysis of oxygen in different gaseous samples. 8
ACCEPTED MANUSCRIPT
3.3. Temperature optimization for O2 adsorption by Cu powder and MnO One of the most fundamental variables for a chemical reaction is the temperature. To evaluate the temperature effect on oxygen removal from nitrogen gas using Cu powder and MnO,
RI PT
temperatures of reactor containing 35 g of each sorbents were indivitualy changed, in the range of 25–250 ℃. For investigation of temperature effect, oxygen-contaminated nitrogen with flow rate of 1400 mL min-1 was passed through the reactor and then injected into the IMS instrument. According to the results shown in Fig. S2 in the Supporting Information, the signal of O2 was
SC
reduced while the reaction temperature increased to 180 and 230 ℃ for MnO and Cu powder, respectively. Therefore, these optimized temperatures were considered for study of the sorbents
M AN U
lifetime.
3.4. Relative lifetime of MnO and Cu powder in stream
For exploration of lifetime of oxygen scavenger compunds, nitrogen gas containing oxygen was passed through the reactor filled with 35 g of each compounds. The temperature was set at 180 and 230 ℃ for MnO and Cu powder, respectively and the flow rate of gas was 1400 mL min-1
TE D
during the adsorption of oxygen. The IMS signals were recorded while Cu powder and MnO were removing oxygen from the nitrogen stream. Herein, the liftime is defined as the time taken to saturate the adsoebent and hence, the O2 signal becomes equal to its initial intensity (before passing the gas through the adsorbents). The results are shown in Fig. S3 (in the Supporting
EP
Information) indicated that relatively copper powder had longer liftime (about 2 times), and
AC C
hence, it was chosen as more suitable oxygen scavenger and used for successive experiments.
3.5. The effect of support type After selection of copper powder, it was necessary to study the effect of support type on the adsorption efficiency. So, the lifetimes of 28 g of copper adsorbent on the different supports including bentonite, alumina, and silicon dioxide were explored. In this regard, adsorption of oxygen from nitrogen gas stream at the temperature of 150℃ and flow rate of 1400 mL min-1 was investigated. As shown in Fig. 4, Cu adsorbent on the supports of bentonite and alumina showed the same virtually performance, while the Cu adsorbent on silicon dioxide presented a
9
ACCEPTED MANUSCRIPT
considerably longer lifetime. This better performance might be due to higher surface area of the support. Hence, silicon dioxide was selected as the adsorbent support for next experiments.
3.6. The effect of particle size
RI PT
To investigate the effect of adsorbent particle size on the sorbent lifetime, 28 g Cu/SiO2 with different ranges of 10-30 and 30-50 mesh size under the same conditions mentioned in the section 3.5 were used. According to the obtained results (data not shown), the size of particles has no significant effect on adsorbent lifetime. Therefore, particles with mesh of 20-40, was used
SC
to eliminate/lower the possible back pressure at high gas flow rates.
M AN U
3.7. Optimization of amount of copper loaded on the adsorbent
Since the copper is active sorbent component, the copper amount in the adsorbent of Cu/SiO2 is a very important parameter which must be optimized. To that end, adsorbents containing 20, 30, 40, 50, 60, and 90 % of copper supported on silicon dioxide were prepared. Then, the lifetimes of these adsorbent were individually examined under the same conditions of temperature 150℃ and flow rate 1400 mL min-1. In these experiments, the lifetime was studied as the time that
TE D
oxygen concentration in nitrogen stream would be controlled below 1000 ppm. As shown in Fig. 5, adsorbent of Cu/SiO2 containing 50% Cu has the longest lifetime for oxygen removal from the nitrogen gas.
EP
3.8. Performance study of optimized adsorbent After preparation and optimization of oxygen sorbent based on Cu supported on silicon dioxide,
AC C
it was used for O2 removal from nitrogen gas used as a carrier gas for GC. Initially, the lifetime of optimized adsorbent at three temperatures of 50, 100, and 150℃ was monitored by NCD-IMS. Based on the obtained results, the temperature of 150℃ was selected for oxygen removal from the nitrogen gas containing about 4250 ppm O2, before introducing into the IMS. Fig. 6 depicts the ion mobility spectra recorded before and after utilizing the prepared O2 adsorbent. In this experiment, the gas flow rate was fixed at 1400 mL min-1. The oxygen concentration was calculated being below 1000 ppm even after 383 min passing through the sorbent. Considering the flow rate of carrier gas that is about 1 mL min-1 in GC, the lifetime of the sorbent would be about 372 days for O2 removal. 10
ACCEPTED MANUSCRIPT
4. Conclusions For the first time, in this research, negative corona discharge ionization ion mobility spectrometry was used for oxygen analysis in nitrogen gas flow and determination of
RI PT
performance of some synthesized oxygen scavenger compounds. The analytical parameters were exhaustively studied and the obtained results revealed the capability of the proposed method for the routine analysis of O2. In the event that problems originated by NOx ions were eliminated, a suitable linear dynamic range in addition to higher sensitivity and selectivity would be resulted
SC
by novel design of NCD-IMS for this application. After synthesis of different oxygen scavengers based on MnO and Cu metal, optimization of some parameters effective on adsorbent efficiency
M AN U
was performed.
Acknowledgments
The authors would like to acknowledge the Isfahan University of Technology (IUT) and Center of Excellence in Sensor and Green Chemistry for financial support of this work. Dr.
References:
TE D
Chiniforoushan is also specially acknowledged for his valuable assistance and discussions.
[1] K. Naito, T. Tsuji, T. Matsui, K. Une, J. Nucl. Sci. Technol. 11 (1974) pp. 22–28. [2] http://www.crawfordcientific.com/downloads/pdf new/Grace/Grace catalogue 2012.pdf.
EP
[3] K. Voldum-Clausen, Analyst 94 (1969) 500–502. [4] Q. Yan, C. Wan, J. Street, D. W. Yan, J. Han, F. Yu, Bioresour. Technol. 147 (2013) 117–123. [5] J. Han, X. He, R. Li, C. Wan, Q. Yan, F. Yu, J. Energy Inst. 87 (2014) 246–252.
AC C
[6] P. K. Sharma, P. K. Seshan, U. S. Patents 5219819 (1993). [7] D. C. Erickson, U. S. Patent 4287170 (1981). [8] H.E. Reinhold, K.S. Knaebel, M. Huber, D.R. King, U. S. Patent 5792239 (1998). [9] M. J. Mitariten, U.S. Patent 8221524 (2011). [10] F.M. Kirk, T.W. Weedon, P. Kirk, Instrumentation, American Technical Publishers, 2010 [11] N. G. Mandal, Anaesthesia & Intensive Care Medicine, 9 (2008) 559–563. [12] F.W. Karasek, W. D. Kilpatrick, M. J. Cohen, Anal. Chem. 43 (1971) 1441–1447. [13] G. A. Eiceman,; Z. Karpas, Ion Mobility Spectrometry, 2nd ed.; CRC Press: Boca Raton, FL, 2005. [14] M. Donegan, M. Browning, J. Liq. Chromatogr. R. T. 35 (2012) 2345–2363. 11
ACCEPTED MANUSCRIPT
[15] M. Tabrizchi, T. Khayamian, N. Taj, Rev. Sci. Instrum. 7 (2000) 2321–2328. [16] M. Tabrizchi, A. Abedi, Int. J. Mass spectrom. 218 (2002) 75–85. [17] R. G. Ewing, M.J.Waltman, Int. J. Ion Mobil. Spectrom. 12 (2009) 65–72. [18] A. Ebrahimi, M. T. Jafari, Talanta 134 (2015) 724–731.
[20] M.T. Jafari, Iranian Patent 79253 (2013).
RI PT
[19] M.T. Jafari, F. Riahi, J.Chromatogr. A 1343 (2014) 63–68.
[21] L. J. Lorenz, R. Culp, R. T. Dixon, Anal. Chem. 42 (1970) 1119–1120. [22] G. E. Spangler and P. A. Lawless, Anal. Chem. 50 (1978) 884–892.
[23] T. Khayamian, M. Tabrizchi, N. Taj, Fresenius J. Anal. Chem. 370 (2001) 1114–1116
SC
[24] L. Biernacki, S. Pokrzywnicki, J. Therm. Anal. Calorim. 55 (1999) 227–232.
[25] B. T. Collins, W. Desisto, R. Kershaw, K. Dvight, A. Wold, J. Less-Common Met. 156 (1989) 341–
M AN U
346.
[26] R. G. Ewing, D. A. Atkinson, G. A. Eiceman, G. J. Ewing, Talanta 54 (2001) 515–529.
AC C
EP
TE D
[27] T. Khayamian, M. Tabrizchi, M.T. Jafari, Talanta 59 (2003) 327–333.
12
ACCEPTED MANUSCRIPT
RI PT
Table 1
Typical operating conditions of NCD-IMS during the analysis of oxygen gas Operating parameters
Setting
Needle voltage
-11.2 kV
- 8.00 kV
SC
Target electrode voltage Drift field Drift gas flow (N2, 99.999%)
M AN U
Discharge gas flow (N2, 99.999%)
320 V cm-1 900 mL min-1 200 mL min-1
Carrier gas flow (N2, 99.999%)
540 mL min-1
Injector temperature
25 ℃
Cell temperature Drift tube length Shutter grid pulse
TE D
Number of IMS averages
AC C
EP
Number of points per ion mobility spectrum
13
150 ℃ 11 cm 0.2 ms 100 500
RI PT
ACCEPTED MANUSCRIPT
Table 2
Comparison of reduced mobility values obtained for some ionic species with those obtained previously by Reduced mobility, Ko, (cm2V-1s-1)
ionic species
Calculated in this work -
2.78 2.59
NO3-
2.52
NxOy-
2.39
2.83 2.61 2.56 –
AC C
EP
TE D
O2-
Reported by Ewing and Waltman [17]
M AN U
NO2
SC
Ewing and Waltman.
14
ACCEPTED MANUSCRIPT
Figure captions:
RI PT
Fig. 1: A) The background ion mobility spectrum of negative corona discharge IMS in pure nitrogen gas (99.999%) and B) The ion mobility spectrum of background in air (as drift/carrier gas in IMS) (blue) compared with that obtained after injection of 25 µL air, when pure nitrogen was used as drift/carrier gas (red). Fig. 2: The ion mobility spectra obtained after injection of NOx gases into the IMS with drift/carrier gases of pure nitrogen (upper) and air (below); the red spectra are obtained as the backgrouns and the blue ones are the spectra obtained after injection of NOx gases into the instrument under the same conditions of apparatus. Fig. 3: A) Signal-Time curve during the air dilution, B) The calibration curve obtained for
different oxygen amounts.
concentration was reached below 1000 ppm.
SC
Fig. 4: The lifetime of copper supported on silicon dioxide, alumina and bentonite. Fig. 5: The lifetime of Cu /SiO2 adsorbents with different amounts of copper, until the oxygen
M AN U
Fig. 6: The ion mobility spectra recorded red) before and blue) after utilizing the prepared O2
AC C
EP
TE D
adsorbent (Cu 50% on silicon dioxide). This figure shows a considerable removing the O2 from the nitrogen (containing 4250 ppm O2), when the prepared adsorbent was used.
15
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1
16
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2
17
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3
18
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 4
19
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5
20
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6
21
ACCEPTED MANUSCRIPT
Highlights:
AC C
EP
TE D
M AN U
SC
RI PT
Analysis of oxygen using negative corona discharge-ion mobility spectrometry was investigated for the first time. Novel designed point-in-cylinder geometry was used to establish the corona discharge without interferences of negative ions. The method was utilized to evaluate the performance of some synthesized oxygen scavengers.
S0003-2670(16)31413-1
DOI:
10.1016/j.aca.2016.12.001
Reference:
ACA 234934
To appear in:
Analytica Chimica Acta
Received Date: 11 September 2016 Revised Date:
26 November 2016
Accepted Date: 1 December 2016
Please cite this article as: K. Azadkish, M.T. Jafari, H.S. Ghaziaskar, Performance Evaluation of Oxygen Adsorbents Using Negative Corona Discharge–Ion Mobility Spectrometry, Analytica Chimica Acta (2017), doi: 10.1016/j.aca.2016.12.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Graphical Abstract
ACCEPTED MANUSCRIPT
Performance Evaluation of Oxygen Adsorbents Using Negative Corona Discharge–Ion Mobility Spectrometry Kamal Azadkish, Mohammad T. Jafari∗, Hassan S. Ghaziaskar
RI PT
Department of Chemistry, Isfahan University of Technology, Isfahan 84156–83111, Iran
Abstract
SC
Trace amounts of oxygen was determined using negative corona discharge as an ionization source for ion mobility spectrometry. A point-in-cylinder geometry with novel design was used -
M AN U
to establish the corona discharge without interferences of negative ions such as NOX . The desirable background spectrum shows only electrons peak, providing the instrument capable of trace analysis of oxygen in gaseous samples. The limit of detection and linear dynamic range with high coefficient of determination (r2=0.9997), were obtained for oxygen as 8.5 and 2814204 ppm, respectively. The relative standard deviations of the method for intraday and interday were obtained 4 and 11 %, respectively. The satisfactory results revealed the ability of
TE D
the negative corona discharge ion mobility spectrometry for investigating the performance of synthesized oxygen adsorbents in nitrogen streams. Two oxygen scavengers of MnO and Cu powder were prepared and the optimum temperature of the reactor containing MnO and Cu powder were obtained as 180 and 230 °C, respectively. Due to higher lifetime of copper powder,
EP
it was selected as the oxygen scavenger and some parameters such as: the type of adsorbent support, the size of adsorbent particles, and the amount of copper were studied for preparation of
AC C
more efficient oxygen adsorbent.
Keywords: Oxygen Analysis; Negative corona discharge; Ion mobility spectrometry; Oxygen adsorbents ∗
Corresponding author: Tel.: +98-31-3391-2351, Fax: +98-31-3391-2350
E-mail: [email protected].
1
ACCEPTED MANUSCRIPT
1. Introduction The determination of oxygen (O2) in different gases such as helium, hydrogen, nitrogen, argon, carbon dioxide, mixed gases, process gases, and ambient air is very important because oxygen content is a major factor influencing gas or fuel quality, internal corrosion, and user safety. In
RI PT
addition, the purification of inert gases is necessary for different applications such as metallurgical processes and chemical reactions that need to have a neutral environment, the nuclear and electronic fields, and laboratory and device applications [1]. For example, oxygen is the most detrimental contaminant for columns of gas chromatography (GC). Due to irreversible
SC
oxidation of stationary phases by oxygen, the destruction of chromatographic columns or reduction of their lifetime is occurred [2]. In the GC with electron-capture detector, the presence
M AN U
of oxygen impurity in the carrier gas, even at trace level, leads to reduce the sensitivity [3]. Some physical and chemical processes are used for purification and filtration of contaminants from gaseous streams. The physical processes include adsorption, refrigeration and low temperature distillation, and the chemical processes contain the gas-solid reactions and using the metallic getters and catalytic processes such as catalytic combination of oxygen and hydrogen [1]. Yan et al. have been studied oxygen catalytic removal from biomass-derived syngas by the Pt, Pd, and
TE D
Cu catalysts [4]. Han et al. applied catalytic oxidation of CuO–CeO2/Al2O3 catalyst for oxygen removal from syngas [5]. On the other hand, utilizing the adsorption/absorption techniques are suitable methods for O2-removal from inert gas streams. In the adsorption/absorption techniques, membrane processes, oxygen scavengers (e.g. MnO, Cu, and FeO) and oxygen acceptors (e.g.
EP
CuCl2, SrO, and Cr2O3) compounds are usually used for oxygen removal from a gas stream under special conditions [6-9].
AC C
So far, some common types of oxygen analyzers such as zirconium oxide analyzers, paramagnetic devices, electrochemical analyzers, and GC with thermal conductivity detector (GC-TCD) have been used for analysis of O2. However, these methods suffer from some drawbacks and limitations, so they cannot be kindly used for routine and real-time analysis of O2. Zirconium oxide analyzers need to operate at temperatures above 810 K. Therefore, the gas stream requires being at an elevated temperature or the instrument needs a sample heater. Therefore, it is usually used for measuring the excess oxygen in burners. Paramagnetic oxygen analyzers are very precise, but, because of its moving parts, they are very delicate. The positionand vibration-sensitive devices need to compensate and temperature control. In electrochemical 2
ACCEPTED MANUSCRIPT
oxygen analyzers, since the oxygen concentration is proportional to generated electrical current by reaction of oxygen as oxidizing agent and electrolyte, other oxidizing gases can react with electrolyte, resulting in poisoning the cell and give false readings [10,11]. Finally, in addition to long run time, the low sensitivity is problematic in the oxygen analysis by GC-TCD method.
RI PT
Ion mobility spectrometry (IMS) is an analytical technique based on measurement of the electrophoretic mobility of ions (affected by ion’s mass, charge, size and shape) through a neutral gas [12]. In this technique, different ions can be formed during an ionization process and then separated inside a drift tube under a constant electric field at ambient pressure. The
SC
application of IMS was traditionally limited to analysis of vapors or gaseous samples mainly of chemical warfare agents, drugs of abuse, and explosives [13]. Contemporary IMS research has
M AN U
expanded enormously to the analysis of gaseous, liquid, and solid samples in many fields including biology, medicine, environmental studies, forensics, pharmaceuticals, and food research. The popularity of IMS as an analytical technique is due to its excellent figures-of-merit including low limits of detection (pg to µg), fast separation times (µs to ms), low cost for handheld or standalone devices, and high throughput (seconds per sample) [13]. Ionization source is one of the key parts in IMS instrument so that a variety of ionization sources have been
TE D
applied for this purpose. Two popular ionization sources of
63
Ni and corona discharge have
mostly been used for analytes in gaseous phase [13, 14]. In addition to non-radioactivity, the corona discharge dedicates higher sensitivity (~ 10 times) related to
63
Ni [15]. However,
application of corona discharge in negative mode has not been expanded due to serious problems
EP
originated form NOx formation in the source. NOx molecules have the highest electron affinity (~ 3.9 eV) among chemical compounds and thus they essentially reduce the sensitivity of IMS
AC C
for other analytes [16]. In addition, analysis of some analytes such as O2, SOx, COx, or halogen gases has not been possible in the presence of NOx species due to overlapping of their signals [17]. In our previous work [18], a novel design of point-in-cylinder geometry was used to -
establish the negative corona discharge without interferences of negative ions such as NOx , for analysis of Ethion as an organophosphorus pesticide. The background ion mobility spectrum in pure nitrogen containing only electrons peak was obtained, promising the possibility of O2 analysis by the instrument. In this work, for the first time the capability of negative corona discharge ion mobility spectrometry (NCD–IMS) for oxygen measurement in nitrogen gas stream has been evaluated. 3
ACCEPTED MANUSCRIPT
The background ion mobility spectra of instrument in two gases of nitrogen and air were obtained and lack of NOx species production was demonstrated. The proposed method for O2 analysis was applied to evaluate the performance of some synthesized oxygen adsorbents.
RI PT
2. Experimental 2.1. Materials
Materials such as silicon dioxide, alumina, and manganese (II) chloride monohydrate were
SC
obtained from Merck Co., Germany. Sodium carbonate, bentonite, copper (II) nitrate trihydrate, were purchased from Sigma-Aldrich Co., USA. To prepare all solutions through entire work four-time distilled water was used. The commercial grade 2,4,6-trinitrotoluene (TNT) was
M AN U
provided from Chemical Industries, Zarrinshahr, Isfahan, Iran and used as the internal standard in IMS.
2.2. Instrumentation
The ion mobility spectrometer used for this research was designed and constructed at Isfahan University of Technology and described previously in detail [18, 19]. The schematic diagram of
TE D
the IMS cell used in this work is shown in Fig. S1, in the Supporting Information. In brief, the IMS instrument included a cell containing the corona discharge needle, two high voltage power supplies, a pulse generator, an analog to digital converter, and a computer. The IMS cell consisted from 15 aluminum rings (8 mm thickness, i.d. 40 and 25 mm for drift and reaction
EP
regions, respectively) separated by PTFE insulators with 1 mm thickness. The length of drift and reaction regions was 11 and 5 cm, respectively and separated from each other by Bradbury-
AC C
Nielsen shutter grid. The detail of corona discharge ionization source without formation of -
negative ions of NOX was presented in an Iranian patent [20]. In this geometry, a sharp needle (0.1 mm diameter) was located inside an aluminum ring (i.d. 8 mm) which used as a cylindrical counter electrode. To remove water vapor or other contaminants, the gas was filtered using a 4Åmolecular sieve (Merck) in a 40 cm length × 25 mm i.d. cylindrical trap, before entering into the IMS cell. Table 1 summarizes all the operation conditions of NCD–IMS under which the mobility spectra were taken.
4
ACCEPTED MANUSCRIPT
2.3. Preparation of different concentrations of oxygen in nitrogen In order to draw the calibration curve, the exponential dilution flask method was used for preparation of different concentrations of oxygen. The usefulness of this method for calibration of gas chromatographic detectors and preparation of dilute gas samples has been discussed
RI PT
elsewhere [21]. In addition, this method has been successfully used for introducing the gas samples into the IMS [22, 23]. Generally, in this research purified air (20% O2 and 80% N2) was trapped in a flask with known volume (1225 mL) and continuously flushed out with the exact flow rate of pure nitrogen (540 mL min-1), as carrier gas. To detect the signal variation related to
SC
changing of oxygen concentration, flask outlet was directly connected to the IMS cell. In this method, the outlet oxygen concentration (C) after time (t) is given by Eq. (1).
M AN U
C = C0 exp (−Qt/V)
(1)
Where C0 is the initial concentration of the oxygen that was equivalent to 20% or 200000 ppm in air, Q is the gas flow rate and V is the flask volume.
2.4. Preparation of oxygen scavenger compounds
TE D
To investigate the performance of oxygen adsorbents for removal of oxygen from nitrogen gas, suitable oxygen scavengers such as MnO and Cu powder were prepared.
2.4.1. MnO Preparation
EP
To prepare the MnO adsorbent according to Eq. (2), appropriate amounts of MnCl2 (H2O) and Na2CO3 with proper stoichiometric ratios were individually dissolved in minimum amount of
AC C
water. The solution of Na2CO3 was slowly added to the solution of MnCl2 to form MnCO3. The precipitate was separated using Buchner funnel under vacuum, and kept in 50 ℃ for 10 h to dryness. The dried MnCO3 was transferred into a tubular reactor (316 stainless steel tube, 4 mm i.d. and 50 mm length) and heated at 400 ℃ for 4 h while passing pure argon gas (99.9995%). In this state, MnCO3 decomposed to MnO and CO2, according to equation (3) [24]. MnCl2 (aq) + Na2CO3 (aq) MnCO3 (s) Δ Ar
MnCO3 (s) + 2 NaCl (aq)
MnO (s) + CO2 (g)
(2) (3)
5
ACCEPTED MANUSCRIPT
2.4.2. Cu powder Preparation For preparation of Cu powder, the copper oxide was prepared according to equation (4) using calcination of copper (II) nitrate at temperature of 500 ℃ for 5 h [25]. Afterward, pure hydrogen
RI PT
gas was passed through the CuO powder with flow rate of 10-20 mL min-1 at temperature of 200 ℃. At this step, Cu powder would be prepared according to Eq. (5) [3]. 2 Cu(NO3)2 → 2 CuO + 4 NO2 + O2
(5)
SC
CuO + H2 → Cu + H2O
(4)
2.5. Cu adsorbents on different supports
To prevent agglomeration of copper powder and to increase the surface area and consequently its
M AN U
performance at lower temperatures, copper adsorbents was dispersed on different supports such as silicon dioxide, bentonite, and alumina, via the impregnation method.
2.5.1. Preparation of Cu adsorbent supported on silicon dioxide
For the preparation of Cu adsorbent supported on silicon dioxide with 40 w% Cu, 45.62 g
TE D
Cu(NO3)2.3H2O was dissolved in water and 18 g silicon dioxide was added to the solution. The solution was stirred at room temperature overnight. After impregnation, water was removed by vacuum drying at 80℃ in a rotary evaporator. The obtained powder was kept in 120 ℃ to dryness. The dried powder was calcinated in air at 500 ℃ for 5h. Finally, the adsorbent was
EP
prepared after passing pure hydrogen gas through the calcinated powder at flow rate of 10-20 mL
AC C
min-1 and temperature of 200 ℃.
2.5.2. Preparation of Cu adsorbent supported on bentonite and alumina Preparation of Cu adsorbent supported on bentonite and alumina (40 w%, Cu content) was accomplished by addition of 18.0 and 19.8 g of alumina and bentonite, to Cu(NO3)2.3H2O solutions containing 12 g Cu. Other steps were carried out similar to section 2.5.1. It is notable that the bentonite used in this work was sodium-type and its loss on ignition was 10%. So we ought to use 19.8 g to achieve 18 g bentonite after calcination.
3. Results and discussion 6
ACCEPTED MANUSCRIPT
3.1. Ion mobility spectra The background ion mobility spectrum of NCD-IMS with pure nitrogen (99.999% and used as drift and carrier gas) is shown in Fig. 1-A. For more clarification of signal to noise ratio, the background amplified 200 times which is shown as inset of the figure. The signal at drift time 6.3
RI PT
ms is related to the trace amount of oxygen impurity in the nitrogen gas. The spectrum shows a very large peak at zero drift time corresponding to electrons. As this figure obviously shows, no -
significant negative ions (e.g. NOx ) were formed during the negative corona discharge by using -
the novel design. Lack of NOx production has some major benefits, which are very important in
SC
both qualitative and quantitative analysis by NCD-IMS. Firstly, high electron affinity of these ions can prevent sufficient ionization of other substances even halogenated compounds [16],
M AN U
reducing the sensitivity of the device, extremely. Secondly, it is impossible to analyze NOX, themselves or SOx, COx, O2 and other chemical compounds with the same drift time as background ions [17]. Finally, NOX ions can form coordination complexes with interest analytes in some cases, resulting in complex spectra [26]. Consequently, very high signal to noise ratio was expected for the oxygen determination when NOx production was inhibited during the process of negative corona discharge.
TE D
Additionally, Fig. 1-B depicts the ion mobility spectrum of background in purified air (as drift/carrier gas in IMS) compared with that obtained after injection of 25 µL purified air, when pure nitrogen was used as drift/carrier gas. As this figure shows, when air was injected into the IMS with N2 drift gas, only one ion peak was originated by ionization in negative corona
EP
discharge. In order to calculate the reduced mobility value (Ko) of this negative ion, 2,4,6trinitrotoluene (TNT) with the Ko=1.55 cm2V-1s-1 [27], was used as standard compound. The Ko
AC C
value was calculated to be 2.59 cm2V-1s-1 which is well compared with 2.61 cm2V-1s-1, reported -
by Ewing and Waltman [17] for O2 ions. In order to show more compelling reasons for this claim, NOx gases, produced by pouring the copper powder in nitric acid solution, were injected into the IMS. Fig. 2 shows the ion mobility spectra obtained after injection of NOx gases, in two cases of drift/carrier gases of pure nitrogen (upper) and purified air (below). These spectra are compared with those of their corresponded backgrounds. To assign the originated ions in these spectra, the reduced mobility values were calculated and tabulated in Table 2. The reduced -
-
-
mobility values reported previously by Ewing and Waltman [17] for NO2 , NO3 , and O2 ions,
7
ACCEPTED MANUSCRIPT
are also shown in this Table. By comparison of these results, it can be concluded that the one ion -
-
peaks produced by air ionization is only O2 and none NOx ions. When we used pure nitrogen -
instead of air, the single ion peak of O2 was also disappeared in the mobility background and
RI PT
only the electron signal would be observed in the spectrum. These favorable results reveal the capability of NCD-IMS with pure nitrogen for qualitative and quantitative analysis of oxygen.
3.2. Analytical performance
As mentioned above, the exponential dilution flask method was used to prepare different oxygen
SC
concentration. The IMS signal originated from O2 was followed at the different times of air dilution process and the obtained results are shown in Fig. 3-A. The analyte concentrations were
M AN U
calculated using the equation (1) at different times and the linear dynamic range was prepared by plotting the IMS signals versus their corresponded O2 amount. The calibration curve (Fig. 3-B) shows an acceptable coefficient of determination (r2=0.9997) for oxygen in the linear dynamic range of 28-14204 ppm. The inset figure was also prepared in order to see better signals of instrument for low concentrations of oxygen (< 850 ppm). The limit of detection defined by signal–to–noise ratio of 3 was calculated to be 8.5 ppm. The relative standard deviations (RSD)
TE D
were achieved, 4 and 11% for one day (n=3) and over a period of three days (n=3), respectively. The favorable analytical results obtained for the analysis of oxygen in this method encourage us to investigate the performance of some synthesized oxygen sorbents, described in following sections.
EP
For more evaluation of the method, the relative errors of the method were calculated for some preapred gaseous solutions entered into the IMS cell. To that end, the solutions with
AC C
various oxygen concentrations of 1105, 5054, and 8171 ppm were prepared according to equation (5).
(5)
Ca = 20 ×
In the equation Ca is the prepared oxygen concentration in percent, Q2 and Q1 are the air and nitrogen gases flow rate, respectively. The IMS-signals were calculated and compared with those obtained by calibration curve. The relative error values were obtained 8.1, 3.3, and 2.2 %, for concentrations of 1105, 5054, and 8171 ppm, respectively. These satisfactory results insure the qualitative analysis of oxygen in different gaseous samples. 8
ACCEPTED MANUSCRIPT
3.3. Temperature optimization for O2 adsorption by Cu powder and MnO One of the most fundamental variables for a chemical reaction is the temperature. To evaluate the temperature effect on oxygen removal from nitrogen gas using Cu powder and MnO,
RI PT
temperatures of reactor containing 35 g of each sorbents were indivitualy changed, in the range of 25–250 ℃. For investigation of temperature effect, oxygen-contaminated nitrogen with flow rate of 1400 mL min-1 was passed through the reactor and then injected into the IMS instrument. According to the results shown in Fig. S2 in the Supporting Information, the signal of O2 was
SC
reduced while the reaction temperature increased to 180 and 230 ℃ for MnO and Cu powder, respectively. Therefore, these optimized temperatures were considered for study of the sorbents
M AN U
lifetime.
3.4. Relative lifetime of MnO and Cu powder in stream
For exploration of lifetime of oxygen scavenger compunds, nitrogen gas containing oxygen was passed through the reactor filled with 35 g of each compounds. The temperature was set at 180 and 230 ℃ for MnO and Cu powder, respectively and the flow rate of gas was 1400 mL min-1
TE D
during the adsorption of oxygen. The IMS signals were recorded while Cu powder and MnO were removing oxygen from the nitrogen stream. Herein, the liftime is defined as the time taken to saturate the adsoebent and hence, the O2 signal becomes equal to its initial intensity (before passing the gas through the adsorbents). The results are shown in Fig. S3 (in the Supporting
EP
Information) indicated that relatively copper powder had longer liftime (about 2 times), and
AC C
hence, it was chosen as more suitable oxygen scavenger and used for successive experiments.
3.5. The effect of support type After selection of copper powder, it was necessary to study the effect of support type on the adsorption efficiency. So, the lifetimes of 28 g of copper adsorbent on the different supports including bentonite, alumina, and silicon dioxide were explored. In this regard, adsorption of oxygen from nitrogen gas stream at the temperature of 150℃ and flow rate of 1400 mL min-1 was investigated. As shown in Fig. 4, Cu adsorbent on the supports of bentonite and alumina showed the same virtually performance, while the Cu adsorbent on silicon dioxide presented a
9
ACCEPTED MANUSCRIPT
considerably longer lifetime. This better performance might be due to higher surface area of the support. Hence, silicon dioxide was selected as the adsorbent support for next experiments.
3.6. The effect of particle size
RI PT
To investigate the effect of adsorbent particle size on the sorbent lifetime, 28 g Cu/SiO2 with different ranges of 10-30 and 30-50 mesh size under the same conditions mentioned in the section 3.5 were used. According to the obtained results (data not shown), the size of particles has no significant effect on adsorbent lifetime. Therefore, particles with mesh of 20-40, was used
SC
to eliminate/lower the possible back pressure at high gas flow rates.
M AN U
3.7. Optimization of amount of copper loaded on the adsorbent
Since the copper is active sorbent component, the copper amount in the adsorbent of Cu/SiO2 is a very important parameter which must be optimized. To that end, adsorbents containing 20, 30, 40, 50, 60, and 90 % of copper supported on silicon dioxide were prepared. Then, the lifetimes of these adsorbent were individually examined under the same conditions of temperature 150℃ and flow rate 1400 mL min-1. In these experiments, the lifetime was studied as the time that
TE D
oxygen concentration in nitrogen stream would be controlled below 1000 ppm. As shown in Fig. 5, adsorbent of Cu/SiO2 containing 50% Cu has the longest lifetime for oxygen removal from the nitrogen gas.
EP
3.8. Performance study of optimized adsorbent After preparation and optimization of oxygen sorbent based on Cu supported on silicon dioxide,
AC C
it was used for O2 removal from nitrogen gas used as a carrier gas for GC. Initially, the lifetime of optimized adsorbent at three temperatures of 50, 100, and 150℃ was monitored by NCD-IMS. Based on the obtained results, the temperature of 150℃ was selected for oxygen removal from the nitrogen gas containing about 4250 ppm O2, before introducing into the IMS. Fig. 6 depicts the ion mobility spectra recorded before and after utilizing the prepared O2 adsorbent. In this experiment, the gas flow rate was fixed at 1400 mL min-1. The oxygen concentration was calculated being below 1000 ppm even after 383 min passing through the sorbent. Considering the flow rate of carrier gas that is about 1 mL min-1 in GC, the lifetime of the sorbent would be about 372 days for O2 removal. 10
ACCEPTED MANUSCRIPT
4. Conclusions For the first time, in this research, negative corona discharge ionization ion mobility spectrometry was used for oxygen analysis in nitrogen gas flow and determination of
RI PT
performance of some synthesized oxygen scavenger compounds. The analytical parameters were exhaustively studied and the obtained results revealed the capability of the proposed method for the routine analysis of O2. In the event that problems originated by NOx ions were eliminated, a suitable linear dynamic range in addition to higher sensitivity and selectivity would be resulted
SC
by novel design of NCD-IMS for this application. After synthesis of different oxygen scavengers based on MnO and Cu metal, optimization of some parameters effective on adsorbent efficiency
M AN U
was performed.
Acknowledgments
The authors would like to acknowledge the Isfahan University of Technology (IUT) and Center of Excellence in Sensor and Green Chemistry for financial support of this work. Dr.
References:
TE D
Chiniforoushan is also specially acknowledged for his valuable assistance and discussions.
[1] K. Naito, T. Tsuji, T. Matsui, K. Une, J. Nucl. Sci. Technol. 11 (1974) pp. 22–28. [2] http://www.crawfordcientific.com/downloads/pdf new/Grace/Grace catalogue 2012.pdf.
EP
[3] K. Voldum-Clausen, Analyst 94 (1969) 500–502. [4] Q. Yan, C. Wan, J. Street, D. W. Yan, J. Han, F. Yu, Bioresour. Technol. 147 (2013) 117–123. [5] J. Han, X. He, R. Li, C. Wan, Q. Yan, F. Yu, J. Energy Inst. 87 (2014) 246–252.
AC C
[6] P. K. Sharma, P. K. Seshan, U. S. Patents 5219819 (1993). [7] D. C. Erickson, U. S. Patent 4287170 (1981). [8] H.E. Reinhold, K.S. Knaebel, M. Huber, D.R. King, U. S. Patent 5792239 (1998). [9] M. J. Mitariten, U.S. Patent 8221524 (2011). [10] F.M. Kirk, T.W. Weedon, P. Kirk, Instrumentation, American Technical Publishers, 2010 [11] N. G. Mandal, Anaesthesia & Intensive Care Medicine, 9 (2008) 559–563. [12] F.W. Karasek, W. D. Kilpatrick, M. J. Cohen, Anal. Chem. 43 (1971) 1441–1447. [13] G. A. Eiceman,; Z. Karpas, Ion Mobility Spectrometry, 2nd ed.; CRC Press: Boca Raton, FL, 2005. [14] M. Donegan, M. Browning, J. Liq. Chromatogr. R. T. 35 (2012) 2345–2363. 11
ACCEPTED MANUSCRIPT
[15] M. Tabrizchi, T. Khayamian, N. Taj, Rev. Sci. Instrum. 7 (2000) 2321–2328. [16] M. Tabrizchi, A. Abedi, Int. J. Mass spectrom. 218 (2002) 75–85. [17] R. G. Ewing, M.J.Waltman, Int. J. Ion Mobil. Spectrom. 12 (2009) 65–72. [18] A. Ebrahimi, M. T. Jafari, Talanta 134 (2015) 724–731.
[20] M.T. Jafari, Iranian Patent 79253 (2013).
RI PT
[19] M.T. Jafari, F. Riahi, J.Chromatogr. A 1343 (2014) 63–68.
[21] L. J. Lorenz, R. Culp, R. T. Dixon, Anal. Chem. 42 (1970) 1119–1120. [22] G. E. Spangler and P. A. Lawless, Anal. Chem. 50 (1978) 884–892.
[23] T. Khayamian, M. Tabrizchi, N. Taj, Fresenius J. Anal. Chem. 370 (2001) 1114–1116
SC
[24] L. Biernacki, S. Pokrzywnicki, J. Therm. Anal. Calorim. 55 (1999) 227–232.
[25] B. T. Collins, W. Desisto, R. Kershaw, K. Dvight, A. Wold, J. Less-Common Met. 156 (1989) 341–
M AN U
346.
[26] R. G. Ewing, D. A. Atkinson, G. A. Eiceman, G. J. Ewing, Talanta 54 (2001) 515–529.
AC C
EP
TE D
[27] T. Khayamian, M. Tabrizchi, M.T. Jafari, Talanta 59 (2003) 327–333.
12
ACCEPTED MANUSCRIPT
RI PT
Table 1
Typical operating conditions of NCD-IMS during the analysis of oxygen gas Operating parameters
Setting
Needle voltage
-11.2 kV
- 8.00 kV
SC
Target electrode voltage Drift field Drift gas flow (N2, 99.999%)
M AN U
Discharge gas flow (N2, 99.999%)
320 V cm-1 900 mL min-1 200 mL min-1
Carrier gas flow (N2, 99.999%)
540 mL min-1
Injector temperature
25 ℃
Cell temperature Drift tube length Shutter grid pulse
TE D
Number of IMS averages
AC C
EP
Number of points per ion mobility spectrum
13
150 ℃ 11 cm 0.2 ms 100 500
RI PT
ACCEPTED MANUSCRIPT
Table 2
Comparison of reduced mobility values obtained for some ionic species with those obtained previously by Reduced mobility, Ko, (cm2V-1s-1)
ionic species
Calculated in this work -
2.78 2.59
NO3-
2.52
NxOy-
2.39
2.83 2.61 2.56 –
AC C
EP
TE D
O2-
Reported by Ewing and Waltman [17]
M AN U
NO2
SC
Ewing and Waltman.
14
ACCEPTED MANUSCRIPT
Figure captions:
RI PT
Fig. 1: A) The background ion mobility spectrum of negative corona discharge IMS in pure nitrogen gas (99.999%) and B) The ion mobility spectrum of background in air (as drift/carrier gas in IMS) (blue) compared with that obtained after injection of 25 µL air, when pure nitrogen was used as drift/carrier gas (red). Fig. 2: The ion mobility spectra obtained after injection of NOx gases into the IMS with drift/carrier gases of pure nitrogen (upper) and air (below); the red spectra are obtained as the backgrouns and the blue ones are the spectra obtained after injection of NOx gases into the instrument under the same conditions of apparatus. Fig. 3: A) Signal-Time curve during the air dilution, B) The calibration curve obtained for
different oxygen amounts.
concentration was reached below 1000 ppm.
SC
Fig. 4: The lifetime of copper supported on silicon dioxide, alumina and bentonite. Fig. 5: The lifetime of Cu /SiO2 adsorbents with different amounts of copper, until the oxygen
M AN U
Fig. 6: The ion mobility spectra recorded red) before and blue) after utilizing the prepared O2
AC C
EP
TE D
adsorbent (Cu 50% on silicon dioxide). This figure shows a considerable removing the O2 from the nitrogen (containing 4250 ppm O2), when the prepared adsorbent was used.
15
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1
16
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2
17
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3
18
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 4
19
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5
20
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6
21
ACCEPTED MANUSCRIPT
Highlights:
AC C
EP
TE D
M AN U
SC
RI PT
Analysis of oxygen using negative corona discharge-ion mobility spectrometry was investigated for the first time. Novel designed point-in-cylinder geometry was used to establish the corona discharge without interferences of negative ions. The method was utilized to evaluate the performance of some synthesized oxygen scavengers.