- Email: [email protected]
Vapor permeation-stepwise injection simultaneous determination of methanol and ethanol in biodiesel with voltammetric detection
Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Vapor permeation-stepwise injection simultaneous determination of methanol and ethanol in biodiesel with voltammetric detection Andrey Shishov n, Anastasia Penkova, Andrey Zabrodin, Konstantin Nikolaev, Maria Dmitrenko, Sergey Ermakov, Andrey Bulatov Institute of Chemistry, St. Petersburg State University, Universitetsky pr. 26, Saint Petersburg 198504, Russia
art ic l e i nf o
a b s t r a c t
Article history: Received 28 February 2015 Received in revised form 12 May 2015 Accepted 18 May 2015
A novel vapor permeation-stepwise injection (VP-SWI) method for the determination of methanol and ethanol in biodiesel samples is discussed. In the current study, stepwise injection analysis was successfully combined with voltammetric detection and vapor permeation. This method is based on the separation of methanol and ethanol from a sample using a vapor permeation module (VPM) with a selective polymer membrane based on poly(phenylene isophtalamide) (PA) containing high amounts of a residual solvent. After the evaporation into the headspace of the VPM, methanol and ethanol were transported, by gas bubbling, through a PA membrane to a mixing chamber equipped with a voltammetric detector. Ethanol was selectively detected at þ0.19 V, and both compounds were detected at þ1.20 V. Current subtractions (using a correction factor) were used for the selective determination of methanol. A linear range between 0.05 and 0.5% (m/m) was established for each analyte. The limits of detection were estimated at 0.02% (m/m) for ethanol and methanol. The sample throughput was 5 samples h 1. The method was successfully applied to the analysis of biodiesel samples. & 2015 Published by Elsevier B.V.
Keywords: Methanol Ethanol Biodiesel Vapor permeation Flow analysis Cyclic voltammetry
1. Introduction Biodiesel occupies a prominent position among alternatives to conventional petrodiesel fuel due to various technical, economic and ecological factors. The technology of biodiesel production includes the transesterification process of vegetable oil (e.g., soy oil) or animal fats (e.g., swine lard) with methanol or ethanol in the presence of a catalyst (alkali, acid or an enzyme) [1–3]. After the transesterification reaction, excess alcohol is removed from the biodiesel by extraction and distillation. The amount of alcohols in biodiesel is regulated by American (ASTM D 6751) and European (EN 14110) biodiesel standards and the permitted maximum level of alcohols in biodiesel samples is 0.2% (m/m). Excess alcohols cause metal corrosion, particularly aluminum corrosion; decrease diesel fuel flash point; and degrade rubber and polymer parts of engines. Various analytical methods have been developed (Table 1) for the determination of alcohols in biodiesel. They are based on gas chromatography (GC) [4–6], near infrared (NIR) [7–11] and UV–vis [10] spectroscopy. GC is the most used technique due to its high accuracy in the quantification of minor components. However, n
Corresponding author. E-mail address: [email protected] (A. Shishov).
baseline drift and overlapping signals can have a detrimental effect on GC accuracy [12]. The GC accuracy can be improved by using a headspace solid-phase microextraction [6]. This technique is characterized by high sensitivity, good reproducibility and recovery. The application of chemometric tools in the NIR spectroscopy allows the determination of methanol and water in biodiesel [8]. NIR and visible spectroscopy are able to predict methanol and glycerol traces in biodiesel samples [10]. Moreover, the technique for the determination of residual alcohol content in biodiesel through determination of its flash point was proposed [13], but this method does not allow the simultaneous determination of methanol and ethanol. In the present research [14], a cyclic voltammetry method was developed for the simultaneous determination of analytes in fuel ethanol. Automation of analysis is an important and rapidly growing trend in modern analytical chemistry. Recently, the automation of analytical procedures based on flow analysis has been developed. Therefore, labor costs for the analysis are decreased, and the volume of sample solutions and reagents are reduced. Currently only one flow method for the determination of methanol in biodiesel samples is described [15]. The proposed method is based on the liquid–liquid extraction of methanol from a sample solution to a phosphate buffer using a membrane unit with a polyvinylidene fluoride membrane. The determination of
http://dx.doi.org/10.1016/j.talanta.2015.05.041 0039-9140/& 2015 Published by Elsevier B.V.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
Table 1 Comparison of the suggested method with previously reported for the determination of methanol and ethanol in biodiesel. Detection technique Analyte
Sample preparation
Sample amount
On-line analysis
Linear range
LOD
Ref.
GC GC
MeOH MeOH
100 μL 1 mL
No No
4 6
MeOH MeOH
– 10 μL
No No
51 ppm –
9 10
FP CV
EtOH, MeOH – EtOH, MeOH Dilution
65 mL –
No No
– 0.005–0.05% (m/ m) – 0.003–0.433% (m/m) to 1% (v/v) 0.1–0.5% (v/v)
0.001% (m/m) –
NIR VIS and NIR
Dilution Headspace solid phase microextraction – –
13 14
VIS
MeOH
200 μL
Yes
0.1% (v/v) 0.028 and 0.045% (v/v) for EtOH, MeOH 0.0002% (m/m)
15
CV
EtOH, MeOH Vapor permeation
1 mL
Yes
0.02% (m/m)
This work
Membrane extraction
0.001–0.200% (m/m) 0.05–0.5% (m/m)
GC – gas chromatography, NIR – near infrared spectroscopy, VIS – visible spectroscopy, FP – flash point determination, CV – cyclic voltammetry.
methanol was then achieved in aqueous solution by converting it to hydrogen peroxide through the use of alcohol oxidase, followed by the use of 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) as an electron donor for horseradish peroxidase, giving the colored ABTS radical cation. In this work, a new flow method for the simultaneous determination of methanol and ethanol in different types of biodiesel with analytes vapor permeation through the PА membrane and cyclic voltammetry detection has been developed. Vapor permeation (VP) is frequently used for the separation of volatile analytes in complex samples [16] and can be coupled with flow methods. The PA was chosen as the membrane material because of its good physico-chemical properties and high selectivity with respect to methanol in pervaporation of tert-butyl methyl ether/ methanol and cyclohexane/methanol mixtures, as seen in previous work [17,18]. It should be noted that the PA membrane studied in this work is different from the PA membrane described in previous works [17–19]. The studied membrane has higher free volume, which significantly effects the transport of low molecular substances through it. The stepwise injection analysis [20–22] was chosen for the automation of methanol and ethanol determination in biodiesel because its manifold allows to mix solutions by a gas
flow. This feature can be used to increase the efficiency of methanol and ethanol penetration during vapor permeation.
2. Experimental 2.1. Manifold and apparatus A stepwise injection manifold (Fig. 1) includes a syringe pump (Sciware systems, Spain), a reversible peristaltic pump MasterFlex L/S (Cole-Parmer, USA) with modified PVC pumping tube (WatsonMarlow, Russia), two six-way solenoid valves (Cole-Parmer Inc., USA), a homemade mixing chamber (MC) (a cylindrical-shaped PTFE tube (20 mm in height and 15 mm in i.d.) equipped with a magnetic stirrer and integrated with the miniaturized home-made Ag/AgCl reference, platinum auxiliary and gold working (ؼ3 mm) electrodes), 797 VA Computrace Analizer (Metrohm, Switzerland), communication tubes (PTFE, 0.5 mm i.d.) and a laboratory-made vapor permeation module. The vapor permeation module (Fig. 1) was constructed from two titanium discs (i.d. 50 mm), held together by four stainless steel bolts. The depths of the chambers, separated by the PA membrane (35 mm diameter
Fig. 1. The VP-SWI manifold for the simultaneous determination of methanol and ethanol in biodiesel samples.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
and 40 mm thickness), were 25 mm. A water-bath, equipped with a thermostat, was used to preserve the temperature of the vapor permeation module. The manifold was operated automatically by means of a computer. The GC GC-MS-QP2010 Ultra gas chromatograph mass spectrometer (Shimadzu, Japan) was used for the determination of analytes and identification of fatty acid methyl esters (FAME) in biodiesel samples. The TG 209 F3 Iris thermo-microbalance (Netzsch, Germany) was used for thermal degradation measurements. Scanning electron microscopy (SEM) micrographs were obtained by a Merlin scanning electron microscope (Zeiss, Germany). 2.2. Reagents and samples Analytical grade chemicals were used throughout the work. 1% (m/m) stock solutions of methanol and ethanol were prepared by reagents (Sigma-Aldrich, USA) dissolving in biodiesel. The biodiesel (B 20) samples were obtained by mixing biofuel and diesel in a ratio of 1:4, respectively. Biofuels were obtained from the following feedstocks: sunflower, canola, corn and palm. These samples had been previously prepared with the use of sodium hydroxide as an alkaline catalyst [23]. To prevent the presence of alcohols, the biofuel samples were washed with water and dried for 2 hours at a temperature of 80 °C. From the GC–MS analysis [24], it was observed that the biofuel samples contained methyl and ethyl esters of n-nonanoic, n-hexadecanoic, n-octadecanoic, 9-octadecenoic and 9,12-octadecadienoic acids. Electrolyte solution (0.5 M NaOH and 0.1% (v/v) CH3OH) was prepared by NaOH (Sigma-Aldrich, USA) and CH3OH dissolving in deionized water (18.2 MΩ cm, Millipore, USA). 2.3. Membrane Poly(phenylene isophtalamide) (commercial sample Fenylons, Minkar, Russia) was used for membrane preparation. The PA membrane (thickness 40 mm) was obtained by casting 5% PA and 0.7% LiCl (Sigma-Aldrich, USA) solution in dimethylacetamide (DMAc) on a glass support with subsequent drying at 40 °C for 2 weeks. 2.4. Procedures 2.4.1. The VP-SWI procedure Initially (Fig. 1), 0.25 mL of electrolyte solution (channel 1) was aspirated through a valve 1 into the holding coil and then into the mixing chamber (channel 3) by the syringe pump. Then 1 mL of the sample (channel 7) was aspirated through the valve 2 into the donor part of VPM by the reversible peristaltic pump using the modified PVC pumping tube. The lifetime of the pumping tube was one week, and after every 50 analyses, it was replaced. The sample did not make contact with the membrane surface. The VPM was thermostated at 70 °C and air (channel 8) was transferred into the VPM by means of the peristaltic pump for 10 min at a rate of 3 mL min 1 through the valve 2 to mix sample and intensification of analytes' evaporation. After evaporation into the headspace of the VPM, methanol and ethanol were transported through the PA membrane by gas bubbling into the MC equipped with the magnetic stirrer and electrodes. The VPM was connected with the MC by means of PTFE tube. After the vapor permeation process and dilution in the electrolyte solution, the concentration of alcohols was measured by cyclic voltammetry. Ethanol was selectively detected at þ0.19 V and both analytes were detected at þ1.20 V. The correction factor was used for the selective determination of methanol. Then solution was then transfered from the MC to waste
3
(channel 4) and the MC was washed with electrolyte solution (channel 1). Finally, the sample was transfered from the donor part of VPM to waste (channel 10) by means of the reversible peristaltic pump and the system was washed with isooctane (channel 9). All parameters of the flow system were controlled by a PC. 2.4.2. The GC procedure for determination of alcohols 0.5 mL of each biodiesel sample was shaken with 5 mL of water for half an hour [25]. Then the solution was centrifuged, and the content of alcohols in the water phase was measured by GC. The GC was fitted with a SPB-624 MS capillary column (Stabilwax polar phase, Crossbondspolyethylene glycol, 30 m, 0.32 mm i.d., 0.5 μm coating). The injector temperature was kept at 250 °C and temperature of the column at 40 °C. Helium was used as a carrier gas at a flow rate of 1.5 mL min 1. A sample volume of 10 μL was injected in splitless mode with 1 min of purge time. 2.4.3. The GC–MS procedure for assessing FAME in biofuel FAME in biofuel was estimated using gas chromatograph mass spectrometry [24]. The GC was fitted with a SPB-624 MS capillary column (6% cyanopropyl phenyl and 94% dimethyl polysiloxane, 30 m, 0.25 mm i.d., 0.25 μm coating). The injector temperature was kept at 220 °C while the interface temperature and the ion source were 200 °C. Helium was used as a carrier gas at a flow rate of 2 mL min 1. The sample volume of 2 μL was injected in splitless mode with 1 min of purge time. The electron ionization source was run at 70 eV. A library search was carried out using NIST, NBS and Wiley GC–MS libraries for the identification of FAME. 2.4.4. Membrane characterizations Membrane morphology was observed by SEM using secondary electrons at 1 kV. To obtain a undeformed cross-section, the PA membrane was immersed in liquid nitrogen for five minutes and fractured perpendicularly to the surface. The behavior of the PA membrane containing residual solvent DMAc under the heating was established by thermogravimetry (TGA) using a thermogravimetric analyzer NETZSCH TG 209 F1 Libra at a heating rate of 10 °C min and a flow of 100 mL min 1 in air atmosphere. The sample (3 mg) was cut from the membrane and heated from 40 to 820 °C. For a sorption test, membrane strips were dried under a vacuum overnight at 40 °C. Then the samples were immersed in ethanol, methanol, biodiesel, 1% (m/m) methanol and ethanol solutions in biodiesel at 20 °C for 7 days. The samples were periodically taken out and blotted with tissue papers and then quickly weighed. The experiments were carried out until the swollen samples had constant weight. Then the samples were placed in vacuum at 40 °C for a week to calculate the amount of desorbing liquids in the samples. The degree of sorption (S0) for polymer/ liquid systems was calculated by the equation: S0 ¼ (m m0)/mp, where m – is the initial mass of the sample, m0 – is the swollen sample mass, mp–is the sample mass after the sorption test and drying. In the process of membrane preparation, the residual solvent (DMAc) was not completely removed from the membrane. A relative amount of residual solvent (SR) was diffused out of the membrane during the sorption test, which was calculated using the following equation: SR ¼ (m0– mp)/mp. The final degree of sorption (S) was calculated: S ¼S0 þ SR. The density (ρ) of PA membrane was determined by a flotation method in the solution of toluene and chloroform at 20 °C. The free volume vf (cm3 g 1) was calculated from the equation: vf ¼ vsp 1.3vw, where vsp ¼1/ρ – is the specific volume, vw – is the Van der Waals volume calculated by Bondi method [26,27].
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
3. Results and discussion 3.1. Characteristics of the PA membrane The presence of a residual solvent in a polymer membrane influences membrane transparency and its mechanical properties. Previously it was shown in pervaporation experiments that the presence of a residual solvent greatly increases flux through the membrane [19]. In this research, the PA membrane with a high residual solvent (DMAc) content was also chosen. Vapor separation by dense polymeric membranes was achieved through a dissolution–diffusion mechanism. Sorption, the first step, plays an important role in the process of permeability of lowmolecular weight substances. Sorption was measured in a liquid phase to demonstrate the maximum sorption ability of the PA membrane in the biodiesel components and to calculate the amount of residual DMAc. The results showed (Fig. 2) that the PA membrane absorbed more alcohols than biodiesel due to the different polarities of studied substances. A relative amount of residual solvent (SR) diffused out from the membrane during sorption experiment was equal to SRMeOH ¼0.33 g/100 g PA, water EtOH ¼0.52 g/100 g PA. These data deSR ¼0.32 g/100 g PA; SR monstrated a high content of residual solvent in the PA membrane. TGA was applied to determine the behavior of the PA membrane containing residual solvent DMAc under the heating. Fig. 3 A shows that the TGA curve has two regions. The first one shows that the weight loss up to 250 °C amounted to 5% (m/m) and this is the result of moisture removal and DMAс, which are not bound by hydrogen bonds with the PA. The second region (from 300 °C) describes the removal of DMAс during the destruction of its complex with PA and the polymer destruction. To study the inner structure of the PA membrane the SEM was applied. The PA membrane exhibits a brittle fracture surface (Fig. 4) with a few fracture lines (sharp white lines in the image) and some plastic deformations (rounded white lines in the image). Fracture surfaces of the studied membrane contain more plastic deformations than PA membrane studied in the work [19]. To explain the obtained result, density and free volume of the PA membranes were determined; these parameters were 1.268 g cm 3 and 0.095 cm3 g 1, respectively. In the previous research, these parameters for another PA type were found as 1.312 g cm 3 [17] and 0.069 g cm 3. The obtained results mean that the PA membrane studied in this paper has more transport channels for the transfer of low-molecular mass penetrants due to
Fig. 2. The PA membrane sorption.
Fig. 3. The TGA curve.
Fig. 4. SEM micrographs of cross-sections of nonporous PA membrane.
lower polymer chain packing as compared to the PA membrane studied in the previous work [27]. 3.2. Cyclic voltammetry measurements As previously reported [14], for the simultaneous determination of ethanol and methanol in fuel ethanol using cyclic voltammetry at a gold electrode, it was recommended to dilute the samples in an electrolyte solution containing 0.5 M NaOH and 0.1% (v/v) of methanol before analysis. To obtain oxidation of ethanol and methanol on gold electrode, gold oxide formation should be reproducible. This problem was excluded by the addition of 0.1% (v/v) methanol to 0.5 M NaOH [14]. In the current study this electrolyte solution was chosen as an acceptor solution. Ethanol was selectively detected at þ0.19 V and both compounds were detected at þ1.20 V (Fig. 5). Bubbles of CO2 were obtained during the alcohols' oxidation followed by the decrease of repeatability. It was necessary to remove any remaining CO2 bubbles from the gold working electrode surface. To solve this problem, it was essential to mix the electrolyte solution. Three techniques of electrolyte solution mixing were studied: an ultrasonic mixing, a magnetic stirrer and vibration. The magnetic stirrer was chosen as the most appropriate for further experiments. In this case any remaining CO2 bubbles were removed from the electrode surface and the highest repeatability
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Table 2 Influence of some interfering compounds on the determination of methanol and ethanol in electrolyte solution without VP (CMeOH ¼ 3 mM, CEtOH ¼ 2 mM). Alcohol
Isopentanol Heptanol Isobutanol Pentanol Butanol Isopropanol Glycerol
Fig. 5. Cyclic voltammograms obtained for the electrolyte solution (a) and different concentrations of methanol and ethanol: 0.05% (b); 0.1% (c); 0.25% (d); and 0.5% (e).
was observed (RSD 7 and 5% for methanol and ethanol, respectively). 3.3. VP-SWI system optimization The main design and operational parameters of the proposed VP-SWI system were optimized with respect to sensitivity and in some cases repeatability as well. All optimization experiments were conducted with methanol and ethanol solutions in biodiesel. The initial set of parameter values used was: volume and concentration of methanol and ethanol solutions in biodiesel – 1 mL, 1% (m/m), the vapor permeation temperature 70 °C; the flow rate of air stream – 5 mL min 1. Three techniques of vapor permeation intensification were studied: heating; gas bubbling through a sample and simultaneous heating; pumping out from the VPM by a syringe pump and simultaneous heating. It was found, that methanol and ethanol were effectively transported through a PA membrane only by gas bubbling and simultaneous heating. The increase of vapor permeation time and temperature (Fig. 6A) had a predictable positive effect on vapor permeation and on the analytical signal as well. Temperature had a significant influence on the efficiency of the vapor permeation process. Its increase promoted a higher vapor pressure of analyte and, therefore, a higher mass transfer through the membrane. To achieve a complete transfer of alcohols through the PA membrane it was necessary to heat the mixture at a temperature of 70 °C for 40 min by gas bubbling. Moreover, the influence of the flow rate on the transportation of alcohols through PA membrane was studied at
Tolerable concentration (mМ) Methanol
Ethanol
0.2 0.2 0.4 0.4 0.5 0.6 0.5
0.9 0.8 1.7 1.5 1.9 2.2 0.7
70 °C. It was found that (Fig. 6B) the analytical signal had constant values upon the following conditions of vapor permeation: the flow rate of air stream – 3 mL min 1 and the vapor permeation time – 40 min. However, to increase the sample throughput, a vapor permeation time equal to 10 min was chosen, which provided the required sensitivity at the permitted maximum level of alcohols in biodiesel (0.2% m/m). 3.4. Analytical performance The calibration plots were constructed using methanol and ethanol solutions in biodiesel in VP-SWIA condition at potential þ0.19 V for ethanol only and at potential þ 1.20 V for methanol and ethanol. A linear range between 0.05 and 0.5% (m/m) was established for each analyte. The found linear relationships foland lowed the equations CEtOH ¼ (I0.19 V þ1.5)/51.3 CMeOH ¼(I(1.20 V)–F I0.19 V–602)/16,409 where C is the concentration of alcohol expressed in % (m/m). The correction factor was calculated using a solution containing only ethanol. For this purpose, the current was determined at potential of 0.19 and 1.2 V. and the correction factor was calculated by the equation: F¼IEtOH 1.20 V /lMeOH 0.19 V ¼ 150. The limits of detection (LOD) are equal to three standard deviations (3 s) and the limits of quantification (LOQ) are equal to ten standard deviations (10 s) of the blank test (n ¼10). They were assessed as 0.02 and 0.05% (m/m) for both analytes, respectively. Linear ranges are suitable for determination of alcohols in biodiesel according to ASTM D 6751. The limits of detection can be decreased 4 times by increasing vapor permeation time up to 40 min, but in this case, sampling frequency would be decreased. The suggested procedure proved satisfactory repeatability of analytical response by the evaluation of the relative standard
Fig. 6. Investigation of VP-SWI experimental conditions: (A) effect of temperature and (B) effect of air flow speed.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Table 3 The results of the determination of methanol and ethanol in biodiesel samples (n¼5, t¼ 2.306, P¼ 0.95). Type of biodiesel
Added MeOH, % (m/m)
Found MeOH, % (m/m) VP-SWI
GC [25]
Added EtOH, % (m/m)
Found EtOH, % (m/m)
Recovery, %
t-test
VP-SWI
GC [25]
MeOH
EtOH
MeOH
EtOH
Palm
0 0.4 0.3
o LOD 0.41 70.02 0.29 7 0.01
oLOD 0.42 7 0.02 0.317 0.01
0 0.5 0.3
o LOD 0.50 70.01 0.317 0.01
o LOD 0.487 0.02 0.30 7 0.02
– 97 96
– 102 102
– 0.45 1.1
– 2 1.5
Canola
0 0.1 0.2
o LOD 0.117 0.01 0.217 0.02
oLOD 0.127 0.02 0.22 7 0.02
0 0.5 0.2
o LOD 0.487 0.02 0.217 0.02
o LOD 0.477 0.01 0.23 7 0.02
– 100 96
– 102 95
– 0.8 1.4
– 0.7 0.9
Corn
0 0.2 0.3
o LOD 0.20 7 0.01 0.30 7 0.01
oLOD 0.217 0.01 0.317 0.01
0 0.5 0.1
o LOD 0.50 70.01 0.117 0.01
o LOD 0.517 0.02 0.107 0.01
– 96 100
– 100 110
– 2.1 1
– 1.1 2
Sunflower
0 0.3 0.4
o LOD 0.30 7 0.01 0.38 7 0.02
oLOD 0.317 0.01 0.39 7 0.02
0 0.2 0.3
o LOD 0.20 70.01 0.30 70.02
o LOD 0.197 0.01 0.29 7 0.01
– 96 100
– 97 104
– 1.2 2.2
– 1.3 0.7
deviation (RSD) from 10 replicate measurements of content of alcohols in real samples with a value of 2.5–10% for methanol and 3– 5% for ethanol. The system throughput, assessed as the sampling frequency, was found to be 5 h 1. Evaluated analytical parameters were also compared with those reported in literature (Table 1).
3.5. Interference effect The effect of potentially interfering compounds, especially alcohols, on the determination of analytes was investigated with and without vapor permeation through a PA membrane. It was performed by the addition of a known concentration of each compound to fixed analyte concentrations in electrolyte solution in the MC (without VP) or in the biodiesel sample (with VP). The tolerable concentration of each interfering alcohol was considered to have less than 5% of relative error in the signal. It was found that various alcohols (Table 2) have a significant influence on voltammetric detection without VP. HРowever, GC analysis did not show the presence of the interfering alcohols in the aqueous phase after VP at concentrations above 10 5%. Thus, it was established that these compounds are not penetrated through the PA membrane under VP-SWI conditions.
4. Conclusions A novel flow method for the simultaneous determination of methanol and ethanol in biodiesel samples using cyclic voltammetry has been developed. To our knowledge, this is the first fully automated method for the simultaneous determination of methanol and ethanol in biodiesel. The stepwise injection analysis with the voltammetric detection and the vapor permeation was first realized. A selective polymer membrane based on poly(phenylene isophtalamide) containing a high amount of a residual solvent was first used for methanol and ethanol separation from biodiesel by means of vapor permeation technique. There was no direct contact of the biodiesel sample with the membrane, which prevented destruction of the membrane. Unlike the cyclic voltammetric method reported in [14], VPSWI prevents the significant influence of complex matrices. To increase methanol and ethanol transportation through the PA membrane, mixing of biodiesel sample in the VPM by bubbling was suggested. The applicability of the method was demonstrated with the help of real sample analysis and the obtained results were compared with the reference method.
Acknowledgment 3.6. Analytical application The proposed procedure was applied for the determination of methanol and ethanol in four biodiesel samples synthesized from different raw materials. The obtained results show that there are no significant differences in the concentration of alcohols obtained by the suggested and the reference method [25] (Table 3). Student's t-test of statistical hypotheses on the equality of the means was used to test the equality of the means of the experimental data obtained by VP-SWI and by the reference method. A hypothesis on the equality of the means of the experimental data obtained by the VP-SWI and by the reference method was taken as a null hypothesis (H0). A hypothesis on the difference of mentioned means was taken as an alternative hypothesis (H1). The level of significance was taken as equal to 0.05. Based on the obtained results, the observed differences did not contradict with the hypothesis H0 and the obtained discrepancies with the 0.05 significance level were insignificant.
This work was supported by the Russian Foundation for Basic Research (Project no. 13-03-00031-a, 14-03-31092mol and 15-3320068). Scientific research was performed at the Center for Chemical Analysis and Materials Research, at the Center of Thermal Analysis and Calorimetry, and at the Interdisciplinary Resource center for Nanotechnology of the St. Petersburg State University and Resource center for Chemical Education.
References [1] A. Demirbas, Prog. Energy Combust. Sci. 31 (2005) 466–487. [2] R. Geris, N.A.C. Santos, B.A. Amaral, I.S. Maia, V.D. Castro, J.R.M. Carvalho, Quim. Nova 30 (2007) 69–73. [3] A.C. Pinto, L.L.N. Guarieiro, M.J.C. Rezende, N.M. Ribeiro, E.A. Torres, W. A. Lopes, P.A. de, P. Pereira, J.B. de Andrade, J. Braz. Chem. Soc. 16 (2005) 1313–1330. [4] M. Mittelbach, G. Roth, A. Bergmann, Chromatographia 42 (1996) 431–434. [5] J.V. Seeley, S.K. Seeley, E.K. Libby, J.D. McCurry, J. Chromatogr. Sci. 45 (2007) 650–656.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎ [6] M.C. Paraschivescu, E.G. Alley, W.T. French, R. Hernandez, K. Armbrust, Bioresour. Technol. 99 (2008) 5901–5905. [7] A.C. Silva, L.F.B.L. Pontes, M.F. Pimentel, M.J.C. Pontes, Talanta 93 (2012) 129–134. [8] P.I. Felizardo, P. Baptista, J.C. Menezes, M.J. Correia, Anal. Chim. Acta 595 (2007) 107–113. [9] R.M. Balabin, E.I. Lomakina, R.Z. Safieva, Fuel 90 (2011) 2007–2015. 10 M.P. Dorado, S. Pinzi, A. de Haro, R. Font, J. Garcia-Olmo, Fuel 90 (2011) 2321–2325. [11] W. Zhang, Renew. Sustain. Energy Rev. 16 (2012) 6048–6058. [12] M.R. Monteiro, A.R.P. Ambrozin, L.M. Liao, A.G. Ferreira, Talanta 77 (2008) 593–605. [13] J.H.F. Boog, E.L.C. Silveira, L.B. de Caland, M. Tubino, Fuel 90 (2011) 905–907. [14] P.F. Pereira, R.M.F. Sousa, R.A.A. Munoz, E.M. Richter, Fuel 103 (2013) 725–729. [15] A.R.T.S. Araujo, M.L.M.F.S. Saraiva, J.L.F.C. Lima, M.G.A. Korn, Anal. Chim. Acta 613 (2008) 177–183. [16] V. Teplyakov, M. Shalygin, Integrated systems involving membrane vapor permeation and applications, in: A. Basile, A. Figoli, M. Khayet (Eds.), Pervaporation, Vapour Permeation and Membrane Distillation, Elsevier, 2015, pp. 177–201, http://dx.doi.org/10.1016/B978-1-78242-246-4.00007-6. [17] A.V. Penkova, G.A. Polotskaya, A.M. Toikka, M. Trchova, M. Slouf, M. Urbanova,
7
J. Brus, L. Brozova, Z. Pientka, Macromol. Mater. Eng. 294 (2009) 432–440. [18] A.V. Penkova, G.A. Polotskaya, V.A. Gavrilova, A.M. Toikka, J.C. Liu, M. Trchova, M. Slouf, Z. Pientka, Sep. Sci. Technol. 45 (2010) 35–41. [19] A.V. Penkova, G.A. Polotskaya, A.M. Toikka, V.V. Kocherbitov, Effect of residual solvent on physicochemical properties of poly(phenylene isophtalamide) membrane, Dry. Technol. 29 (2011) 633–641. [20] C.S. Vakh, A.V. Bulatov, A.Y. Shishov, A.V. Zabrodin, L.N. Moskvin, Fuel 135 (2014) 198–204. [21] A. Bulatov, K. Medinskaia, D. Aseeva, S. Garmonov, L. Moskvin, Talanta 133 (2015) 66–70. [22] M. Falkova, M. Alexovic, M. Pushina, A. Bulatov, L. Moskvin, V. Andruch, Microchem. J. 116 (2014) 98–106. [23] P.V. Tekade, O.A. Mahodaya, G.R. Khandeshwar, B.D. Joshi, Sci. Revs. Chem. Commun. 2 (2012) 208–211. [24] S. Basumatary, P. Barua, D.C. Deka, J. Chem. Pharm. Res. 5 (2013) 172–179. [25] R.E. Pauls, R.W. McCoy, J. Chromatogr. Sci. 19 (1981) 558–561. [26] A. Bondi, Liquids and Glasses, Wiley, New York, 1968. [27] G.A. Polotskaya, A.V. Penkova, A.M. Toikka, Z. Pientka, L. Brozova, M. Bleha, Sep. Sci. Technol. 42 (2007) 333–347.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Vapor permeation-stepwise injection simultaneous determination of methanol and ethanol in biodiesel with voltammetric detection Andrey Shishov n, Anastasia Penkova, Andrey Zabrodin, Konstantin Nikolaev, Maria Dmitrenko, Sergey Ermakov, Andrey Bulatov Institute of Chemistry, St. Petersburg State University, Universitetsky pr. 26, Saint Petersburg 198504, Russia
art ic l e i nf o
a b s t r a c t
Article history: Received 28 February 2015 Received in revised form 12 May 2015 Accepted 18 May 2015
A novel vapor permeation-stepwise injection (VP-SWI) method for the determination of methanol and ethanol in biodiesel samples is discussed. In the current study, stepwise injection analysis was successfully combined with voltammetric detection and vapor permeation. This method is based on the separation of methanol and ethanol from a sample using a vapor permeation module (VPM) with a selective polymer membrane based on poly(phenylene isophtalamide) (PA) containing high amounts of a residual solvent. After the evaporation into the headspace of the VPM, methanol and ethanol were transported, by gas bubbling, through a PA membrane to a mixing chamber equipped with a voltammetric detector. Ethanol was selectively detected at þ0.19 V, and both compounds were detected at þ1.20 V. Current subtractions (using a correction factor) were used for the selective determination of methanol. A linear range between 0.05 and 0.5% (m/m) was established for each analyte. The limits of detection were estimated at 0.02% (m/m) for ethanol and methanol. The sample throughput was 5 samples h 1. The method was successfully applied to the analysis of biodiesel samples. & 2015 Published by Elsevier B.V.
Keywords: Methanol Ethanol Biodiesel Vapor permeation Flow analysis Cyclic voltammetry
1. Introduction Biodiesel occupies a prominent position among alternatives to conventional petrodiesel fuel due to various technical, economic and ecological factors. The technology of biodiesel production includes the transesterification process of vegetable oil (e.g., soy oil) or animal fats (e.g., swine lard) with methanol or ethanol in the presence of a catalyst (alkali, acid or an enzyme) [1–3]. After the transesterification reaction, excess alcohol is removed from the biodiesel by extraction and distillation. The amount of alcohols in biodiesel is regulated by American (ASTM D 6751) and European (EN 14110) biodiesel standards and the permitted maximum level of alcohols in biodiesel samples is 0.2% (m/m). Excess alcohols cause metal corrosion, particularly aluminum corrosion; decrease diesel fuel flash point; and degrade rubber and polymer parts of engines. Various analytical methods have been developed (Table 1) for the determination of alcohols in biodiesel. They are based on gas chromatography (GC) [4–6], near infrared (NIR) [7–11] and UV–vis [10] spectroscopy. GC is the most used technique due to its high accuracy in the quantification of minor components. However, n
Corresponding author. E-mail address: [email protected] (A. Shishov).
baseline drift and overlapping signals can have a detrimental effect on GC accuracy [12]. The GC accuracy can be improved by using a headspace solid-phase microextraction [6]. This technique is characterized by high sensitivity, good reproducibility and recovery. The application of chemometric tools in the NIR spectroscopy allows the determination of methanol and water in biodiesel [8]. NIR and visible spectroscopy are able to predict methanol and glycerol traces in biodiesel samples [10]. Moreover, the technique for the determination of residual alcohol content in biodiesel through determination of its flash point was proposed [13], but this method does not allow the simultaneous determination of methanol and ethanol. In the present research [14], a cyclic voltammetry method was developed for the simultaneous determination of analytes in fuel ethanol. Automation of analysis is an important and rapidly growing trend in modern analytical chemistry. Recently, the automation of analytical procedures based on flow analysis has been developed. Therefore, labor costs for the analysis are decreased, and the volume of sample solutions and reagents are reduced. Currently only one flow method for the determination of methanol in biodiesel samples is described [15]. The proposed method is based on the liquid–liquid extraction of methanol from a sample solution to a phosphate buffer using a membrane unit with a polyvinylidene fluoride membrane. The determination of
http://dx.doi.org/10.1016/j.talanta.2015.05.041 0039-9140/& 2015 Published by Elsevier B.V.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
Table 1 Comparison of the suggested method with previously reported for the determination of methanol and ethanol in biodiesel. Detection technique Analyte
Sample preparation
Sample amount
On-line analysis
Linear range
LOD
Ref.
GC GC
MeOH MeOH
100 μL 1 mL
No No
4 6
MeOH MeOH
– 10 μL
No No
51 ppm –
9 10
FP CV
EtOH, MeOH – EtOH, MeOH Dilution
65 mL –
No No
– 0.005–0.05% (m/ m) – 0.003–0.433% (m/m) to 1% (v/v) 0.1–0.5% (v/v)
0.001% (m/m) –
NIR VIS and NIR
Dilution Headspace solid phase microextraction – –
13 14
VIS
MeOH
200 μL
Yes
0.1% (v/v) 0.028 and 0.045% (v/v) for EtOH, MeOH 0.0002% (m/m)
15
CV
EtOH, MeOH Vapor permeation
1 mL
Yes
0.02% (m/m)
This work
Membrane extraction
0.001–0.200% (m/m) 0.05–0.5% (m/m)
GC – gas chromatography, NIR – near infrared spectroscopy, VIS – visible spectroscopy, FP – flash point determination, CV – cyclic voltammetry.
methanol was then achieved in aqueous solution by converting it to hydrogen peroxide through the use of alcohol oxidase, followed by the use of 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) as an electron donor for horseradish peroxidase, giving the colored ABTS radical cation. In this work, a new flow method for the simultaneous determination of methanol and ethanol in different types of biodiesel with analytes vapor permeation through the PА membrane and cyclic voltammetry detection has been developed. Vapor permeation (VP) is frequently used for the separation of volatile analytes in complex samples [16] and can be coupled with flow methods. The PA was chosen as the membrane material because of its good physico-chemical properties and high selectivity with respect to methanol in pervaporation of tert-butyl methyl ether/ methanol and cyclohexane/methanol mixtures, as seen in previous work [17,18]. It should be noted that the PA membrane studied in this work is different from the PA membrane described in previous works [17–19]. The studied membrane has higher free volume, which significantly effects the transport of low molecular substances through it. The stepwise injection analysis [20–22] was chosen for the automation of methanol and ethanol determination in biodiesel because its manifold allows to mix solutions by a gas
flow. This feature can be used to increase the efficiency of methanol and ethanol penetration during vapor permeation.
2. Experimental 2.1. Manifold and apparatus A stepwise injection manifold (Fig. 1) includes a syringe pump (Sciware systems, Spain), a reversible peristaltic pump MasterFlex L/S (Cole-Parmer, USA) with modified PVC pumping tube (WatsonMarlow, Russia), two six-way solenoid valves (Cole-Parmer Inc., USA), a homemade mixing chamber (MC) (a cylindrical-shaped PTFE tube (20 mm in height and 15 mm in i.d.) equipped with a magnetic stirrer and integrated with the miniaturized home-made Ag/AgCl reference, platinum auxiliary and gold working (ؼ3 mm) electrodes), 797 VA Computrace Analizer (Metrohm, Switzerland), communication tubes (PTFE, 0.5 mm i.d.) and a laboratory-made vapor permeation module. The vapor permeation module (Fig. 1) was constructed from two titanium discs (i.d. 50 mm), held together by four stainless steel bolts. The depths of the chambers, separated by the PA membrane (35 mm diameter
Fig. 1. The VP-SWI manifold for the simultaneous determination of methanol and ethanol in biodiesel samples.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
and 40 mm thickness), were 25 mm. A water-bath, equipped with a thermostat, was used to preserve the temperature of the vapor permeation module. The manifold was operated automatically by means of a computer. The GC GC-MS-QP2010 Ultra gas chromatograph mass spectrometer (Shimadzu, Japan) was used for the determination of analytes and identification of fatty acid methyl esters (FAME) in biodiesel samples. The TG 209 F3 Iris thermo-microbalance (Netzsch, Germany) was used for thermal degradation measurements. Scanning electron microscopy (SEM) micrographs were obtained by a Merlin scanning electron microscope (Zeiss, Germany). 2.2. Reagents and samples Analytical grade chemicals were used throughout the work. 1% (m/m) stock solutions of methanol and ethanol were prepared by reagents (Sigma-Aldrich, USA) dissolving in biodiesel. The biodiesel (B 20) samples were obtained by mixing biofuel and diesel in a ratio of 1:4, respectively. Biofuels were obtained from the following feedstocks: sunflower, canola, corn and palm. These samples had been previously prepared with the use of sodium hydroxide as an alkaline catalyst [23]. To prevent the presence of alcohols, the biofuel samples were washed with water and dried for 2 hours at a temperature of 80 °C. From the GC–MS analysis [24], it was observed that the biofuel samples contained methyl and ethyl esters of n-nonanoic, n-hexadecanoic, n-octadecanoic, 9-octadecenoic and 9,12-octadecadienoic acids. Electrolyte solution (0.5 M NaOH and 0.1% (v/v) CH3OH) was prepared by NaOH (Sigma-Aldrich, USA) and CH3OH dissolving in deionized water (18.2 MΩ cm, Millipore, USA). 2.3. Membrane Poly(phenylene isophtalamide) (commercial sample Fenylons, Minkar, Russia) was used for membrane preparation. The PA membrane (thickness 40 mm) was obtained by casting 5% PA and 0.7% LiCl (Sigma-Aldrich, USA) solution in dimethylacetamide (DMAc) on a glass support with subsequent drying at 40 °C for 2 weeks. 2.4. Procedures 2.4.1. The VP-SWI procedure Initially (Fig. 1), 0.25 mL of electrolyte solution (channel 1) was aspirated through a valve 1 into the holding coil and then into the mixing chamber (channel 3) by the syringe pump. Then 1 mL of the sample (channel 7) was aspirated through the valve 2 into the donor part of VPM by the reversible peristaltic pump using the modified PVC pumping tube. The lifetime of the pumping tube was one week, and after every 50 analyses, it was replaced. The sample did not make contact with the membrane surface. The VPM was thermostated at 70 °C and air (channel 8) was transferred into the VPM by means of the peristaltic pump for 10 min at a rate of 3 mL min 1 through the valve 2 to mix sample and intensification of analytes' evaporation. After evaporation into the headspace of the VPM, methanol and ethanol were transported through the PA membrane by gas bubbling into the MC equipped with the magnetic stirrer and electrodes. The VPM was connected with the MC by means of PTFE tube. After the vapor permeation process and dilution in the electrolyte solution, the concentration of alcohols was measured by cyclic voltammetry. Ethanol was selectively detected at þ0.19 V and both analytes were detected at þ1.20 V. The correction factor was used for the selective determination of methanol. Then solution was then transfered from the MC to waste
3
(channel 4) and the MC was washed with electrolyte solution (channel 1). Finally, the sample was transfered from the donor part of VPM to waste (channel 10) by means of the reversible peristaltic pump and the system was washed with isooctane (channel 9). All parameters of the flow system were controlled by a PC. 2.4.2. The GC procedure for determination of alcohols 0.5 mL of each biodiesel sample was shaken with 5 mL of water for half an hour [25]. Then the solution was centrifuged, and the content of alcohols in the water phase was measured by GC. The GC was fitted with a SPB-624 MS capillary column (Stabilwax polar phase, Crossbondspolyethylene glycol, 30 m, 0.32 mm i.d., 0.5 μm coating). The injector temperature was kept at 250 °C and temperature of the column at 40 °C. Helium was used as a carrier gas at a flow rate of 1.5 mL min 1. A sample volume of 10 μL was injected in splitless mode with 1 min of purge time. 2.4.3. The GC–MS procedure for assessing FAME in biofuel FAME in biofuel was estimated using gas chromatograph mass spectrometry [24]. The GC was fitted with a SPB-624 MS capillary column (6% cyanopropyl phenyl and 94% dimethyl polysiloxane, 30 m, 0.25 mm i.d., 0.25 μm coating). The injector temperature was kept at 220 °C while the interface temperature and the ion source were 200 °C. Helium was used as a carrier gas at a flow rate of 2 mL min 1. The sample volume of 2 μL was injected in splitless mode with 1 min of purge time. The electron ionization source was run at 70 eV. A library search was carried out using NIST, NBS and Wiley GC–MS libraries for the identification of FAME. 2.4.4. Membrane characterizations Membrane morphology was observed by SEM using secondary electrons at 1 kV. To obtain a undeformed cross-section, the PA membrane was immersed in liquid nitrogen for five minutes and fractured perpendicularly to the surface. The behavior of the PA membrane containing residual solvent DMAc under the heating was established by thermogravimetry (TGA) using a thermogravimetric analyzer NETZSCH TG 209 F1 Libra at a heating rate of 10 °C min and a flow of 100 mL min 1 in air atmosphere. The sample (3 mg) was cut from the membrane and heated from 40 to 820 °C. For a sorption test, membrane strips were dried under a vacuum overnight at 40 °C. Then the samples were immersed in ethanol, methanol, biodiesel, 1% (m/m) methanol and ethanol solutions in biodiesel at 20 °C for 7 days. The samples were periodically taken out and blotted with tissue papers and then quickly weighed. The experiments were carried out until the swollen samples had constant weight. Then the samples were placed in vacuum at 40 °C for a week to calculate the amount of desorbing liquids in the samples. The degree of sorption (S0) for polymer/ liquid systems was calculated by the equation: S0 ¼ (m m0)/mp, where m – is the initial mass of the sample, m0 – is the swollen sample mass, mp–is the sample mass after the sorption test and drying. In the process of membrane preparation, the residual solvent (DMAc) was not completely removed from the membrane. A relative amount of residual solvent (SR) was diffused out of the membrane during the sorption test, which was calculated using the following equation: SR ¼ (m0– mp)/mp. The final degree of sorption (S) was calculated: S ¼S0 þ SR. The density (ρ) of PA membrane was determined by a flotation method in the solution of toluene and chloroform at 20 °C. The free volume vf (cm3 g 1) was calculated from the equation: vf ¼ vsp 1.3vw, where vsp ¼1/ρ – is the specific volume, vw – is the Van der Waals volume calculated by Bondi method [26,27].
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
3. Results and discussion 3.1. Characteristics of the PA membrane The presence of a residual solvent in a polymer membrane influences membrane transparency and its mechanical properties. Previously it was shown in pervaporation experiments that the presence of a residual solvent greatly increases flux through the membrane [19]. In this research, the PA membrane with a high residual solvent (DMAc) content was also chosen. Vapor separation by dense polymeric membranes was achieved through a dissolution–diffusion mechanism. Sorption, the first step, plays an important role in the process of permeability of lowmolecular weight substances. Sorption was measured in a liquid phase to demonstrate the maximum sorption ability of the PA membrane in the biodiesel components and to calculate the amount of residual DMAc. The results showed (Fig. 2) that the PA membrane absorbed more alcohols than biodiesel due to the different polarities of studied substances. A relative amount of residual solvent (SR) diffused out from the membrane during sorption experiment was equal to SRMeOH ¼0.33 g/100 g PA, water EtOH ¼0.52 g/100 g PA. These data deSR ¼0.32 g/100 g PA; SR monstrated a high content of residual solvent in the PA membrane. TGA was applied to determine the behavior of the PA membrane containing residual solvent DMAc under the heating. Fig. 3 A shows that the TGA curve has two regions. The first one shows that the weight loss up to 250 °C amounted to 5% (m/m) and this is the result of moisture removal and DMAс, which are not bound by hydrogen bonds with the PA. The second region (from 300 °C) describes the removal of DMAс during the destruction of its complex with PA and the polymer destruction. To study the inner structure of the PA membrane the SEM was applied. The PA membrane exhibits a brittle fracture surface (Fig. 4) with a few fracture lines (sharp white lines in the image) and some plastic deformations (rounded white lines in the image). Fracture surfaces of the studied membrane contain more plastic deformations than PA membrane studied in the work [19]. To explain the obtained result, density and free volume of the PA membranes were determined; these parameters were 1.268 g cm 3 and 0.095 cm3 g 1, respectively. In the previous research, these parameters for another PA type were found as 1.312 g cm 3 [17] and 0.069 g cm 3. The obtained results mean that the PA membrane studied in this paper has more transport channels for the transfer of low-molecular mass penetrants due to
Fig. 2. The PA membrane sorption.
Fig. 3. The TGA curve.
Fig. 4. SEM micrographs of cross-sections of nonporous PA membrane.
lower polymer chain packing as compared to the PA membrane studied in the previous work [27]. 3.2. Cyclic voltammetry measurements As previously reported [14], for the simultaneous determination of ethanol and methanol in fuel ethanol using cyclic voltammetry at a gold electrode, it was recommended to dilute the samples in an electrolyte solution containing 0.5 M NaOH and 0.1% (v/v) of methanol before analysis. To obtain oxidation of ethanol and methanol on gold electrode, gold oxide formation should be reproducible. This problem was excluded by the addition of 0.1% (v/v) methanol to 0.5 M NaOH [14]. In the current study this electrolyte solution was chosen as an acceptor solution. Ethanol was selectively detected at þ0.19 V and both compounds were detected at þ1.20 V (Fig. 5). Bubbles of CO2 were obtained during the alcohols' oxidation followed by the decrease of repeatability. It was necessary to remove any remaining CO2 bubbles from the gold working electrode surface. To solve this problem, it was essential to mix the electrolyte solution. Three techniques of electrolyte solution mixing were studied: an ultrasonic mixing, a magnetic stirrer and vibration. The magnetic stirrer was chosen as the most appropriate for further experiments. In this case any remaining CO2 bubbles were removed from the electrode surface and the highest repeatability
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Table 2 Influence of some interfering compounds on the determination of methanol and ethanol in electrolyte solution without VP (CMeOH ¼ 3 mM, CEtOH ¼ 2 mM). Alcohol
Isopentanol Heptanol Isobutanol Pentanol Butanol Isopropanol Glycerol
Fig. 5. Cyclic voltammograms obtained for the electrolyte solution (a) and different concentrations of methanol and ethanol: 0.05% (b); 0.1% (c); 0.25% (d); and 0.5% (e).
was observed (RSD 7 and 5% for methanol and ethanol, respectively). 3.3. VP-SWI system optimization The main design and operational parameters of the proposed VP-SWI system were optimized with respect to sensitivity and in some cases repeatability as well. All optimization experiments were conducted with methanol and ethanol solutions in biodiesel. The initial set of parameter values used was: volume and concentration of methanol and ethanol solutions in biodiesel – 1 mL, 1% (m/m), the vapor permeation temperature 70 °C; the flow rate of air stream – 5 mL min 1. Three techniques of vapor permeation intensification were studied: heating; gas bubbling through a sample and simultaneous heating; pumping out from the VPM by a syringe pump and simultaneous heating. It was found, that methanol and ethanol were effectively transported through a PA membrane only by gas bubbling and simultaneous heating. The increase of vapor permeation time and temperature (Fig. 6A) had a predictable positive effect on vapor permeation and on the analytical signal as well. Temperature had a significant influence on the efficiency of the vapor permeation process. Its increase promoted a higher vapor pressure of analyte and, therefore, a higher mass transfer through the membrane. To achieve a complete transfer of alcohols through the PA membrane it was necessary to heat the mixture at a temperature of 70 °C for 40 min by gas bubbling. Moreover, the influence of the flow rate on the transportation of alcohols through PA membrane was studied at
Tolerable concentration (mМ) Methanol
Ethanol
0.2 0.2 0.4 0.4 0.5 0.6 0.5
0.9 0.8 1.7 1.5 1.9 2.2 0.7
70 °C. It was found that (Fig. 6B) the analytical signal had constant values upon the following conditions of vapor permeation: the flow rate of air stream – 3 mL min 1 and the vapor permeation time – 40 min. However, to increase the sample throughput, a vapor permeation time equal to 10 min was chosen, which provided the required sensitivity at the permitted maximum level of alcohols in biodiesel (0.2% m/m). 3.4. Analytical performance The calibration plots were constructed using methanol and ethanol solutions in biodiesel in VP-SWIA condition at potential þ0.19 V for ethanol only and at potential þ 1.20 V for methanol and ethanol. A linear range between 0.05 and 0.5% (m/m) was established for each analyte. The found linear relationships foland lowed the equations CEtOH ¼ (I0.19 V þ1.5)/51.3 CMeOH ¼(I(1.20 V)–F I0.19 V–602)/16,409 where C is the concentration of alcohol expressed in % (m/m). The correction factor was calculated using a solution containing only ethanol. For this purpose, the current was determined at potential of 0.19 and 1.2 V. and the correction factor was calculated by the equation: F¼IEtOH 1.20 V /lMeOH 0.19 V ¼ 150. The limits of detection (LOD) are equal to three standard deviations (3 s) and the limits of quantification (LOQ) are equal to ten standard deviations (10 s) of the blank test (n ¼10). They were assessed as 0.02 and 0.05% (m/m) for both analytes, respectively. Linear ranges are suitable for determination of alcohols in biodiesel according to ASTM D 6751. The limits of detection can be decreased 4 times by increasing vapor permeation time up to 40 min, but in this case, sampling frequency would be decreased. The suggested procedure proved satisfactory repeatability of analytical response by the evaluation of the relative standard
Fig. 6. Investigation of VP-SWI experimental conditions: (A) effect of temperature and (B) effect of air flow speed.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Table 3 The results of the determination of methanol and ethanol in biodiesel samples (n¼5, t¼ 2.306, P¼ 0.95). Type of biodiesel
Added MeOH, % (m/m)
Found MeOH, % (m/m) VP-SWI
GC [25]
Added EtOH, % (m/m)
Found EtOH, % (m/m)
Recovery, %
t-test
VP-SWI
GC [25]
MeOH
EtOH
MeOH
EtOH
Palm
0 0.4 0.3
o LOD 0.41 70.02 0.29 7 0.01
oLOD 0.42 7 0.02 0.317 0.01
0 0.5 0.3
o LOD 0.50 70.01 0.317 0.01
o LOD 0.487 0.02 0.30 7 0.02
– 97 96
– 102 102
– 0.45 1.1
– 2 1.5
Canola
0 0.1 0.2
o LOD 0.117 0.01 0.217 0.02
oLOD 0.127 0.02 0.22 7 0.02
0 0.5 0.2
o LOD 0.487 0.02 0.217 0.02
o LOD 0.477 0.01 0.23 7 0.02
– 100 96
– 102 95
– 0.8 1.4
– 0.7 0.9
Corn
0 0.2 0.3
o LOD 0.20 7 0.01 0.30 7 0.01
oLOD 0.217 0.01 0.317 0.01
0 0.5 0.1
o LOD 0.50 70.01 0.117 0.01
o LOD 0.517 0.02 0.107 0.01
– 96 100
– 100 110
– 2.1 1
– 1.1 2
Sunflower
0 0.3 0.4
o LOD 0.30 7 0.01 0.38 7 0.02
oLOD 0.317 0.01 0.39 7 0.02
0 0.2 0.3
o LOD 0.20 70.01 0.30 70.02
o LOD 0.197 0.01 0.29 7 0.01
– 96 100
– 97 104
– 1.2 2.2
– 1.3 0.7
deviation (RSD) from 10 replicate measurements of content of alcohols in real samples with a value of 2.5–10% for methanol and 3– 5% for ethanol. The system throughput, assessed as the sampling frequency, was found to be 5 h 1. Evaluated analytical parameters were also compared with those reported in literature (Table 1).
3.5. Interference effect The effect of potentially interfering compounds, especially alcohols, on the determination of analytes was investigated with and without vapor permeation through a PA membrane. It was performed by the addition of a known concentration of each compound to fixed analyte concentrations in electrolyte solution in the MC (without VP) or in the biodiesel sample (with VP). The tolerable concentration of each interfering alcohol was considered to have less than 5% of relative error in the signal. It was found that various alcohols (Table 2) have a significant influence on voltammetric detection without VP. HРowever, GC analysis did not show the presence of the interfering alcohols in the aqueous phase after VP at concentrations above 10 5%. Thus, it was established that these compounds are not penetrated through the PA membrane under VP-SWI conditions.
4. Conclusions A novel flow method for the simultaneous determination of methanol and ethanol in biodiesel samples using cyclic voltammetry has been developed. To our knowledge, this is the first fully automated method for the simultaneous determination of methanol and ethanol in biodiesel. The stepwise injection analysis with the voltammetric detection and the vapor permeation was first realized. A selective polymer membrane based on poly(phenylene isophtalamide) containing a high amount of a residual solvent was first used for methanol and ethanol separation from biodiesel by means of vapor permeation technique. There was no direct contact of the biodiesel sample with the membrane, which prevented destruction of the membrane. Unlike the cyclic voltammetric method reported in [14], VPSWI prevents the significant influence of complex matrices. To increase methanol and ethanol transportation through the PA membrane, mixing of biodiesel sample in the VPM by bubbling was suggested. The applicability of the method was demonstrated with the help of real sample analysis and the obtained results were compared with the reference method.
Acknowledgment 3.6. Analytical application The proposed procedure was applied for the determination of methanol and ethanol in four biodiesel samples synthesized from different raw materials. The obtained results show that there are no significant differences in the concentration of alcohols obtained by the suggested and the reference method [25] (Table 3). Student's t-test of statistical hypotheses on the equality of the means was used to test the equality of the means of the experimental data obtained by VP-SWI and by the reference method. A hypothesis on the equality of the means of the experimental data obtained by the VP-SWI and by the reference method was taken as a null hypothesis (H0). A hypothesis on the difference of mentioned means was taken as an alternative hypothesis (H1). The level of significance was taken as equal to 0.05. Based on the obtained results, the observed differences did not contradict with the hypothesis H0 and the obtained discrepancies with the 0.05 significance level were insignificant.
This work was supported by the Russian Foundation for Basic Research (Project no. 13-03-00031-a, 14-03-31092mol and 15-3320068). Scientific research was performed at the Center for Chemical Analysis and Materials Research, at the Center of Thermal Analysis and Calorimetry, and at the Interdisciplinary Resource center for Nanotechnology of the St. Petersburg State University and Resource center for Chemical Education.
References [1] A. Demirbas, Prog. Energy Combust. Sci. 31 (2005) 466–487. [2] R. Geris, N.A.C. Santos, B.A. Amaral, I.S. Maia, V.D. Castro, J.R.M. Carvalho, Quim. Nova 30 (2007) 69–73. [3] A.C. Pinto, L.L.N. Guarieiro, M.J.C. Rezende, N.M. Ribeiro, E.A. Torres, W. A. Lopes, P.A. de, P. Pereira, J.B. de Andrade, J. Braz. Chem. Soc. 16 (2005) 1313–1330. [4] M. Mittelbach, G. Roth, A. Bergmann, Chromatographia 42 (1996) 431–434. [5] J.V. Seeley, S.K. Seeley, E.K. Libby, J.D. McCurry, J. Chromatogr. Sci. 45 (2007) 650–656.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i
A. Shishov et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎ [6] M.C. Paraschivescu, E.G. Alley, W.T. French, R. Hernandez, K. Armbrust, Bioresour. Technol. 99 (2008) 5901–5905. [7] A.C. Silva, L.F.B.L. Pontes, M.F. Pimentel, M.J.C. Pontes, Talanta 93 (2012) 129–134. [8] P.I. Felizardo, P. Baptista, J.C. Menezes, M.J. Correia, Anal. Chim. Acta 595 (2007) 107–113. [9] R.M. Balabin, E.I. Lomakina, R.Z. Safieva, Fuel 90 (2011) 2007–2015. 10 M.P. Dorado, S. Pinzi, A. de Haro, R. Font, J. Garcia-Olmo, Fuel 90 (2011) 2321–2325. [11] W. Zhang, Renew. Sustain. Energy Rev. 16 (2012) 6048–6058. [12] M.R. Monteiro, A.R.P. Ambrozin, L.M. Liao, A.G. Ferreira, Talanta 77 (2008) 593–605. [13] J.H.F. Boog, E.L.C. Silveira, L.B. de Caland, M. Tubino, Fuel 90 (2011) 905–907. [14] P.F. Pereira, R.M.F. Sousa, R.A.A. Munoz, E.M. Richter, Fuel 103 (2013) 725–729. [15] A.R.T.S. Araujo, M.L.M.F.S. Saraiva, J.L.F.C. Lima, M.G.A. Korn, Anal. Chim. Acta 613 (2008) 177–183. [16] V. Teplyakov, M. Shalygin, Integrated systems involving membrane vapor permeation and applications, in: A. Basile, A. Figoli, M. Khayet (Eds.), Pervaporation, Vapour Permeation and Membrane Distillation, Elsevier, 2015, pp. 177–201, http://dx.doi.org/10.1016/B978-1-78242-246-4.00007-6. [17] A.V. Penkova, G.A. Polotskaya, A.M. Toikka, M. Trchova, M. Slouf, M. Urbanova,
7
J. Brus, L. Brozova, Z. Pientka, Macromol. Mater. Eng. 294 (2009) 432–440. [18] A.V. Penkova, G.A. Polotskaya, V.A. Gavrilova, A.M. Toikka, J.C. Liu, M. Trchova, M. Slouf, Z. Pientka, Sep. Sci. Technol. 45 (2010) 35–41. [19] A.V. Penkova, G.A. Polotskaya, A.M. Toikka, V.V. Kocherbitov, Effect of residual solvent on physicochemical properties of poly(phenylene isophtalamide) membrane, Dry. Technol. 29 (2011) 633–641. [20] C.S. Vakh, A.V. Bulatov, A.Y. Shishov, A.V. Zabrodin, L.N. Moskvin, Fuel 135 (2014) 198–204. [21] A. Bulatov, K. Medinskaia, D. Aseeva, S. Garmonov, L. Moskvin, Talanta 133 (2015) 66–70. [22] M. Falkova, M. Alexovic, M. Pushina, A. Bulatov, L. Moskvin, V. Andruch, Microchem. J. 116 (2014) 98–106. [23] P.V. Tekade, O.A. Mahodaya, G.R. Khandeshwar, B.D. Joshi, Sci. Revs. Chem. Commun. 2 (2012) 208–211. [24] S. Basumatary, P. Barua, D.C. Deka, J. Chem. Pharm. Res. 5 (2013) 172–179. [25] R.E. Pauls, R.W. McCoy, J. Chromatogr. Sci. 19 (1981) 558–561. [26] A. Bondi, Liquids and Glasses, Wiley, New York, 1968. [27] G.A. Polotskaya, A.V. Penkova, A.M. Toikka, Z. Pientka, L. Brozova, M. Bleha, Sep. Sci. Technol. 42 (2007) 333–347.
Please cite this article as: A. Shishov, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.05.041i