Development of a Novel Leak-Free Constant-Pressure Cylinder for

Oct 20, 2017 - The development of accurate CRMs relies on the maintenance of their composition. However, a PCPC might allow hydrocarbons to leak owing...
11 downloads 10 Views 660KB Size
Download PDF
Technical Note Cite This: Anal. Chem. 2017, 89, 11924-11928

pubs.acs.org/ac

Development of a Novel Leak-Free Constant-Pressure Cylinder for Certified Reference Materials of Liquid Hydrocarbon Mixtures Yong Doo Kim,†,‡ Ji Hwan Kang,†,‡ Hyun Kil Bae,† Namgoo Kang,†,§ Sang Hyub Oh,†,§ Jin-Hong Lee,‡ Jin Chun Woo,† and Sangil Lee*,†,§ †

Center for Gas Analysis, Korea Research Institute of Standards and Science (KRISS), Daejeon 34113, Republic of Korea Department of Environmental Engineering, Chungnam National University, Daejeon 34134, Republic of Korea § Science of Measurement, University of Science and Technology, Daejeon 34113, Republic of Korea ‡

S Supporting Information *

ABSTRACT: Liquid hydrocarbon mixtures such as liquefied petroleum gas and liquefied natural gas are becoming integral parts of the world’s energy system. Certified reference materials (CRMs) of liquid hydrocarbon mixtures are necessary to allow assessment of the accuracy and traceability of the compositions of such materials. A piston-type constant-pressure cylinder (PCPC) comprising chambers for a pressurizing gas (helium) and liquid (hydrocarbons) separated by a piston can be used to develop accurate and traceable liquid hydrocarbon mixture CRMs. The development of accurate CRMs relies on the maintenance of their composition. However, a PCPC might allow hydrocarbons to leak owing to the imperfect seal of the piston. In this study, a novel leak-free bellows-type constantpressure cylinder (BCPC) is designed and evaluated by comparison with PCPCs. Liquid hydrocarbon mixtures consisting of ethane, propane, propene, isobutane, n-butane, 1-butene, and isopentane were prepared in both types of constant pressure cylinders and then monitored to check leakages between the gas and liquid chambers. Overall, notable leakage occurred from and into both chambers in the PCPCs, whereas no leakage occurred in the BCPCs in the three months after their gravimetric preparation. The BCPCs maintained no leakage even 10 months after their preparation, whereas the PCPCs showed significantly increasing leakage during the same period.

C

hydrocarbon mixture is much higher than the vapor pressure of the mixture. A liquid hydrocarbon mixture in a PCPC remains in the liquid phase without any evaporation during its use.1 Liquid hydrocarbon mixtures in both types of cylinder have been tested to evaluate changes in their initial composition as they are withdrawn for analysis (i.e., as the volume of mixture decreases).1 Significant compositional changes occurred in the standard gas cylinders, whereas PCPCs showed only minor changes.1 For example, the amount-of-substance fractions (i.e., mole fractions) of ethane and propane in two standard gas cylinders decreased by 20−22% and 7−11%, respectively, but remained unchanged for PCPCs. As a result, PCPCs are widely used to develop accurate and traceable standards for liquid hydrocarbon mixtures. International comparison studies have been designed to assess the preparative and analytical capabilities of liquid hydrocarbon mixture standards developed in PCPCs;3,4 however, the composition of liquid hydrocarbon mixtures in a PCPC might change with time, given that the

ertified reference materials of liquid hydrocarbon mixtures comprising small-chain hydrocarbon components (C1− C5) at cmol mol−1 levels can be gravimetrically prepared in either standard gas cylinders or constant-pressure cylinders.1,2 In a standard gas cylinder, mixtures are pressurized by inert helium to remain as liquid; however, as the mixture is withdrawn through the dip tube, its volume in the liquid phase decreases, and then, the pressure of helium also decreases (as helium occupies the volume previously occupied by the removed liquid hydrocarbons). The changing hydrocarbon volume and helium pressure alter the composition of the liquid hydrocarbon mixture away from the initially prepared values.1 To reduce the change in helium pressure, a standard gas cylinder with a dual port valve can be used to prepare liquid hydrocarbon mixtures. Liquid hydrocarbons can be withdrawn from one port, while additional helium can be introduced through the other to pressurize the mixture. The second type of cylinder is a piston-type constant pressure cylinder (PCPC) that has two separate chambers for both components. Unlike a standard gas cylinder, the liquid hydrocarbon mixture is completely separated from the pressurizing helium by a piston with a rubber O-ring. The pressure exerted by the helium through the piston onto the © 2017 American Chemical Society

Received: September 21, 2017 Accepted: October 20, 2017 Published: October 20, 2017 11924

DOI: 10.1021/acs.analchem.7b03858 Anal. Chem. 2017, 89, 11924−11928

Technical Note

Analytical Chemistry piston does not guarantee a perfect seal.3 In this study, we developed a novel leak-free bellows-type constant-pressure cylinder (BCPC) for liquid hydrocarbon mixture standards. In contrast to the PCPC that separates two chambers with a piston, the BCPC has a bellows-type liquid chamber (without a seal between the two chambers) that is completely separated from the gas chamber containing the pressurizing helium. We prepared liquid hydrocarbon mixtures in two different types of constant-pressure cylinders and then compared their performance, focusing on leakage between the two chambers.

1 L (Welker, U.S.A.). Figure 1 shows it equipped with pressure gauges and safety valves for both chambers. Hydrocarbon Reagents. Before preparation of the hydrocarbon mixtures, neat reagents were analyzed for impurities by gas chromatography (GC) with flame ionization detection (FID; 6890N, Agilent Technologies, U.S.A.) for hydrocarbons, GC (6890N) with an atomic emission detector (AED; 2390AA, JAS, Germany) for helium, and a Karl Fischer coulometer (831 KF, Metrohm, Switzerland) for moisture. The purities of hydrocarbons are summarized in Table S-1. Preparation of Liquid Hydrocarbon Mixtures. The constant-pressure cylinders were checked for leaks before use. They were evacuated to a vacuum state on the order of 10−5 Pa using a turbo pump (ATP-80, Adixen, France) while being heated at ∼70 °C with a heating jacket. After evacuation, the cylinders were left in the laboratory for about 24 h to reach thermal equilibrium at room temperature. Liquid hydrocarbon mixtures were prepared by sequentially introducing reagents using a custom-made injection system (Figure S-2). The flexible tube attached to the sample cylinder allowed rough monitoring of the injected amount. After each reagent was injected, the interior of the injection system was evacuated and purged with nitrogen. The reagent with the lowest saturation vapor pressure was injected first. The mass amount of each injected reagent was gravimetrically determined using an automated top-pan balance (XP-26003L, Metler-Toledo, Switzerland) with 26 kg of capacity and 1 mg of resolution.2,5,6 Liquid hydrocarbon mixtures were prepared in both PCPCs and BCPCs to evaluate possible leakage between the two chambers of each cylinder. Helium was injected into the gas chamber to pressurize the liquid chamber to about 2400 kPa in order to prevent the liquid mixtures from evaporating. The cylinder was rotated vertically and horizontally for 3 h to ensure homogeneous mixing of all the substances. Chromatographic Analysis. To evaluate helium leakage into the liquid chamber, helium was analyzed using GC (GC/ TCD; 6890N, Agilent Technologies, U.S.A.) equipped with a Molesieve 5A stainless steel column (4 m length, 3.2 mm diameter, 80/100 mesh size) with a 1 mL sample loop. The isothermal GC oven temperature was set at 50 °C. The TCD was set at 250 °C with an argon carrier gas flow rate of 40 mL/ min to a reference and 20 mL/min to the sample cell. The matrix differs between the primary standard gas mixtures (PSMs) in nitrogen and the samples in liquid hydrocarbons used for quantification. To overcome the matrix effect7,8 during analysis, samples were introduced at a constant flow rate by using a flow controller (Model N-1189−1, Swagelok, U.S.A.) and monitored with a bubble flow meter. To evaluate hydrocarbon leakage into the gas chamber, hydrocarbons were analyzed with a GC/FID (7890N, Agilent Technologies, U.S.A.) equipped with a porous layer open tubular (PLOT) Al2O3/KCL capillary column (50 m length, 0.53 mm diameter, 10 μm thickness) and a 0.25 mL sample loop. The GC oven temperature was operated isothermally at 110 °C. The FID temperature was set at 250 °C with a helium carrier flow rate of 5 mL/min. The sample was set at a split ratio of 70:1. Again, the matrix effect (between PSMs in nitrogen and samples in helium used for quantification) during analysis was overcome by introducing samples at a constant flow rate using a flow controller (Model N-1189−1, Swagelok, U.S.A.) instead of a mass flow controller. As above, the flow was monitored using a bubble flow meter.

■

EXPERIMENTAL METHOD Bellows and Piston-Type Constant-Pressure Cylinders. To overcome the disadvantages of the PCPC (Figure 1),

Figure 1. Schematic diagram of a commercial piston-type constant pressure cylinder (PCPC). A: liquid chamber, B: gas chamber, C: piston with sliding O-ring, D: gas and liquid valve, E: pressure gauge, F: safety valve.

Figure 2. Schematic diagram of a bellows-type constant-pressure cylinder (BCPC). A: liquid chamber, B: gas chamber, C: sampling valve, D stop plate, E: magnetic indicator, F: O-ring, G: pressurized gas inlet, H: safety valve, I: pressure gauge, K: weighing base. Numerical values (mm) indicate the dimension of each part.

we designed and built the BCPC shown in Figure 2. The liquid chamber (A in Figure 2) comprised stainless steel bellows manufactured by Korea-Bellows (Korea) instead of a piston, and the O-ring (F) was used only to seal the gas chamber (B). The internal structure of the cylinder prior to its assembly is shown in Figure S-1. Both chambers had inlet valves (C and G), while only the gas chamber had a pressure gauge (I) and safety valve (H). A stop plate (D) prevented expansion of the bellows beyond the nominal maximum volume (700 mL) of the liquid chamber. A pair of magnets (E) indicated contraction and expansion of the bellows: one is attached to the bottom plate of the bellows, and the other (outside the gas chamber) visibly followed its movement. A base plate (K) installed on the outside of the bellows allowed stable measurement of mass using a high-precision balance. A commercially available PCPC was used for the comparison, comprising a stainless-steel cylinder with an internal volume of 11925

DOI: 10.1021/acs.analchem.7b03858 Anal. Chem. 2017, 89, 11924−11928

Technical Note

Analytical Chemistry

in normal length; the maximum filling pressure was 500 kPa. They can be compressed to 10 cm (J in Figure 2). The external diameter and length of the cylinder were 8 and 50 cm, respectively (without the weighing base in Figure 2). A pressure gauge and safety valve were installed in the gas chamber (B in Figure 2) but not the liquid chamber (A in Figure 2) in order to eliminate the presence of an unnecessary dead space in the chamber, which can hinder homogenization of the mixtures. The liquid and gas chambers were tested for leakage by monitoring pressure change at both vacuum pressure and 3000 kPa (filled with helium): no pressure change was found in either chamber. To establish the working pressure of helium for pressuring liquid mixtures, the effective saturation vapor pressure of the hydrocarbon mixtures was calculated using the GasVLe program (DNVGL, U.K.).11 The effective saturation vapor pressure of each liquid mixture listed in Table 1 was estimated as ∼700 kPa. The working pressure (2400 kPa) of helium, which is about three times larger than the effective saturation vapor pressure, was high enough to prevent liquid-to-gas phase conversion in the liquid chamber. During the normal working of shrinking and expanding the bellows, the filling capacity (or volume) of the liquid chamber ranged from ∼200 to ∼700 mL. In addition, a base plate (K in Figure 2) was attached securely at the bottom of the cylinder to load the BCPC onto the automatic top-pan balance. The leaked amount-of-substance fractions of helium into the liquid chamber are summarized in Table 2. For the two PCPCs, helium was detected at (0.1206 ± 0.0016) cmol mol−1 (PCPC1) and (0.1998 ± 0.0054) cmol mol−1 (PCPC-2) in the liquid chamber at 68 and 90 days after the gravimetric preparation, respectively, indicating that helium started leaking to the liquid chamber shortly after preparation. A second leak check at 254 days after the first found increases in the leaked amount-ofsubstance fraction of helium of (0.2486 ± 0.0077) cmol mol−1 (PCPC-1) and (0.1676 ± 0.0063) cmol mol−1 (PCPC-2), reaching ∼0.37 cmol mol−1 for both cylinders. The initial leakage rates before the first leak check (Table 3) were estimated as (1.77 × 10−3 ± 2.61 × 10−5) cmol mol−1 day−1 (PCPC-1) and (2.22 × 10−3 ± 2.47 × 10−5) cmol mol−1 day−1 (PCPC-2), while those between the first and second checks were estimated as (0.98 × 10−3 ± 3.07 × 10−5) cmol mol−1 day−1 (PCPC-1) and (0.66 × 10−3 ± 2.50 × 10−5) cmol mol−1 day−1 (PCPC-2). The initial leakage rate of helium is a factor of 2 or 3 higher than that between the first and second checks. In contrast, for two BCPCs, no helium was detected in the liquid chamber with a limit of detection (LOD) of 1.6 μmol mol−1 even after ∼320 days after gravimetric preparation. The LOD was determined by dividing a signal-to-noise ratio of 3 by the analytical sensitivity of a primary standard gas mixture. The leaked amount-of-substance fractions of hydrocarbons into the gas chamber are summarized in Table 4. For two PCPCs, all hydrocarbons were detected in the gas chamber within less than 3 months after gravimetric preparation, indicating the commencement of hydrocarbon leakage to the gas chamber shortly after preparation. The detected amount-of-

Uncertainty Estimation. The amount-of-substance fraction of each substance and its uncertainty (Table 1) were Table 1. Amount-of-Substance Fractions and Expanded Uncertainties (k = 2) of Hydrocarbons in Gravimetrically Prepared BCPCs and PCPCs cylinder

substance

amount-of-substance fraction and expanded uncertainties (cmol mol−1)

BCPC-1

ethane propene propane isobutane 1-butene n-butane isopentane ethane propene propane isobutane 1-butene n-butane isopentane ethane propene propane isobutane 1-butene n-butane isopentane ethane propene propane isobutane 1-butene n-butane isopentane

2.161 ± 0.023 9.651 ± 0.022 69.737 ± 0.284 4.428 ± 0.023 3.444 ± 0.017 9.721 ± 0.057 0.823 ± 0.016 2.071 ± 0.030 9.765 ± 0.051 70.462 ± 0.369 3.854 ± 0.021 3.095 ± 0.024 9.877 ± 0.059 0.844 ± 0.017 2.045 ± 0.032 9.177 ± 0.054 70.338 ± 0.292 4.265 ± 0.031 3.265 ± 0.015 9.891 ± 0.078 0.987 ± 0.039 2.458 ± 0.033 8.188 ± 0.038 70.316 ± 0.198 4.338 ± 0.029 3.458 ± 0.021 10.206 ± 0.071 1.001 ± 0.039

BCPC-2

PCPC-1

PCPC-2

calculated based on results from the gravimetric preparation and purity analysis in accordance with ISO 6142−1.1 Uncertainties from each step were propagated and combined in accordance with the Guide to the Expression of Uncertainty in Measurement (GUM),9 and the calculations were performed with GUM Workbench Pro (Metrodata GmbH, Germany). For all GC analyses, PSMs traceable to SI units were analyzed before and after samples. Uncertainties in the results of GC analysis were calculated by combining the uncertainties from PSMs (calibration standards) and the repeatability of the GC analyses in accordance with GUM. A more detailed discussion of uncertainties in analytical procedures of chromatographic analysis can be found in ref 10.

■

RESULTS AND DISCUSSION The BCPC was designed to eliminate the leakage that affects commercial PCPCs; it was manufactured as shown in Figure 2. The bellows (A in Figure 2) were 7 cm in diameter and 30 cm

Table 2. Helium Amount-of-Substance Fractions (cmol mol−1) and Expanded Uncertainties (k = 2) in the Liquid Chamber for PCPCs cylinder

PCPC-1

PCPC-1

PCPC-2

PCPC-2

elapsed time (days)

68 0.1206 ± 0.0016

322 0.3692 ± 0.0075

90 0.1998 ± 0.0054

344 0.3674 ± 0.0032

11926

DOI: 10.1021/acs.analchem.7b03858 Anal. Chem. 2017, 89, 11924−11928

Technical Note

Analytical Chemistry

Table 3. Leakage Rate (cmol cmol−1 day−1) and Expanded Uncertainties (k = 2) of He from the Gas Chamber to the Liquid Chamber for PCPCs cylinder

PCPC-1

PCPC-1

PCPC-2

PCPC-2

elapsed time (days)

68 1.77 × 10−3 ± 2.61 × 10−5

322 0.98 × 10−3 ± 3.07 × 10−5

90 2.22 × 10−3 ± 2.47 × 10−5

344 0.66 × 10−3 ± 2.50 × 10−5

Table 4. Hydrocarbon Amount-of-Substance Fractions (μmol μmol−1) and Expanded Uncertainties (k = 2) Leaked from the Liquid Chamber to the Gas Chamber for PCPCs cylinder component

PCPC-1

elapsed time

ethane propane propene isobutane n-butane 1-butene isopentane

PCPC-1

PCPC-2

PCPC-2

57 days

320 days

79 days

342 days

12.0 ± 0.2 61.6 ± 1.1 31.6 ± 0.6 1.30 ± 0.03 24.2 ± 0.4 3.54 ± 0.06 81.9 ± 1.3

242 ± 2.3 2062 ± 15 802 ± 6.5 34.0 ± 0.6 143.0 ± 1.6 80.7 ± 1.1 100.9 ± 1.4

37.8 ± 0.6 187.3 ± 2.9 83.6 ± 1.3 2.15 ± 0.02 25.9 ± 0.4 8.65 ± 0.14 49.6 ± 0.8

297 ± 2.4 2131 ± 13 753 ± 5.9 36.6 ± 0.5 153.4 ± 1.0 94.1 ± 1.3 71.3 ± 1.0

Table 5. Leakage Rates (μmol μmol−1 day−1) of Hydrocarbons Uncertainties (k = 2) from the Liquid Chamber to the Gas Chamber for PCPCs component ethane propane propene isobutane n-butane 1-butene isopentane

cylinder

PCPC-1

PCPC-1

PCPC-2

PCPC-2

elapsed time

(0−57) days

(57−320) days

(0−79) days

(79−342) days

0.211 1.081 0.554 0.023 0.425 0.062 1.437

± ± ± ± ± ± ±

0.005 0.027 0.014 0.001 0.010 0.002 0.034

0.875 7.606 2.929 0.124 0.452 0.293 0.072

± ± ± ± ± ± ±

0.010 0.070 0.029 0.002 0.007 0.004 0.007

0.478 2.371 1.058 0.0272 0.328 0.109 0.628

± ± ± ± ± ± ±

0.010 0.047 0.021 0.0004 0.007 0.002 0.013

0.986 7.390 2.545 0.131 0.485 0.325 0.083

± ± ± ± ± ± ±

0.011 0.064 0.027 0.002 0.005 0.005 0.005

modification of a proven high-precision automatic cylinder weighing system.5,6 Leakage evaluation tests showed that the BCPCs had no leakage in either direction between the liquid and gas chambers, whereas the PCPCs suffered from significant leakage between the chambers. The amount-of-substance fraction of helium leaked into the liquid chamber reached ∼0.4 cmol mol−1 about 10 months after preparation. The amount-of-substance fraction of hydrocarbons leaking to the gas chamber ranged from ∼30 to ∼2013 μmol mol−1. Of the seven considered hydrocarbons, propane leaked the fastest, followed by propene, ethane, n-butane, 1-butene, isobutane, and isopentane. Given that the stability of CRMs is a major source of uncertainty in the development of accurate CRMs, it is therefore essential to maintain the gravimetrically determined amount-of-substance fractions in them to allow the distribution of accurate and traceable standards. The results indicate that the newly developed leak-free BCPC can help reduce uncertainties in developing certified reference materials of liquid hydrocarbon mixtures as no leakage is found in the BCPC.

substance fractions were much higher with the GC-FID LOD (0.05 μmol mol−1). The LOD was determined by dividing a signal-to-noise ratio of 3 by the analytical sensitivity of a primary standard gas mixture. The amount-of-substance fractions of all hydrocarbons increased significantly 263 days after the first leak check. For example, propane in PCPC-1 increased from (61.6 ± 1.1) μmol mol−1 at the first check to (2062 ± 15) μmol mol−1 at the second, 263 days later. Leakage rates for all hydrocarbons are summarized in Table 5. In contrast to helium, the initial leakage rates of all hydrocarbons are much lower than those between the first and second check, except for isopentane. For example, the initial leakage rate of propane (PCPC-1) until the first leak check was estimated as ∼1.1 μmol mol−1 day−1, while that between the first and second check was estimated as ∼7.6 μmol mol−1 day−1. Propane has the highest leakage rate followed by propene, ethane, n-butane, 1-butene, isobutane, and isopentane, which follows the order of the partial vapor pressure estimated by the amount-ofsubstance fraction and vapor pressure.

■

■

CONCLUSION A new bellows-type constant-pressure cylinder (BCPC) was designed and constructed to reduce the leakage found in the commercial piston-type constant-pressure cylinders (PCPCs) used for certified reference materials of hydrocarbon liquid mixtures. Characteristics of penetration between chambers were compared for the two cylinder types. The liquid chamber of the newly developed BCPC had a capacity of ∼700 mL, which can be contracted to ∼200 mL. A base plate installed at the bottom of the cylinder facilitated accurate weighing without any

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03858. Impurities in hydrocarbon reagents; schematic showing two chambers in a bellows-type constant-pressure cylinder before assembly; gas and liquid feeding system 11927

DOI: 10.1021/acs.analchem.7b03858 Anal. Chem. 2017, 89, 11924−11928

Technical Note

Analytical Chemistry

■

for preparing liquid hydrocarbon mixture in constantpressure cylinders (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] ORCID

Sangil Lee: 0000-0001-7912-2841 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by grants from the Korea Research Institute of Standards and Science under the Basic R&D Project of Establishment of National Gas Analysis Measurement Standards and Improvement of Calibration and Measurement (17011010).

■

REFERENCES

(1) Benesch, R.; Jacksier, T. Anal. Chem. 2001, 73, 379−383. (2) International Organization for Standardization. Gas analysis Preparation of calibration gas mixturesPart 1: Gravimetric method for Class I mixtures; ISO 6142−1:2015; International Organization for Standardization: Geneva, Switzerland, 2015. (3) Brown, A. S.; Downey, M. L.; Milton, M. J. T.; van der Veen, A. M. H.; Zalewska, E. T.; Li, J. Metrologia 2013, 50, 08015. (4) Bureau International des Poids et Measures. CCQM-K119. The BIPM Key Comparison Database [Online], 2015. http://kcdb.bipm. org/AppendixB/KCDB_ApB_info.asp?cmp_idy=1363&cmp_cod= CCQM%2DK119&page= (accessed September 2017). (5) Milton, M. J. T.; Vargha, G. M.; Brown, A. S. Metrologia 2011, 48, R1−R9. (6) Kim, M. E.; Kim, Y. D.; Kang, J. H.; Heo, G. S.; Lee, D. S.; Lee, S. Talanta 2016, 150, 516−524. (7) Min, D.; Kang, N.; Moon, D. M.; Lee, J. B.; Lee, D. S.; Kim, J. S. Talanta 2009, 80, 422−427. (8) Lim, J. S.; Lee, J.; Moon, D.; Tshilongo, J.; Qiao, H.; Shuguo, H.; Tiqiang, Z.; Kelley, M.; Rhoderick, G. C.; Konopelko, L. A.; Kolobova, A. V.; Vasserman, I. I.; Zavyalov, S. V.; Gromova, E. V.; Efremova, O. V. Metrologia 2017, 54, 08017. (9) Joint Committee for Guides in Metrology. Evaluation of Measurement DataGuide to the Expression of Uncertainty in Measurement; JCGM 100:2008 (GUM 1995 with minor corrections); Joint Committee for Guides in Metrology, 2008. (10) Konieczka, K.; Namiesnik, J. J. Chromatogr. A. 2010, 1217, 882− 891. (11) DNV GL. DNV GL Home Page. https://www.dnvgl.com (accessed October 2017).

11928

DOI: 10.1021/acs.analchem.7b03858 Anal. Chem. 2017, 89, 11924−11928