- Email: [email protected]
Selective gas-chromatographic detection using an ion-selective electrode—II1
~auuwa, wrz,
vol.
19, pp. 539 to 547.
P:rtwmozt Press. Printed In. Northem
Inland
SELECTIVE GAS-CHROMATOGRAPHIC DETECTION USING AN ION-SELECTIVE ELECTRODE-II* SELECTIVE
DETECTION
OF FLUORINE
COMPOUNDS
TSUGIO KOJIMA, MITWJNOJO ICHJ.SE and YOSHWTSU SEO Department of Industrial Chemistry, Faculty of Engineering, Kyoto University, Kyoto, Japan (Received 7 June 1971. Accepted 17 September 1971)
Snmmary-Components in samples are separated on a gas chromatography column using hydrogen as carrier gas. The individual wmportents from the cohnnn are passed through a platinum tube heated at 1000°, where they undergo hydrogenulysis, and fluorine compounds are converted into hydrogen fluoride. The hydrogen tiuoride is dissolved in a slow stream of an absorption solution, and the fluoride ion concentration in the resulting solution is monitored in a flow-ceil with a fluoride ion electrode. The potentiometr~ output of the cell is converted into a signal, which is proportional to the concentration of fluoride ion, by an antilogarithmic converter, and rewrded. The response of the detector to fluorine compounds was about 10,000 times that to an equal quantity of other organic wmpounds, and 5 x 1O-rt mole of fluorobenzene could be detected. SEPARATE components from a gas chromatographic column are generally identified by determining the retention value. In practice, however, identification by the retention value alone is difficult and the final confirmation is often possible only after information on the elemental composition, functional groups and structure is obtained by simultaneously using other analytical means. One method used for this purpose is to transfer the eluates of pure components from the chromatograph into a spectrometer (infrared, mass or nuclear magnetic resonance), which thus produces the spectrum of each pure component in turn. The most effective is the combination of gas chromatograph and mass spectrometer. Another method, which dispenses with expensive instruments uses the retention value and additional info~ation obtained from the gas chromatogram above. Volatile injected compounds are reacted with non-volatile reagents to form nonvolatile products in the gas chromatographic system and a comparison of chromatograms run in the presence and absence of the reagent is used to identify the kind of functional group or the structure of the carbon skeleton. Another method of this type is based on the selective detection of the compound containing a specific element or having a specific functional group, by means of a selective detector. We have searchedI+ for a method of this type which wilf give information on the elemental composition as well as the retention value. This is achieved by the selective detection of compounds containing halogens, sulphur, nitrogen and oxygen, with a reaction column and a selective detector. A new detection system for a gas chromatograph was recently devised, using an ion-selective electrode as a detecting element.* This system will not only detect organic compounds selectively but also furnish information on the composition.
* Part I: Bumeki Kagaku, 1971,20,20. 539
540
Tsuo~o KOJJMA,MITSUNOJO ICHISB and YOSHIMITSU SEO
It is difficult to detect fluorine compounds selectiveIy, and the only method6 known so far is to monitor the intensity of the atomic emission spectrum of fluorine, using a d.c. discharge detector. The present method uses a simpler apparatus. EXPERIMENTAL Principle of the method
The sample components are separated on a gas chromatographic cohmm using hydrogen as carrier gas and the effluent is passed through a platinum tube at fOOO”wbere the separated components undergo hydrogenolysis and the fluorine in the fluorine compound is converted into hydrogen fluoride. The decomposition gases are introduced into an absorption tube where solution (flowing at a constant rate) absorbs the hydrogen tluoride. The solution emerging from the absorption tube is passed into a micro-cell equipped with a fluoride ion electrode. Changes in the fluoride ion concentration in the solution are detected by the co~~~nd~g changes in the ion electrode potential.‘ When the ionic strength is kept constant, a Nernst relationship exists between the electrode potential and the logarithm of the ion concentration. The difference iu potential between the ion electrode and a reference electrode is fed to an antilog converter circuit and a signal directly proportional to the ion concentration is recorded. A chromatogram composed of only the peaks due to fluorine compounds is obtained. Apparattrcsandprocedure
A schematic diagram is shown in Fig. 1: (a) is the carrier gas source (hydrogen at a flow-rate of 25-48 mllmin), (b) and (c) represent a Yanagimoto Model GCG-II gas chromatograph, and (d) is an electric furnace, 200 mm in length, for elemental analysis which was operated at 1000”. A platinum
iquid
FIG. l.-Schematic
diagram of the apparatus.
tube, 2 mlongand 1.5 mminside~ameter, was used as a pyroiysis tube (e); (f} is a glass absorption tube consisting of a gas-liquid contact area, 1 mm inside diameter and 0.15 m high, and a gas-liquid separation area, 13 mm inside diameter and 25 mm high. The pyrolysis tube and absorption tube are co~ected with 1 mm bore polyethylene tubing. Absorption solution (g) is pumped by a micropump (h) at a constant flow-rate (O-6mlfmin) via a stream buffer (i) to the absorption tube where it absorbs hydrogen fluoride etc from the hydrogenolysis gases. The unabsorbed gases such as methane are separated, together with the carrier gas, from the absorption solution in the gas-liquid separator and vented from the system. The emerging absorption solution is led through a measuring cell equipped with a fluoride ion electrode (j) as a detecting element and is then discarded. The absorption sotution and the reference electrode (Q (a sleevetype saturated calomel electrode) are connected by means of a saturated potassium chloride solution-agar bridge. The measuring cell is shown in detail in Fig. 2. Since the volume of the cell is about @OSml, the dead time is about 5.0 set when the flow rate of the absorption solution is 0.6 ml/min. The fluoride ion electrode used was an Orion Model 94-09A constructed from a single crysta1 membrane of lanthanum 5uoride and it shows a potential E for a fluoride ion activity a,_ in the vicinity of the electrode, in accordance with the Nernst equation as follows: F = EO - 2.303 Flog
a,_
(0
Selective gas-chromatographic detection
541
FIG. 2.---&h detaiks. (a) Fluoride ion electrode body; (b) silicone rubber packing; (c) polyethylene plate; (4 glass capillary If the ionic strength of the abso~tion sotution is kept constant, a negative potential proportional to the logarithm of the fiuoride ion ~n~nt~tion C,.. is obtained: E = E,’ - 0.0592 log C,_
(at 257
(2)
The circuit for measuring the electrode potential is shown in Fig. 3, where (a) is a voltage follower Using a @T-2a amplifier ~~b~~/Nexus Research) as an impedance converter, (b) is an adjustablecoefficient inverter (Philbrick PF-85AU inversion type negative feedback coupling) as an adapter for (c), an antilog converter, (Philbrick/Nexus Model 4351); (df is a 1-mV recorder (Shimadxu R201) and (e) is the signal input.
(a AE
FIG, 3i--Circuit for ion electrode detector, In order to use the tluoride ion electrode in the Nernstian range, a small amount of sodium fluoride was added to the absorption solution to make the inanition of fluoride ion approximately IO-ah#. If this concentration of fluoride ion is C,, the electrode potential is E=E,‘-
0.0592 Iog C,
The potential E’ when h drogen fluoride is absorbed in thii soIu~on and the ~n~ntmtion fluoride ion is increased by drC is given by E’ = &,’ _ 0*0592log (C, + AC) 11
(3) of (4)
542
TSUGSO KOJJMA,MRWNOJO
Icmsa and Y-mu
SEO
From equations (3) and (4), the change in the potential, AE, corresponding to the change in the fluoride ion concentration is given by AE = (E’ - E) = -0*0592 log (I + $$)
IS)
In this apparatus, the potential change is recorded after antilog conversion to give an output signal proportional to the concentration, as in the case of an ordinary detector, the antiIog converter used having the following conversion characteristics:
a=-i6*1
loo’%,,
(6)
The change in potential, AE, enters as input efll fo the anti@ converter (f) after undergo& zfedz conversion in the voltage follower (u) and 1s then ampltfied A trmes m the amplmer (&). , e,, = -A.hE (7) The following equation is derived from equations (5), (6) and (7): e. = -
A. 1O-o*6,A
. AE
logt--d = --I+ 0~5.A.O*0592log (l + $f) Adjustment of the gain A inthe amplifier @I in such a manner as to make the coefficient of the second term of the right-hand side of equation (8) unity, that is to make A = 33.78, gives the relationship
and the output e. of the antilog converter begins to change in direct proportion to the change in the fluoride ion concentration in the absorption solution. RESULTS
AND
DISCUSSION
Samples containing hydrocarbons, nitrogen and oxygen compounds in addition to fluorine compounds were prepared and the possibility of selective detection of fluorine compounds was investigated.
In Fig. 4 is shown a chromatogram obtained by injecting into the column O*l0.2 ~1 of sample consisting of a mixture of O*Olmole each of ethylmercaptan, n-octane, 1,Zdibromoethane and 1-nitropropane and 0.1 mmole of ethyl trifluoroacetate. In Fig. 4, (a) is an ordinary non-selective chromatogram obtained by passing the effluent component from the separation column through a hydrogen flame ionization detector while by-passing the hydrogenolysis tube, and the number in parentheses indicates the mole ratio; (b) is a chromatogram obtained by subjecting the same sample to hydrogenolysis absorbing the decomposition gas in the TISAB solution’ (a mixture of 57 ml of glacial acetic acid, 58 g of sodium chloride and O-3 g of sodium citrate, diluted to 2 1.) which is about 1ObM in sodium fluoride as mentioned above. The ethyl ~fluoroace~te is selectively detected in the measuring cell. In measurements of fluoride ion concentration with the aid of the fluoride ion electrode, it should be noted that the interfering OH- ions will show their effect as the pH of the solution rises whereas the apparent F- concentration will decrease owing to formation of HF or HF,- as the pH drops. The useful range of pH is 5-7
Selective gas-cbromatographicdetection n-C,H,,
543
( I1
x6 C,H,SH
( I)
xi
(0)
CH2BrCH2f3r
Ii
CF,COOC,H,
0
( I 1
(0.01)
( I ; Mole ratio
IO
5 min
FW. 4.-Chromatogam of a test mixture. Column-Durapak-Carbowax 4OO/PorasilC (100-120 mesh), 1 m; temperature-90”; carrier gas-H, 28 ml/m& (a) Non-selective detection by use of FID; (b) selective detection of fluotie compound by use of fluoride ion electrode detector.
when the concentration of F- is lObM, according to Bock,g and a TISAB solution of pH 5 is used as an absorption solution. By use of the same procedure, O-1mmole in 1,1,2,2-tetrachloro-1 ,Zdifluoroethane was detected selectively in a mixture with O-01 mole each of ethyl iodide, n-propyl acetate, 1,l ,Ztrichloroethane and ethyl thiocyanate. Aromatic compounds A chromatogram of a mixture of O-01 mole each of ethylbenzene, bromobenzene, benzonitrile and thioanisole and O-1mmole of o-fluorotoluene is shown in Fig. 5. For further confirmation, a mixture of O-01 mole each of o-xylene, o-chlorotoluene, methyl benzoate, phenyl isothiocyanate and 0.1 mmole of benzotrifluoride was chromatographed. It was confirmed from these experiments that selective detection of fluorine compounds is complete.
544
TSUGIO KOIIMA, M~UNOJO Icmse and YOSHIMITSUSao
Quantitative results
The results shown in Fig. 4 and Fig. 5 indicated that fluorine compounds could be detected selectively, and the quantitativeness of response of the detector was next investigated. The results are shown in Table I.
(b)
0
I
I
5
IO
min
FIG. 5.-Chromatogram of a test mixture. Column temperature 128” ; seeFig. 4 for other conditions. (a) Non-selective detection by use of FID, (6) selective detection of fluorine compound by use of fluoride ion
electrode detector. Fluorobenzene, o-fluorotoluene, and benzotrifluoride were chosen as the first group of compounds; a sample containing 0.01 mole each of these compounds was prepared, diluted lO,OOO-foldwith n-hexane, and 2 ~1 were injected into a column (Durapak-Carbowax 4OO/Porasil C, lo@-120 mesh, 1 m, 609, and chromatographed. The relative size of each peak area was determined by the method of cutting out the recorded peaks and weighing the paper. Taking benzotrifluoride as the standard, the ratio of the peak area of each of two other components to the peak area of the standard was measured repeatedly. 1,1,2,2-Tetrachloro-l , I-difluoroethane, o-fluorotoluene, and p-chlorofluorobenzene were chosen as the second group of compounds; a sample containing 0.01 mole each of these compounds and diluted lO,OOO-fold with n-hexane, was chromatographed as before. The same procedure as for the first
Average
Ratio of peak areas
atoms
flUOlilX
Number of
Mole ratio
Sample
0323
0,322
1
1
0.32 0.32 O-32 0.34 0.31 0.32 0.33 o-33 O-32 0.32
1
1
0.33 0*36 0.32 0.31 O-32 ’ 0*3f 0.31 0+31 @33 iO.32
ecu
Fist group
1
1
I
O-470
0.49 0.48 o-41 o-47 051 o-50 045 044 O-49 O-46
1
1
O-985
X:Z 0.50 0.51 O-45 0.47 0.48 0.47 0.478
14
G&&F*
1.02 0.98 0.99 o-93 o-99 0.99 0.97 l-04 0.98 O-96
,
0+49 0.49
1
1
“4
TABLE ~.---COMPARISON OF RESPONSES OBTAINED PROM SEVERAL FLUORINB COMPOUNDS
1
6
t
Third group
1.008
::t 1.03 1.05 O-97 0.97
0.96
1.02 1.03 0‘97
2
1
F
‘, 0
\F
.
Fpa a -c. g
!!I 3
I
I!
f 0. 8
546
TSUGIOKOJIMA, Mn-smOlO
ICHlsH and Yosmhmsu
SE0
group was repeated with 1,1,2,2-tetrachloro-1,2-~fluoroe~ane as the standard. Perfluoro-compounds were chosen as the third group of compounds; a sample containing l/350 mole of perfluoromethylcyclohexane, l/l50 mole of perfluorobenzene and l/50 mole of m-difluorobenzene was prepared and diluted 1000-fold with n-hexane, and 0.8 pi was chromatographed, The perfluorobenzene was used as the standard. If the hydrogenolysis reaction, the absorption of hydrogen fluoride, and the response of the detector to fluoride ions are quantitative, the ratio of the peak areas of the three components for each group should be proportional to the number of fluorine atoms in each molecule, The average of the experimentally determined ratio of peak areas is close to the theore~cal value. It is thus shown that ftuorine bonded to an aromatic nucleus (which has been said to be particularly resistant to quantitative decomposition), pertluoro-compounds of the third group, and compounds containing other halogens in addition to fluorine can be deter~ned quantitatively. As mentioned previously, a propo~ionality to the fluoride ion concentration in the absorption solution should exist in the present apparatus in the range where the ionic strength of the absorption solution is constant. To confirm this, experiments were carried out with fluoro~nzene as a sample. The output e, of the antilog converter is confined to the range from 0 to -10 V. Therefore, supposing the concentration of the fluoride ion added to the absorption solution is IBM, the permissible range of the fluoride ion concentration in the absorption solution becomes 10-S-lOSM from equation (9). The ionic strength p of the absorption solution is O-820, and this changes merely to 0,821 when the fluoride ion concentration reaches lW_4& upon absorption of hydrogen fluoride. As a consequence, the changes in the activity coefficient of the fluoride ion may be ignored. Moreover, the changes in the pH are negligible. Therefore, if the amount of sample is adjusted so that the fluoride ion concentration falls within the range lo”-lO-sM, there should exist a linear relationship between the amount of sample injected and the peak area. Accordingly, solutions of fluorubenzene in n-hexane with a concentration of 04001-0~05M were prepared, 2 ~1 of each were injected into the column, and the relationship between the amount of sample injected and the peak area was examined. An absolute calibration curve was obtained and the existence of a linear relationship between the amount of sample injected (2 x 10-10-10-7 mole) and the peak area in the concentration range indicated above was found. CONCLUSION
A gas c~omatographic detector with using an ion-selective electrode has been developed and applied to the selective detection of fluorine compounds. It has the following advantages : the limit of detection is 5 x lO--lr mole (for fluorobenzene), the detector sensitivity is high, the degree of selectivity is 10,OOQor more with an extremely high sensitivity, the drift of the base line is O-02 mVj’hr with good stability. The response time of the detector is about 10 set when the concentration of Fis about lO-sJ4 and several seconds when it is more than IO-*M. Such slow responses distort the early peaks slightly but the effect is negligible for peaks having retention times greater than 5 min. A platinum tube was used as a hydrogenolysis tube in the present study, but could be replaced by a quartz tube.
Selective gas-chromatographic detection Zusammenfassnng-Die Proben werden an einer Gasehromatographiesriule mit Wasserstoff als Triigergas in ihre Komponenten getrennt. Die einzelnen von der S&de kommenden Komponenten werden durch ein auf 1000” geheiztes Platinrohr geleitet; dort werden sie hydrierend gespalt&. Fluorverbindung& werden in Fluorwasserstoff ~~~ef~rt. Der Fluo~~~~toff wird in einem langsamen Strom e&r Absorptionsliisung gelost und die Fluoridko~tmtion in der L&mg in einer Durc~~zelle mit einer Fluoridionen-Elektrode verfolgt. Das potentiometrische Verhalten der Zelle wird mit Hilfe eines antilo~~t~i~hen Umsetzers in ein Signal umgesetzt, das der Fluoridkonzentration proportional ist, und retistriert. Der Detektor sorach auf Fiuo~erb~d~~~ etwa 10 000&l starker an als auf eine gleiche Menge anderer or&nischer 5 s 1O-*1Mel Fluorbenzol komxten nachgewiesen Verbindungen; werden. R&tune--& &pare des composants dans des &hantillons sur une colonne de chromatographie en phase gaxense en u&ant l’hydrogi?ne comme gaz vecteur. On fait passer les composants individuels provenant de la colome a travers un tube de platine chat&5 a IOOO”, oli ils subissent ~hydro~nol~, et les composes Auor& sent convertis en acide fluorhydrique. L’acide fluorhydrique est dissous darts un lent courant dune solution &absorption, et la concentration en ion fluornre dans la solution r&&ante est contr6lBe dans une cellule a &oulement avec une electrode a ion fluorure. LQnission potentiom~~que de la cellule est convertie en un signal, qui eat proportionnel a la concentration en ion fluorure, par un conve~~ur ant~og~it~ique, et enregistre. La reponse du detecteur aux composes fluores est d’environ 10000 fois celle fonrnie par tine Cgale quantite d’autres composes organiques, et l’on a pu detecter 5 x lo-l1 mole de fluorobenzene. REFERENCES 1. W. Funasaka, T. Kojima and Y. Seo, Bunseki Kagaku, 1968,17,464. 2. T. Koiima. Y. Seo and J. Nishida. ibid.. 1968.17.1496. 3. T. Ko&& M, Ichise and Y. Seo,Sibid.,S1969,.18,~1460. 4. Idem, ibid., 1971,20,20. 5. R. S. Braman and A. Dynako, Anal. Chem., 1968,40,95.
6. E. Pungor, i&d., 1967,39 (13), 28A. 7. M. S. Frant and J. W. Ross, ibid., 1968,48,1169. 8. R. Bock and S. Strecker, 2. Anal. Chem., 1968,2X5,322.
547
vol.
19, pp. 539 to 547.
P:rtwmozt Press. Printed In. Northem
Inland
SELECTIVE GAS-CHROMATOGRAPHIC DETECTION USING AN ION-SELECTIVE ELECTRODE-II* SELECTIVE
DETECTION
OF FLUORINE
COMPOUNDS
TSUGIO KOJIMA, MITWJNOJO ICHJ.SE and YOSHWTSU SEO Department of Industrial Chemistry, Faculty of Engineering, Kyoto University, Kyoto, Japan (Received 7 June 1971. Accepted 17 September 1971)
Snmmary-Components in samples are separated on a gas chromatography column using hydrogen as carrier gas. The individual wmportents from the cohnnn are passed through a platinum tube heated at 1000°, where they undergo hydrogenulysis, and fluorine compounds are converted into hydrogen fluoride. The hydrogen tiuoride is dissolved in a slow stream of an absorption solution, and the fluoride ion concentration in the resulting solution is monitored in a flow-ceil with a fluoride ion electrode. The potentiometr~ output of the cell is converted into a signal, which is proportional to the concentration of fluoride ion, by an antilogarithmic converter, and rewrded. The response of the detector to fluorine compounds was about 10,000 times that to an equal quantity of other organic wmpounds, and 5 x 1O-rt mole of fluorobenzene could be detected. SEPARATE components from a gas chromatographic column are generally identified by determining the retention value. In practice, however, identification by the retention value alone is difficult and the final confirmation is often possible only after information on the elemental composition, functional groups and structure is obtained by simultaneously using other analytical means. One method used for this purpose is to transfer the eluates of pure components from the chromatograph into a spectrometer (infrared, mass or nuclear magnetic resonance), which thus produces the spectrum of each pure component in turn. The most effective is the combination of gas chromatograph and mass spectrometer. Another method, which dispenses with expensive instruments uses the retention value and additional info~ation obtained from the gas chromatogram above. Volatile injected compounds are reacted with non-volatile reagents to form nonvolatile products in the gas chromatographic system and a comparison of chromatograms run in the presence and absence of the reagent is used to identify the kind of functional group or the structure of the carbon skeleton. Another method of this type is based on the selective detection of the compound containing a specific element or having a specific functional group, by means of a selective detector. We have searchedI+ for a method of this type which wilf give information on the elemental composition as well as the retention value. This is achieved by the selective detection of compounds containing halogens, sulphur, nitrogen and oxygen, with a reaction column and a selective detector. A new detection system for a gas chromatograph was recently devised, using an ion-selective electrode as a detecting element.* This system will not only detect organic compounds selectively but also furnish information on the composition.
* Part I: Bumeki Kagaku, 1971,20,20. 539
540
Tsuo~o KOJJMA,MITSUNOJO ICHISB and YOSHIMITSU SEO
It is difficult to detect fluorine compounds selectiveIy, and the only method6 known so far is to monitor the intensity of the atomic emission spectrum of fluorine, using a d.c. discharge detector. The present method uses a simpler apparatus. EXPERIMENTAL Principle of the method
The sample components are separated on a gas chromatographic cohmm using hydrogen as carrier gas and the effluent is passed through a platinum tube at fOOO”wbere the separated components undergo hydrogenolysis and the fluorine in the fluorine compound is converted into hydrogen fluoride. The decomposition gases are introduced into an absorption tube where solution (flowing at a constant rate) absorbs the hydrogen tluoride. The solution emerging from the absorption tube is passed into a micro-cell equipped with a fluoride ion electrode. Changes in the fluoride ion concentration in the solution are detected by the co~~~nd~g changes in the ion electrode potential.‘ When the ionic strength is kept constant, a Nernst relationship exists between the electrode potential and the logarithm of the ion concentration. The difference iu potential between the ion electrode and a reference electrode is fed to an antilog converter circuit and a signal directly proportional to the ion concentration is recorded. A chromatogram composed of only the peaks due to fluorine compounds is obtained. Apparattrcsandprocedure
A schematic diagram is shown in Fig. 1: (a) is the carrier gas source (hydrogen at a flow-rate of 25-48 mllmin), (b) and (c) represent a Yanagimoto Model GCG-II gas chromatograph, and (d) is an electric furnace, 200 mm in length, for elemental analysis which was operated at 1000”. A platinum
iquid
FIG. l.-Schematic
diagram of the apparatus.
tube, 2 mlongand 1.5 mminside~ameter, was used as a pyroiysis tube (e); (f} is a glass absorption tube consisting of a gas-liquid contact area, 1 mm inside diameter and 0.15 m high, and a gas-liquid separation area, 13 mm inside diameter and 25 mm high. The pyrolysis tube and absorption tube are co~ected with 1 mm bore polyethylene tubing. Absorption solution (g) is pumped by a micropump (h) at a constant flow-rate (O-6mlfmin) via a stream buffer (i) to the absorption tube where it absorbs hydrogen fluoride etc from the hydrogenolysis gases. The unabsorbed gases such as methane are separated, together with the carrier gas, from the absorption solution in the gas-liquid separator and vented from the system. The emerging absorption solution is led through a measuring cell equipped with a fluoride ion electrode (j) as a detecting element and is then discarded. The absorption sotution and the reference electrode (Q (a sleevetype saturated calomel electrode) are connected by means of a saturated potassium chloride solution-agar bridge. The measuring cell is shown in detail in Fig. 2. Since the volume of the cell is about @OSml, the dead time is about 5.0 set when the flow rate of the absorption solution is 0.6 ml/min. The fluoride ion electrode used was an Orion Model 94-09A constructed from a single crysta1 membrane of lanthanum 5uoride and it shows a potential E for a fluoride ion activity a,_ in the vicinity of the electrode, in accordance with the Nernst equation as follows: F = EO - 2.303 Flog
a,_
(0
Selective gas-chromatographic detection
541
FIG. 2.---&h detaiks. (a) Fluoride ion electrode body; (b) silicone rubber packing; (c) polyethylene plate; (4 glass capillary If the ionic strength of the abso~tion sotution is kept constant, a negative potential proportional to the logarithm of the fiuoride ion ~n~nt~tion C,.. is obtained: E = E,’ - 0.0592 log C,_
(at 257
(2)
The circuit for measuring the electrode potential is shown in Fig. 3, where (a) is a voltage follower Using a @T-2a amplifier ~~b~~/Nexus Research) as an impedance converter, (b) is an adjustablecoefficient inverter (Philbrick PF-85AU inversion type negative feedback coupling) as an adapter for (c), an antilog converter, (Philbrick/Nexus Model 4351); (df is a 1-mV recorder (Shimadxu R201) and (e) is the signal input.
(a AE
FIG, 3i--Circuit for ion electrode detector, In order to use the tluoride ion electrode in the Nernstian range, a small amount of sodium fluoride was added to the absorption solution to make the inanition of fluoride ion approximately IO-ah#. If this concentration of fluoride ion is C,, the electrode potential is E=E,‘-
0.0592 Iog C,
The potential E’ when h drogen fluoride is absorbed in thii soIu~on and the ~n~ntmtion fluoride ion is increased by drC is given by E’ = &,’ _ 0*0592log (C, + AC) 11
(3) of (4)
542
TSUGSO KOJJMA,MRWNOJO
Icmsa and Y-mu
SEO
From equations (3) and (4), the change in the potential, AE, corresponding to the change in the fluoride ion concentration is given by AE = (E’ - E) = -0*0592 log (I + $$)
IS)
In this apparatus, the potential change is recorded after antilog conversion to give an output signal proportional to the concentration, as in the case of an ordinary detector, the antiIog converter used having the following conversion characteristics:
a=-i6*1
loo’%,,
(6)
The change in potential, AE, enters as input efll fo the anti@ converter (f) after undergo& zfedz conversion in the voltage follower (u) and 1s then ampltfied A trmes m the amplmer (&). , e,, = -A.hE (7) The following equation is derived from equations (5), (6) and (7): e. = -
A. 1O-o*6,A
. AE
logt--d = --I+ 0~5.A.O*0592log (l + $f) Adjustment of the gain A inthe amplifier @I in such a manner as to make the coefficient of the second term of the right-hand side of equation (8) unity, that is to make A = 33.78, gives the relationship
and the output e. of the antilog converter begins to change in direct proportion to the change in the fluoride ion concentration in the absorption solution. RESULTS
AND
DISCUSSION
Samples containing hydrocarbons, nitrogen and oxygen compounds in addition to fluorine compounds were prepared and the possibility of selective detection of fluorine compounds was investigated.
In Fig. 4 is shown a chromatogram obtained by injecting into the column O*l0.2 ~1 of sample consisting of a mixture of O*Olmole each of ethylmercaptan, n-octane, 1,Zdibromoethane and 1-nitropropane and 0.1 mmole of ethyl trifluoroacetate. In Fig. 4, (a) is an ordinary non-selective chromatogram obtained by passing the effluent component from the separation column through a hydrogen flame ionization detector while by-passing the hydrogenolysis tube, and the number in parentheses indicates the mole ratio; (b) is a chromatogram obtained by subjecting the same sample to hydrogenolysis absorbing the decomposition gas in the TISAB solution’ (a mixture of 57 ml of glacial acetic acid, 58 g of sodium chloride and O-3 g of sodium citrate, diluted to 2 1.) which is about 1ObM in sodium fluoride as mentioned above. The ethyl ~fluoroace~te is selectively detected in the measuring cell. In measurements of fluoride ion concentration with the aid of the fluoride ion electrode, it should be noted that the interfering OH- ions will show their effect as the pH of the solution rises whereas the apparent F- concentration will decrease owing to formation of HF or HF,- as the pH drops. The useful range of pH is 5-7
Selective gas-cbromatographicdetection n-C,H,,
543
( I1
x6 C,H,SH
( I)
xi
(0)
CH2BrCH2f3r
Ii
CF,COOC,H,
0
( I 1
(0.01)
( I ; Mole ratio
IO
5 min
FW. 4.-Chromatogam of a test mixture. Column-Durapak-Carbowax 4OO/PorasilC (100-120 mesh), 1 m; temperature-90”; carrier gas-H, 28 ml/m& (a) Non-selective detection by use of FID; (b) selective detection of fluotie compound by use of fluoride ion electrode detector.
when the concentration of F- is lObM, according to Bock,g and a TISAB solution of pH 5 is used as an absorption solution. By use of the same procedure, O-1mmole in 1,1,2,2-tetrachloro-1 ,Zdifluoroethane was detected selectively in a mixture with O-01 mole each of ethyl iodide, n-propyl acetate, 1,l ,Ztrichloroethane and ethyl thiocyanate. Aromatic compounds A chromatogram of a mixture of O-01 mole each of ethylbenzene, bromobenzene, benzonitrile and thioanisole and O-1mmole of o-fluorotoluene is shown in Fig. 5. For further confirmation, a mixture of O-01 mole each of o-xylene, o-chlorotoluene, methyl benzoate, phenyl isothiocyanate and 0.1 mmole of benzotrifluoride was chromatographed. It was confirmed from these experiments that selective detection of fluorine compounds is complete.
544
TSUGIO KOIIMA, M~UNOJO Icmse and YOSHIMITSUSao
Quantitative results
The results shown in Fig. 4 and Fig. 5 indicated that fluorine compounds could be detected selectively, and the quantitativeness of response of the detector was next investigated. The results are shown in Table I.
(b)
0
I
I
5
IO
min
FIG. 5.-Chromatogram of a test mixture. Column temperature 128” ; seeFig. 4 for other conditions. (a) Non-selective detection by use of FID, (6) selective detection of fluorine compound by use of fluoride ion
electrode detector. Fluorobenzene, o-fluorotoluene, and benzotrifluoride were chosen as the first group of compounds; a sample containing 0.01 mole each of these compounds was prepared, diluted lO,OOO-foldwith n-hexane, and 2 ~1 were injected into a column (Durapak-Carbowax 4OO/Porasil C, lo@-120 mesh, 1 m, 609, and chromatographed. The relative size of each peak area was determined by the method of cutting out the recorded peaks and weighing the paper. Taking benzotrifluoride as the standard, the ratio of the peak area of each of two other components to the peak area of the standard was measured repeatedly. 1,1,2,2-Tetrachloro-l , I-difluoroethane, o-fluorotoluene, and p-chlorofluorobenzene were chosen as the second group of compounds; a sample containing 0.01 mole each of these compounds and diluted lO,OOO-fold with n-hexane, was chromatographed as before. The same procedure as for the first
Average
Ratio of peak areas
atoms
flUOlilX
Number of
Mole ratio
Sample
0323
0,322
1
1
0.32 0.32 O-32 0.34 0.31 0.32 0.33 o-33 O-32 0.32
1
1
0.33 0*36 0.32 0.31 O-32 ’ 0*3f 0.31 0+31 @33 iO.32
ecu
Fist group
1
1
I
O-470
0.49 0.48 o-41 o-47 051 o-50 045 044 O-49 O-46
1
1
O-985
X:Z 0.50 0.51 O-45 0.47 0.48 0.47 0.478
14
G&&F*
1.02 0.98 0.99 o-93 o-99 0.99 0.97 l-04 0.98 O-96
,
0+49 0.49
1
1
“4
TABLE ~.---COMPARISON OF RESPONSES OBTAINED PROM SEVERAL FLUORINB COMPOUNDS
1
6
t
Third group
1.008
::t 1.03 1.05 O-97 0.97
0.96
1.02 1.03 0‘97
2
1
F
‘, 0
\F
.
Fpa a -c. g
!!I 3
I
I!
f 0. 8
546
TSUGIOKOJIMA, Mn-smOlO
ICHlsH and Yosmhmsu
SE0
group was repeated with 1,1,2,2-tetrachloro-1,2-~fluoroe~ane as the standard. Perfluoro-compounds were chosen as the third group of compounds; a sample containing l/350 mole of perfluoromethylcyclohexane, l/l50 mole of perfluorobenzene and l/50 mole of m-difluorobenzene was prepared and diluted 1000-fold with n-hexane, and 0.8 pi was chromatographed, The perfluorobenzene was used as the standard. If the hydrogenolysis reaction, the absorption of hydrogen fluoride, and the response of the detector to fluoride ions are quantitative, the ratio of the peak areas of the three components for each group should be proportional to the number of fluorine atoms in each molecule, The average of the experimentally determined ratio of peak areas is close to the theore~cal value. It is thus shown that ftuorine bonded to an aromatic nucleus (which has been said to be particularly resistant to quantitative decomposition), pertluoro-compounds of the third group, and compounds containing other halogens in addition to fluorine can be deter~ned quantitatively. As mentioned previously, a propo~ionality to the fluoride ion concentration in the absorption solution should exist in the present apparatus in the range where the ionic strength of the absorption solution is constant. To confirm this, experiments were carried out with fluoro~nzene as a sample. The output e, of the antilog converter is confined to the range from 0 to -10 V. Therefore, supposing the concentration of the fluoride ion added to the absorption solution is IBM, the permissible range of the fluoride ion concentration in the absorption solution becomes 10-S-lOSM from equation (9). The ionic strength p of the absorption solution is O-820, and this changes merely to 0,821 when the fluoride ion concentration reaches lW_4& upon absorption of hydrogen fluoride. As a consequence, the changes in the activity coefficient of the fluoride ion may be ignored. Moreover, the changes in the pH are negligible. Therefore, if the amount of sample is adjusted so that the fluoride ion concentration falls within the range lo”-lO-sM, there should exist a linear relationship between the amount of sample injected and the peak area. Accordingly, solutions of fluorubenzene in n-hexane with a concentration of 04001-0~05M were prepared, 2 ~1 of each were injected into the column, and the relationship between the amount of sample injected and the peak area was examined. An absolute calibration curve was obtained and the existence of a linear relationship between the amount of sample injected (2 x 10-10-10-7 mole) and the peak area in the concentration range indicated above was found. CONCLUSION
A gas c~omatographic detector with using an ion-selective electrode has been developed and applied to the selective detection of fluorine compounds. It has the following advantages : the limit of detection is 5 x lO--lr mole (for fluorobenzene), the detector sensitivity is high, the degree of selectivity is 10,OOQor more with an extremely high sensitivity, the drift of the base line is O-02 mVj’hr with good stability. The response time of the detector is about 10 set when the concentration of Fis about lO-sJ4 and several seconds when it is more than IO-*M. Such slow responses distort the early peaks slightly but the effect is negligible for peaks having retention times greater than 5 min. A platinum tube was used as a hydrogenolysis tube in the present study, but could be replaced by a quartz tube.
Selective gas-chromatographic detection Zusammenfassnng-Die Proben werden an einer Gasehromatographiesriule mit Wasserstoff als Triigergas in ihre Komponenten getrennt. Die einzelnen von der S&de kommenden Komponenten werden durch ein auf 1000” geheiztes Platinrohr geleitet; dort werden sie hydrierend gespalt&. Fluorverbindung& werden in Fluorwasserstoff ~~~ef~rt. Der Fluo~~~~toff wird in einem langsamen Strom e&r Absorptionsliisung gelost und die Fluoridko~tmtion in der L&mg in einer Durc~~zelle mit einer Fluoridionen-Elektrode verfolgt. Das potentiometrische Verhalten der Zelle wird mit Hilfe eines antilo~~t~i~hen Umsetzers in ein Signal umgesetzt, das der Fluoridkonzentration proportional ist, und retistriert. Der Detektor sorach auf Fiuo~erb~d~~~ etwa 10 000&l starker an als auf eine gleiche Menge anderer or&nischer 5 s 1O-*1Mel Fluorbenzol komxten nachgewiesen Verbindungen; werden. R&tune--& &pare des composants dans des &hantillons sur une colonne de chromatographie en phase gaxense en u&ant l’hydrogi?ne comme gaz vecteur. On fait passer les composants individuels provenant de la colome a travers un tube de platine chat&5 a IOOO”, oli ils subissent ~hydro~nol~, et les composes Auor& sent convertis en acide fluorhydrique. L’acide fluorhydrique est dissous darts un lent courant dune solution &absorption, et la concentration en ion fluornre dans la solution r&&ante est contr6lBe dans une cellule a &oulement avec une electrode a ion fluorure. LQnission potentiom~~que de la cellule est convertie en un signal, qui eat proportionnel a la concentration en ion fluorure, par un conve~~ur ant~og~it~ique, et enregistre. La reponse du detecteur aux composes fluores est d’environ 10000 fois celle fonrnie par tine Cgale quantite d’autres composes organiques, et l’on a pu detecter 5 x lo-l1 mole de fluorobenzene. REFERENCES 1. W. Funasaka, T. Kojima and Y. Seo, Bunseki Kagaku, 1968,17,464. 2. T. Koiima. Y. Seo and J. Nishida. ibid.. 1968.17.1496. 3. T. Ko&& M, Ichise and Y. Seo,Sibid.,S1969,.18,~1460. 4. Idem, ibid., 1971,20,20. 5. R. S. Braman and A. Dynako, Anal. Chem., 1968,40,95.
6. E. Pungor, i&d., 1967,39 (13), 28A. 7. M. S. Frant and J. W. Ross, ibid., 1968,48,1169. 8. R. Bock and S. Strecker, 2. Anal. Chem., 1968,2X5,322.
547