Daniel M. Ottenstei n and Walter R. Supina and Patri ck W. Byrnes Supelco, I nc. Fi sher Sci enti f i c Co. Supelco Park 711 Forbes Avenue Pi ttsburgh, Pa. 15219 Bel l ef onte, Pennsylvania 16823 I NTRODUCTI ON Carbosieve B is a unique adsorbent f or gas sol i d chromatography which can be used f or the separati on of: (1) Permanent gases (2) Sul fur and ni trogen oxi des (3) Li ght hydrocarbons C to C (4) Certai n lower mol ecul ar wefght compounds such as formaldehyde, methanol and water (5) Water from most organi cs Because of i ts i nertness i t can be used f or the determi nati on i n the parts per mi l - lion and parts per bi l l i on concentrati on l evel . Due to its unique structure, i t w i l l el ute compounds in an order di f f erent from that obtai ned wi th other adsorbents or packings. Car bosi eve e i s manufactured by SUPELCO, I NC. Kaiser' descri bed Zl atki s et al . have al so descri bed i ts chromatographic properti es for i ts use f or the determi nati on of trace quanti ti es of water and as a general purpose GSC packing'. l i ght gas anal ysi s3. Carbosieve B is a hi ghl y pure carbon wi th a surface area of approximately 1000m2/g and pore radi us i n the range of 10-12 Angstroms, making i t essenti al l y a carbon molecular sieve. Its hi gh puri ty el i mi nates the adsorpti on problems that are nor- mally encountered wi th conventi onal carbon and other adsorbents; thi s is an impor- tant feature. i nstances, are consi derabl y di f f erent from that obtai ned wi th ot hd adsorbents. Its unique surface and structure cause separati ons which, i n many Thi s adsorbent i s hi ghl y nonpol ar; .Rohrschneider Constants show thi s very wel l and are as fol l ows: Benzene Ethanol MEK Nitromethane Pyri di ne (XI (Y 1 ( 2) (U) (SI -1.03 -0.84 -0.62 -1.93 -1.51 A l l val ues are negati ve i ndi cati ng that Carbosieve B is even more nonpol ar than squalane. compounds, even before methane. Because of its nonpol ar character, water is el uted before al l organi c APPLICATIONS PERMANENT GAS SEPARATIONS Permanent gases such as H2, 02, N2, CO, CH and C02 can be separated wi th 4 To do thi s i t is necessary to temperature program the column from are not easi l y separated so that a 9 f t. x Carbosieve B. room temperature to 175OC. 118" stai nl ess steel column is neede3, packed wi th 120/140 mesh Carbosieve B. column is hel d i sothermal l y f or four minutes, then programmed at the rate of 3O0C/min. O2 and N The 30 to 115C. Under these condi ti ons, oxygen and argon are not separated from each otter. 3 shows a chromatogram of CO, m4 and C02 i n ai r. SEPARATION OF HYDROCARBONS The components are el uted i n the order of H 2, 02, N2, CO, CH and C02. Figure Carbosieve B can be used to separate hydrocarbons i n the C - C range. TO do thi s, the column must be operated above ambient temperature. 'de C54and hi gher molecular weight hydrocarbons are not el uted from the columns. They appear to be permanently adsorbed by the Carbosieve B. Conceivably they could be el uted i f the column temperature were rai sed suf f i ci entl y, but thi s would resul t i n pyrol ysi s of the sample. range, the compounds are el uted by the degree of unsat- urati on, wi th saturated compounds el uted l ast. This i s i n contrast to conventional pol ar columns where the order would be (a) paraf f i ni c (b) ol ef i ni c (c) acetyl eni c. The pol ar columns would be usef ul f or traces of paraf f i ns i n ol ef i ns or ol ef i ns i n acetyl eni c compounds, but f or traces of acetyl ene i n ol ef i ns, or ol ef i ns i n paraf f i ns, Carbosieve B is unique. In the C1 - C 3 C Hydrocarbons 2 The C2' s are separated i n the order acetyl ene, ethyl ene and ethane. These com- ponents alolig wi t h methane are shown separated i n Figure 4 at 150'C wi th a 3 f t. x 2mm gl ass column. C3 Hydrocarbons order of methylacetylene, allene, propylene, and propane. Cyclopropane and al l ene are not separated from each other and are el uted together. C3's along wi th methane and the C2' s i s shown in Fi gure 5 wi th a 3 f t. column at ZOO' C. The C ' s are al so separated by the degree of saturati on. They are el uted i n the 3 The separati on of the C Hydrocarbons 4 I n prel i mi nary work wi th the C ' s, i t has been possi bl e to el ute is0 and n-butane, but not the unsaturated C ' s . 4 are not resol ved wi th a 3 f t. x 2mm column at 215'C. i n Figure 6 at 215C. The $so and n-butane peaks are rel ati vel y broad and Separati on of C1 - C4 is shown Acetylene i n Ethylene Zl atki s3 has shown that traces of acetyl ene can be determined i n ethylene. pri nci pal advantage of Carbosieve B f or thi s anal ysi s is that acetyl ene i s el uted before ethylene. This allows a l arge sample to be used wi thout swamping the acety- l ene peak, as is the case where ethyl ene is el uted f i rst. The The separati on shown i n Figure 2 was carri ed out wi t h a 3 f t . x 2mm column, Carbosieve B, 45/60 mesh at 14O"C, flow 4Omllmin. N and a flame i oni zati on detector was used here. Sample size 1.5m1, sensi ti vi ty 2 x 18-" A.F.S., lmv recorder. SEPARATION OF PERMANENT GASES AND METHANE, ETHYLENE AND ACETYLENE I f it i s not necessary to separate O2 and N but only ai r, CO, CH , C02, acety- methane, ethane, ethy- 2 lene, ethyl ene and ethane, the separati on can be carri ed out wi t h a 4 'ft. x 1/8" column wi th Carbosieve B, 60180 mesh programmed from 30 to l l OC as shown i n Figure 7. If O2 and N2 are al so to be separated along wi th C ' s el uted i n order of acetyl ene, ethyl ene and ethane. 2 lene and acetyl ene, the 9 f t. x 1/8" column used i n can be used with the 31 SEPARATION OF WATER, FORMALDEHYDE AND METHANOL Kaiser2 f i r st reported thi s di f f i cul t separati on; i t can be readi l y made wi th Carbosieve B as shown i n Figure 1. ti on contai ni ng 36-38% formaldehyde, 10% methanol and the remainder water. condi ti ons used are as follows: Column, 4 f t. x 3mmI.D. gl ass packed wi th 60/80 mesh Carbosieve B; Column Temp., 20OoC; Flow rate, 7fml/min. Hel i um: Sample si ze, 0.5 mi crol i ters. The separati on was made of a formaldehyde solu- The A thermal conducti vi ty detector i s used here because i t responds to al l three Note that the order of el uti on f or the three components components. A flame detector i s less sui tabl e since water i s not seen at al l and formaldehyde only poorly. di f f ers from that obtained wi th porous polymers. TRACE ANALYSIS Trace Gases Kaiser2 has shown that trace anal ysi s i s possi bl e wi th Carbosieve B, and that i t is possi bl e to detect parts per bi l l i on of hydrocarbons using reversi on gas chro- matography. This technique consi sts of pul l i ng a l arge, known volume of the sample through the column, al l owi ng the i mpuri ti es to col l ect. Then carrier gas i s turned on and the column i s heated, by a moving oven, which causes the sample to move through the column. This technique allows f or the sample t o be hi ghl y concentrated i n the column. Since Carbosieve B does not bl eed, it can be used wi th ul tra hi gh sensi ti vi ty detectors such as helium i oni zati on. TRACE WATER ANALYSIS Kaiser' has shown that trace quanti ti es of water can be determined using a Carbosieve B column. Because of the l ow af f i ni ty f or water of i ts carbon surface, water i s quickly el uted from a Carbosieve B column wi thout tai l i ng even before methane. from bei ng adsorbed on the wal l of the column tube. is less adsorpti ve than gl ass or stai nl ess steel tubi ng To carry out trace water anal ysi s, care must be taken to prevent water Kaiser has shown that quartz SULFUR GASES Sul f ur di oxi de is readi l y el uted f r m Carbosieve B columns wi thout i ndi cati on 9 of tai l i ng as shown in Fi gure 8. to 30 ppm. Bel ow that l evel the SO2 i s l ost on the column.2 Consequently, Carbosieve B can be recommended onl y f or hi gher concentrati ons of SO2. badl y at rel ati vel y hi gh concentrati on s o Carbosieve B cannot be recommended f or H S anal ysi s at this ti me. Kaiser has found that SO can be determined dawn Hydrogen sul f i de tails 2 OXIDES OF NITROGEN Both ni trous oxide (N 0) and ni tri c oxide (NO) can be separated from each other as wel l as other gases wi t 2 Carbosieve B. No i nformati on of ni trogen di oxi de (NO2) i s avai l abl e as yet. Ni trous oxide is readi l y separated from other compounds and is el uted af ter CO and bef ore acetyl ene. Thfs separati on was carri ed out at 100C wi$h a 3 f t. x 118'' stai nl ess steel column wi th 60180 mesh Carbosieve B. Figure 9 shows N 0 separated from ai r, methane and C02. 32 Ni tri c oxide is el uted from Carbosieve more rapi dl y than N 0. A t room tempera- 2 ture i t i s qui ckl y el uted af ter N a 6 ft. x 1/8" stai nl ess steel co?hmn wi th 60180 mesh Carbosieve B. and before CO. This i s shown i n Figure 10 wi th Speci al precauti ons must be observed when NO i s to be separated. The column must be conditioned wi th hydrogen to prevent the NO peak from tai l i ng. This can be done wi th ei ther pure hydrogen or a mixed gas such as ni trogen or helium and hydro- gen. f or condi ti oni ng. essenti al l y the same resul ts. wi th trace quanti ti es. We have tri ed a 94% N2-6% H2 mixture and found thi s comparable to pure hydrogen It should be noted that none of the work with NO deal t Conditioning f or one hour at 200, 300 and 400C seemed to gi ve I f the column i s to be operated at room temperature f or NO and then temperature programmed to hi gher temperatures, the carri er gas should contai n hydrogen. column i s heated wi thout hydrogen i n the carrier gas, then the NO peak wi l l tai l . This was found wi th hel i um as the carrier gas. ri orated f or NO when the column was repeatedl y temperature programmed to 200C to el ute other components. expl oratory, and at thi s poi nt we are not certai n that i t can be reproduced. I f the We found the column gradually dete- Our work wi th NO and the hydrogen deacti vati on has been Both NO and N20 can be separated from 02, N2, CO, CO and methane with a 6 f t. 2 x 118" column. The column i s operated at room temperature ti l l CO i s el uted, then the column is programmed qui ckl y to 150C to el ute methane, C02 and N20. When the column i s programmed, the carri er gas must be dry. MISCELLANEOUS RETENTI ON DATA Data prepared by Kai ser i n Table 1 compares Carbosieve B, Porapak Q and Molec- ul ar Sieve SA. Note that the val ues are retenti on volumes, not retenti on ti mes. Water is shown to el ute bef ore methane and CO on the Carbosieve B but it i s evi dent that i t takes qui te some t i me f or butane to be el uted, even at 250C. 2 The data i n Table 2 were obtai ned i n our l aboratory usi ng a 3 f t. x 118" O.D. stai nl ess steel column packed wi th 60180 Carbosieve B and operated at 40mlfmin. and are absol ute retenti on times. HANDLING OF CARBOSIEVE B Carbosieve B is suppl i ed i n f i f teen cc evacuated, seal ed vi al s. when the vi al i s opened, unused material should be transf erred to a gl ass bottl e and ti ghtl y closed. To avoid contamination, Carbosieve B should not be exposed to the atmosphere f or pro- longed peri ods of ti me. Because of i ts hi gh surface area, it tends to readi l y pick up contaminants from the air; these can be removed but i t requi res extensi ve condi- ti oni ng of the column. To avoi d oxi dati on and damage to i ts surface, do not heat Carbosieve to over 2OO0C i n air; do not heat columns above thi s temperature i f the carrier gas i s not f ree of oxygen. Columns that are to be temperature programmed shoul d be conditioned f or a number of hours at or sl i ghtl y above the hi ghest tem- perature anti ci pated. f Carbosieve B shoul d be used wi th a carrier gas that is f ree of oxygen to avoid oxi dati on of i ts surface. Oxygen can be removed catal yti cal l y by means of speci al l y designed traps which are i nstal l ed i n the carrier gas l i ne. I f oxi dati on shoul d occur, the peak w i l l show tai l i ng. I f temperature programming is used, i t i s imperative that the carri er gas al so be dry. Use of a devi ce such as a molecular si eve trap to remove water is recom- mended. I f the carrier gas i s not dry, an i rregul ar base l i ne wi l l be obtained when the column i s programmed wi th a broad peak f or water seen i n the vi ci ni ty of the co2 33 peak. This i s a problem wi th a thermal conducti vi ty detector, but not a flame which does not respond to water. A dryi ng' cartri dge should be used t o remove water from the carrier gas. One must assume that al l carrier gases have somemoi sture i n them. to i nstal l a s mal l dryi ng trap af ter the flow control l er i n the i nstrument. seen i n several i nstances where it is the flow control l er bl eedi ng moi sture that causes the base l i ne problem even though the carrier gas was being dri ed. One can also heat the carrier gas tubi ng from the cyl i nder to the chromatograph to remove moi sture from the wal l s of the tubing, but thi s i s unnecessary i f the trap i s pl aced af ter the con- trol l er. Once thi s is done, heat the column to the maximum temperature which you expect t o reach and hold the column at that temperature f or several hours t o condi ti on it. We al so f i nd i t necessary We have A short secti on of tubi ng f i l l ed wi th Molecular Si eve 5A is adequate f or thi s. The problem of moisture i n the carri er gas is i l l ustrated i n Fi gure 11 which The si ze and posi ti on of the water peak w i l l shows the peak f or water af ter C02. depend on the amount of water present. REFERENCES (1) Kaiser, R., Chromatographia 2, 453 (1969) (2) Kaiser, R., Chromatographia 3, 38 (1970) 3) Zl atki s, A., Kaufman, H. R. and Durbin, D. E., J . Chromatog. Sci . 8, 416 (1970) 4) Kaiser, R. , Private Communication \ 34 TABLE 1 Molecular Si eve Carbosieve B Porapak Q 5A O2 N2 c02 CH4 c2H2 c0 ZH4 ZH6 C3H8 qH1O gH12 CHOH CH30H C2H50H H2 150C 1.22 1.22 1.75 9.09 4.71 21.6 40.5 64.1 - - - 27.1 65.7 - 4.59 250C - - - 1.80 1.03 4.08 6.59 9.42 54.9 126 .O - 3.96 8.35 38.3 0.51 l5ooc - 0.28 0.28 0.28 0.78 0.40 1.11 1.04 1.23 3.09 7.83 16.72 3.78 5.59 12.35 2.61 Speci f i c Retention Volume Vg at 150C Methane 3 . 0 ml f p Ethane 47.4ml/gm Propane 506. l ml f gm 15OoC 0.32 0.56 1.34 1.01 43.0 - 63.9 7.36 - - - - - - - I n the tabl e above (prepared by Kai ser) retenti on volumes f or a number of compounds are compared .for several types of columns. I n al l cases column l ength was one meter. 35 TABLE 2 RETENTION TIMES (ABSOLUTE) HZ O2 N2 m4 c02 N2 H2S s02 Argon NO CO Acetylene E thy l ene Ethane Methyl Acetylene Cyclopropane Allene Propylene Propane Water Formaldehyde Methanol Column Temperature Ti me i n Minutes l0O0C 15OoC 2000c 0.23 0.23 0.23 0.23 0.23 0.23 0.5 0.92 1.45 2.4 4.7 5.4 9.8 - - - - - - - - - 0. 2 0. 2 0. 2 0. 2 0. 2 0. 2 0.3 0.4 0.6 0.85 0.450 1.6 2.6 5.0 - 12.8 15.7 - - 0.15 0.35 1.0 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.27 0.3 0.4 0.3 0.7 1.0 1.6 3.7 4.2 4.5 6 .O 8.6 0.15 0.2 0.45 Column: 3 f t. x 118'' S.S., 60180 mesh Carbosieve B Flow: 40mlfmin. 36 _ . YlNUlES FIGURE 7 - Water-FormaldehydeMethenol Separation 0 2 4 6 8 10 MINUTES FIGURE 3 - Separation of O1 N2 CO. CH4 and C02 FIGURE 2 - Trace Acetylene in Ethylene METHANE ;I I I ETHYLENE ACETYLENE I ' ! .5 1.0 1.5 ' 2.0 2.5 0 r MINUTES FIGURE 4 - Separation of Methane and Cz's I i / I 0 1 2 3 4 5 6 MINUTES FIGURE 5 - Separation of C,'s, C2f and C3's FIGURE 6 - Separation of n-Paraffins - C, - C4 c I ' 0 I 2 MINUTES FIGURE 7 - Separation of Air, CO, CH4 COz C p z Cp, and 'p6 FIGURE 8 - Separation of Ai r and soz 38 1 0 I 2 MINUTES FIGURE 9 - Separation of Ai r. CH4 C02 and N20. 2 i 0 1 2 3 MINUTES FIGURE 10- Separation of Nr NO and CO . .. I , 1 1 I 2 3 4 5 6 7 YINUTES FIGURE 11 - Separation of Or Nz CO, CH, COz and H20.
Physical Organic Chemistry — 3: Plenary Lectures Presented at the Third IUPAC Conference on Physical Organic Chemistry, Montpellier, France, 6 - 10 September, 1976