Basic gel permeation chromatography studies

JOURNAL OF CHROMATOGRAPWY

BASIC II.

GEL

PERMEATION

BENZENE

J. G.

RESULTS

543

CHROMATOGRAPHY WITH

SMALL

STUDIES

MOLECULES

I-IENDRLCKSON

The Dow

Chemical

(Received

August

Com+any 25tl1,

I;ree$orl,

Texas

(U.S.A.)

1947)

1NTROI)UCTION Gel permeation chromatography (GPC) studies from this laboratory have included papers on the elution data from simple solutes in tetrahydrofuranl and chloroform2 using 40 A gel in the GPC columns. The studies based on tetrahydrofuran led to rules to predict elution, and basically the molecular sizes as seen by GPC were proportional to the sum of the contributing chain lengths. In this early work, hydrogen bonding to the solvent was measured, and almost all non-hindered OH compounds studied were strongly bonded to the solvent. About 130 compounds were studied in tetrahydrofuran, and the correlations used produced a good agreement between a measure of the molecular sizes observed in GPC and the theoretical values for a range of molecular types. The data with chloroform as the solvent showed that it acts as a hydrogen bond donor, and almost no revisions of the concepts used in the tetrahydrofuran work had to be made to explain the chloroform results. Results in benzene are of interest, since this solvent is classified as neither a hydrogen bond donor or acceptor; yet it is a good GPC solvent and gel swelling agent. An evaluation was also desired as to whether benzene was a rather fine, good or poor choice for studying small molecules, since earlier work could be interpreted either way. CORTIS-JONES~ studied several solutes and solvents with a range of GPC-type, low permeability gels. Prom his results with ten oxygenated compounds in benzene, he concluded that the hydroxyl group caused relative exclusion of the compounds that contained this grouping. EXPERIMENTAL Prior techniques were used in this studylp”. Basically, 3/S in. O.D. steel tubing 4 ft. long was packed with 40 A styrenedivinylbenzene GPC gel. The gel was permeable to alkanes below a molecular weight of 450, The GPC runs were made using a commercial machine, made by Waters Associates Incorporated*. Benzene at 1.0 ml/min was the elution solvent, and typically 0.1 ml of 4 o/osolute in benzene was injected using an automatic injection system. Molecular size calibration was basecl on a mathematical comparison of the elution volume of new solutes to that of the normal alkanes and the ethers studied. The calibration data were fitted to an equation of the form: E =

Vz +

S x

(log

(No.

of C atoms

f

A!)) J, Ckromalog,,

32 (IgGS)

543-558

J. G. WENDRLCXSON

544

For atoms other than carbon, their effective chain lengths are given in Table I, using a theoretical model based on additive structural elements. The calibration TABLE

I

A COMPARISON OF EPFECTIVE CHAIN LENGTHS Based on the clution volumes of walkanos.

IN

GPC

Ge?aeYal di fersnces

A toms

(No. -CN,

1.0 -

c--o-c

1.0

I.0

1.0

0.67

-

0.67

c=o

1.0

1.0

0.65

0.65

0.61

0.61

0

0

-Cl

1.08

I.37 I*54 0.67

0.54

0.22 -_

-1 -01-I

o-43 0.55 0.67

c==c

-Br

--COOH Ar OH

-/-2J3_5n

4-2.85n

2.33 +2,85a

-_

0.77

-

0

0

0,61 -

-1.5

--

0.91

None None 0.35

3*54h 5.7”

j-J.0

0.67 -I-4.oh

-NH,

0.68

of c units)

I.5

G.2”

0

3.76”

-l-2.85a m Hyclrogcn bonding observed in TI-IF and not in benzene increases molecular size. I) Plus a correction of 0.15 times sorbing groups present times length of longest chain.

equations were based on the constants, given in Table II. In reporting peak widths, tangents were drawn to each side of the GPC curve and the number of milliliters of elution at the base of the triangle is reported in the tables under the heading Elution (ml), at’ base. DISCUSSION

Benzene was studied as a GPC solvent to see if rules could be derived to predict solute elution volumes. As in prior studies, molecular size was determined by comTABLE

II

CALIBRATION

CONSTANTS

Column

Vi?2

A I3 C D

98.68

165.89 137.26

n Theoretical J.

Clrromatog.,

S 37. I4 70*88 56,oG (see text)

Ii!

TPP

0.0 2.03 x.16

5oo 990 990 990

plates per ft. using ut-pentane as solute.

32 (1968)

543-558

Lengtlt 8 12 12 22

(ft.)

I

16

‘4 15

13

11 12

IO

S 9

3

2

76.0

Di-octyl ether tr-Do&cane jz-Decane tr-Heptane

. ::;;

4.67

El

9.13

9.0

4.73

5.10

6.99

12.22

12.0

0.06

0.10

-0.01

0.22 0.13

-0.01

16.66

16.67

0.1

0.06

-0.3r

-0.04

5.10

6.06

12.17 10.16 6.69

0.16 0.17 0.16

16.50

4.63

6.0 . 5-o 4-67

7.0

10.0

12.0

16.67

-

-

-

d

Error (06 -

-

Obs.

Tkeory

Compared ckaitz ten@ (No. ojc urrits)

a (Molecularsize in tetrahydrofuran)- (the size in benzene in No. of C units).

( 7;:6

4-41)

n-Heptane wPentaue Di-ethyl ether

F:

3.97

4.16 4-24) 4-25) q-20)

105.4 ro7-5

+Pentane Dðyl ether

3.89

3.Sr 3.96 3.sz 4.57

3-70

3-34 3.60

At base

Di-octyl ether wDodecane n-Nonane

101.5

n-Hexane

99.2

85.9

84.2

55.3 62.5 67-4

P-4.000 (~%W20 WC,&JzO

To peak

EMion (ml)

A

-

-

-

--

-

-

-

-

-

-

-

CT - C$

-

Ad

T)

A

A A

A A A

B B B

B B B B

B B B

COh~W

GPC GEL Run conditions: 3/S in. 0-D. columns; benzene solvent at I ml/min at 24’, using a 12 ft. column, except elution data in parentheses are from an S ft. column.

T~ELUTIONOFNOW.~HYDROCI\RBONSANDETHERSINBENZENEUSING~O

-..

546

--

J.

C..

Gi 2333NJJRICICSON

paring elution volumes of the solutes with a series of normal paraffins and ethers. The results were expressed as an effective chain length. of so many carbon atoms long, and the effective length of heteroatoms were figured from their bond lengths, bond angles and covalent radii. Calibration standards are reported in Table III. Two GPC columns having 8 ft. total length were studied extensively with tetrahydrofuran elution solvent, and then they were studied with benzene solvent. The allcanes and ethers eluted slightly earlier in the new solvent, indicating that benzene was the better gel swelling agent of the two solvents. The ethers and allcanes eluted on the same calibration line, which indicates the absence of partition, adsorption and bonding to the solvent with these ethers. Also, the light gases studied eluted only very slightly after propane, in agreement with the results observed with tetrahydrofuran as the elution solvent. Hence, those compounds eluting after or smaller than a three carbon atom size in the current worlc were assumed to be sorbed-either adsorbed, absorbed, partitioned, permeated, chemisorbed or otherwise delayed due to contact with the gel surface. Summarizing, sorption played an important role in the elution volumes observed in benzene. The literature on sorption chromatography is developing” to show that one can at times predict elutions that are delayed by sorption, and one technique is to predict elution by using factors in a mathematical relationship, which uses one factor for each functional group present. This approach was taken in the current work. However, certain precautions were taken-for example, the degree of steric hindrance of a phenol was a factor in the sorption observed with it, as one might anticipate. Also,‘sorption was too complex to unravel into its parts at this time. Benzene derivatives and other aromatic compounds were studied, and their elution volumes were compared to the calibration standards used. A size of 2.2 carbon atoms long was used for the effective size of the phenyl group, and the correlations thus obtained were in good agreement with the observed values, as shown by the data in Table IV. A group of carbonyl compounds was studied in benzene, and the observed elution volumes deviated from the tetrahydrofuran resultsin benzene, these cornpounds eluted later than the qz-alkane standards having the same apparent molecular size. However, useful correlations were obtained by not figuring in the oxygen atoms into the chain length. This technique gave the results in Table V, The reasons for this phenomenon are not understood, but sorption appears to play a major role. Nevertheless, these corrections corresponded to only a IO o/oto 15 o/oerror in molecular size, which is less than a 5 o/ochange in predicted elution volumes. A group of halogenated compounds also eluted later in benzene than in tetrahydrofuran, but not very much later. As expected, chloroform appeared later and hence smaller than did carbon tetrachloride in benzene. Since no molecule looked bigger than expected, one may conclude that there was no evidence for hydrogen bonding of these compounds or for self-association in benzene. Correlations were tried on the data, and the one that appeared to work best was to place a value on each halide atom at half of the size found in the tetrahydrofuran work. This data is compared in Table I, and the observed results using these chain lengths are given in Table VI. The results agree fairly well with the theory. A group of alcohols and amines were studied in benzene to detect their elution pattern. An initial tabulation of the data for the alcohols indicated also that the oxygen J. Chromalo~., 32 (1968) 543-558

9

peak

81.0

Anisole Phenetole 107.6

I 14.6

84.8

Dicumyl peroxide

102.7

I 7 ( ,,5

(

To

Ehrlim (ml)

96.0

III.

Naphthalene Biphenyl Diphenylmethane Diphenoxyethane

Toluene Ethylbenzene p-Xylene Cumene

Run conditions: see Table

GEL

4-56 4.43

6.02

4-94

3.57 4.87

XI.73

5.40 7-73

4-40

3.26 4.61

II.90

5.40 7.65

2.90 3.95

3.18

6.20) 5-92) 5.08

5.56

2.99 4.28 4.02

06s.

5.20

3.20 4.20 4.20

ThOry

--I

1;

---I

At base

&GPC

Cotnpayedchaittlet@ (No. ojc units)

THE ELUTION OF AROMATIC HYDROCARBONS AND ETHERS IN BENZENE USING 40

TABLE IV

B

1.76

0.76

0.0

-0.61 -0.26

0.r7

-0.08

0.94

B

1.29 I.49 -

-0.28

-0.45

A” It B B B

A

Log

A

coltlm~

0.56 0.58 0.33

CT--Cl3

0.08 -0.18 0.36

AA

2=)

-0.20

A

Eyyo~(06 -

gS.r 94.2 73-S

Dimethyl adipate Dimethyl sebac-ate

9

11

j.ss 6.30 7.01

6.0 6.2

102.2 100.6

lsobutyl methyl ketone Dimethyl terephthalate qz-Heptaldehyde

4-03 4% 4.39

4-o 4.0 4-o

4.17 4-62

4.40 5-30 6.40

4.34

L&IO

4-46

110.4 110.3 1oS.6

3-o 3-o 3.0

4.66) 4.34 4.52

12.0

s.0

7-o

S.?3 12.39

3-47 3.19 3.rS

Obs.

( 75.6 II5.3 101.5

Tlzeory

At base

Compared chaitt leta$lb (No ofc tfltits!

GPC GEL

To peak

;

IO

USING 40 if

Eltttiolt (ml)

a-butyraldehyde a-Butanone Ethyl acetate

Acetone Acetone Acetone

Contfiotrttd

IN BENZENE

4 z

3

2

I

Run

Run conditions : see Table III.

THE ELUTION OF CARBONYL CONPOUNDS

TABLE V

0.23 o-39

0.01

-O-I2 0.10

0.03 0.05 o-39

o-47 0.19 0.1s

A

Error (Ob dd

T)

0.71

0.66 0.94 0.95

CT-CB

B C

B

B

B B B

A B c

Colt1tttis

N

b.l

iz m

I

!I

o\ m

Q

_ n

OFORGANIC

15 26

j

4

3

I 2

4.55 4-93

Dibromobenzene-(P) Dichlorobenzene-(P) 102.0

j.SO)

103.9

5.45)

Chlorobenzene Bromobenzene Iodobenzene

5.28)

-) -1

r,4-Dichlorobutane Benzene hexachIoride ( 74-o ( 70-7

5-44) -) -) -)

Methylchloroform Tetrachloroethane Pentachloroethane Perchloroethane

2.51

ChIoroform Carbon tetrachloride Ethylene dichloride

3.57 3.29

2-74 2.88 2.97

(5.29

5.09

4.0-t

3.20 3.51 3.61

2.74 3.09

I.77 2.65

Theory

3.22

2.sg

3.06 2.86 2.86

4.61 5.66

3.64 3.60 3-53 3.69

2.76 3.1s 3.02

2.Sj

1.90

Oh.

(NO. Oft 26&S)

Compared ckain le@k

Methyl iodide Ethyl bromide

To peak

AGPCGEL

At base

USING40

Elation (ml)

HALIDES INBENZENE

Run conditions: see Table III.

THEELUTION

TABLE VI

-0.68 -0.07

-0.11

0.32 -0.02

0.07 0.37)

-0.35

0.22

0.09

0.44

o-44 -0.06

0.25

0.13 0.17

_!I

Error (Ub AA

T)

1.35 1.42 r.4r

1.52

5.16 3.26 1.63

1.29

2.rj

O-55 O-54

CT--Cl3

C

C

A A A

A A

A A A A

A A A

A A

.

Cozutt6n

;op

ul

.

cn m

x

2

;::

4.88 4-70

97.3 98.7

wDecano1 a-Decanol

I3 ‘4

I2

7-o 7-o 7-o

4.65) 4.77) 4-40

1 68.7 ( 69-S 101.0

3-Heptanol n-Heptanol a-Heptanol Diethylene glycol monomethyl ether 2,2,4-Trimethylpentanol IO.0

10.0

4.70

4-Y

93.5 82.2

4-o 3.2 4.0

8 9 IO II

4.60 5.22) 4.52

(110.7 82-7 112.7

5.43

8.47

6-33 6.83

6.41 5.99 6.20

3.97 2.69 **97

I.51 I-45 2.76 2.93

1.0 2.0 3.0 3-o

Wt.-Butyl Benzyl alcohol alcohol IZ-Butanol

5-47) 4-54 4-52 5.47)

Obs.

Theory

2 7

( 92-o 127-3 117.6 ( 81.3

At base

Methanol Ethanol Isopropyl alcohol d?ropanol

To peak

(No. of c tmi2s)

Coilrpared chair8length

2 3 4

I

Rrrtr Compouwd

Elufiorr (ml)

THEELUTfONOF ALCOHOLS IN BENZENE USING40 A GPC GEL

lo -%

Run conditions: see Tabl
TABLE VII

*

w

I.53

- I.57

-

-0.67 -1.17

-0.59 - 1.01 -0.80

-0.51 -0.03 - 1.03

O.jI -o-55 -0.24 -0.07

d

Error (Ob -

-0.07

-0.03

0.3s -0.42

0.16 + o-04 0.25

-0.03 0.27 -0.43

-0.2 j + 0.06 0.38

dd

T)

5.11

3.75 4-30 4.09

j.OS 4.33 4-13 3-64

CT--CR

C

B

B B

A A B

A B B

A B B A

Colnar~

Q

4.

Aniline

Pyridine

Benzylamine Dimethylaniline

2

3

4

9

3

I 2

128.8

To peak

83.5 110.0

64.9

89.9

95.5

110.0

t 76.2 117.7

101.9

129.3

125.0

To peak

Etlrtiort (tttt)

2.56 4.11

2.39

2.LI

I.39

Obs.

6.23

IS.37

-1.63

-1.60

6.So

8.40 4.52

20.4

-3-37 -4-65 -3.25 -0.39 5.37 6.97 -

d

- 1.48 - I.95 -2.49 4.02 2-75 4.11 7.81

3-70

1.25

0.72

Obs.

T)

-O.Ij

AA

0.22

-0.4s

AA

T)

o-47=

0.3%

--0.99 O.Oj -o.3ga

0.29

-0slg

Q-35 -0.30

Error (Ob -

-0.09

-0.63

c?.rg

-o.ag

-0.81

A

Error (Ob -

7-4 7.4 7.40 8.20

2.2 -2 i.2

Theory

(No. of c mik)

Compamd chailr leught

3.20 4.20

2.2

2.2

2.2

Theory

[No. of C units)

Connfiaredchaitz letzgth

12.95)

6.25 6.13 6.16

At base

GPC GEL

5.53) 5.40

7-24 6.57) 6.76)

,4t base

GPC GEL

USING 40 A

(

86.6 ( 54.6

(

d

Elutiorr (ml)

USING40

a One OH group is sterically hindered.

Bisphenol “A” (p,p’) Bisphenol “A” (p,p’) o,p’-Bisphenol “A” z,6-Diisopropylphenol 6,4,6’,4’-Tetramethyl-2,2’dihydrosybiphenyl 4,6,4’,6’-Tetra-tert.-bntyl-2,2’dihydroxybiphenyl

PhenoI p-Cresol p-tert.-Butylphenol

Run conditions: see Table III.

THEELUTLONOFPHENOLSINBENZENE

TABLE IX

j

Aniline

I

THEELUTIONOFABUNESINBENZENE Run conditions: see Table ID.

TABLE VIII

6.81

CT-CS

5.27 4-57 0.90

CT-G

D

D

A B B B

D B D

Cohlr~m

A

B A A

COliWZ

552

J. G. HENDRICKSON

atom of the OH groups could be ignored in measuring the effective chain length, as was done with the ketones and esters studied. However, a slightly larger correction was needed for the alcohols than was made for the carbonyl compounds. Plotting the data l.ed to the conclusion that a correction of 0.15 carbon atoms chain length times the length of the longest chain length attached to the OH group gave an improved correlation. Further, it was observed that this correction peaked out at about IO carbon atoms and declined thereafter. These results are tabulated in Table VII, and the refined results are compared to the observed values in the column of Table VII headed by the symbols dd. Amines also conformed to this general picture, with some reservations. The nitrogen of the amine could be ignored from measurements of the effective chain length in benzene, provided that it was not buried, such as in nitrobenzene or diethylaniline. The 0.15 carbon atoms correction for the longest chain also appeared to be applicable to the amines, although this is based on a relatively small number of tests. These results with amines are given in Table VIII. With phenols in benzene, elution occurred much later than in tetrahydrofuran, and the difference was typically about seven carbon atoms in size. This would consist of four carbon atoms length due to the absence of hydrogen bonding, 0.65 C units due to the smaller effect of the aromatic ring in determining molecular size, 0.65 C units due to not measuring the oxygen atom, and about 1.5 carbon atoms length beyond these factors. With the substituted phenols, elution was still later than was anticipated with the above four corrections aplied to phenols, and the delay paralleled that observed with alcohols. EIence an 0.~5 C units times the chain length correction has also been figured into the revised chain lengths for the phenol data, and the results in Table IX agree closely with the observed values, Organic acids in benzene were also studied, and the results indicated a high degree of sorption. Typically, the portion of the GPC curve that eluted first was very close to normal in shape and width, for organic acids in benzene; but these acids eluted over a range of 15 to 25 ml, which was several fold more than the other solutes of the current study. Also, the elution position of the acids changed considerably from column to column, and the position of elution was changed by other polar solutes present. For example, acetic acid eluted 5 ml earlier in the presence of phenol, compared to when it was injected alone, Still, rules were sought to predict the elution of these acids studied. The elution correlations finally chosen were in line with the prior observations. The oxygen atoms were not figured into the effective chain lengths, and a correction of 1.5 carbon atoms lengths was applied for the acidic hydrogen present. Further, a correction of 0.3 carbon atoms per carbon atom in the longest chain attached to the COOH group appeared appropriate. (One may reason that the COON group is a bidentate type of absorbing group). Using these guidelines, the results in Table X were obtained, and the agreement of theoretical and observed values was fairly good. Epoxidcs were also studied, and in this case the best correlations were obtained by combining two prior principles; these were not to figure in the oxygen in a group that could be sorbed and to substract 0.67 C units for the small ring structure. These concepts led to the results in Table XI. A group of other cyclic compounds were studied, and they gave results in line with the molecular sizes observed in tetrahydrofuran. These results are also given in Table XII. In tetrahydrofuran, unsaturated compounds showed almost no deviation from J. Clzromato~., 32 (1968) 543-558

I

Propylene oxide ’ Phenyl f$ycidyl ether Butyl glycidyl ether p,p’-Bisphenol diglycidyl ether

Run conditions: s&tTable III.

THEELUTIONOFEPOSIDESINBENZEME

2 3 4 -.

so.3

112.2

131.7 115.6 20.5a 23.4a 1S.6a 15.1’

i\GPCGEL

2.0

6.0 7-o 16.0

-2.90

-3-31 - I.90

3.69 14.og

-1.72

3.10

0.27 0.08

0.8

-0.01

0.10

‘19.7 106.4 97.6 82.3

To peak

Ellttio~l (id)

GEL

44I 5.07 4-56 5-69

At base

USING 40 k\Gpc

.:

3-o 5-57 7.67 ‘4-4’

Theory

2.46 4.57 7.16 13-40

Obs.

Compared chain Ieqtk (No. qf C ,wits)

-o-54 - 1.00 -0.51 - 1.01

A

Error (Ob -

0.13 -0.33 o. 16 0.32

AA

T)

5-95

0.57

CT--Cn

a Considerable tailing was observed, and the width of the leading edge was similar to that of other compounds studied.

Acetic acid Caproic acid Heptanoic acid Palmitic acid

TABLE XI

2 3 4

I

Run conditions: see Table III.

THEELUTIONOFORG:\NICACIDSINBENZEWEUSING_~O

TABLE X

B B B B

co!lltrls

B B B B

UNSATURATED,

113.4

( 67.3 82.7

( 65.6

Heptene-r Decene-r

Isooctane

13 14

15

12

4.28)

4.41) 4.56

4.82 4-Q 4.21

114.2 g1.S go.6

Styrene 2-Methylpentene-I 4-Methylpentene-I

II

IO

9

5-67)

4.90

5-30

5.11

4.90

4-75

112.9

4-75 4.59 544)

107.1 110.0 103-S

( 76.2

I 10.7

At base

I-93

2.55

8.0

10.0

7.0

4.10 6.0 6.0

3.43 4.10

3.40

7.77

6.53 9-54

3.33 5-3r 5.63

2.06 3.34

2.92

2.29 I.90

2.85 2.85

4.02

3-47 3-97

06s.

3.40 3.40

2.85

Theory

(No. ojc units)

USING

-0.23

-0.46 -0.45

-o-77 -0.69 -0.37

-1.37 -0.76

-0.92 -0.54 -0.65 - 0.48

0.62 o-57 0.62

A

Error (06 -

40 A GPC GEL

Cotnpaved chain Zen$h

COMPOUNDS IN BENZENE

108.8 ( 79.2

Furan Furan Ally1alcohol a-Butene-I-01

Tetrahydrofuran Cyclohexane Cyclohexane

To peak

EZztiion(wl)

AND BRANCHED

Propargyl chlotide Phenylacetylene

S

3

2

I

Run conditions: see Table III.

THE ELUTION OF CYCLIC,

TABLE XII

-0.06 -0.05

-0.37 -0.29 0.03

i-Y7 0.0

0.26 -0.25 -0.08

-0.12

AA

T)

0.05 0.10

CT-&

B B B

A

C

: C

D

B B A

CoZmn

9

r2

I1

ro

Y m cu

p”

.ti?

S Er,

A

(w N

a$ :

I7

16

Ij

a 13 g I4

s

0

yi

. 102.8

92.6 134-7

5-o

4-72

4-23

1.0

5.73

5.96 0.72

6.65 9.84 rr.06 7-62

6.67 IO-33 10.0 s-54 6.0

5.73

0.40

2.40 3.12 6.36

0.73

-- 0.04 -0.2s

-0.49 1.06 -0.92

-0.02

0.36 -0.27

0.12

0.19 -0.08

0.32

-0.r

0.07

-0.2

_!I

0.06

0.39 0.24

Ad

EYYOY(Ob. 3 T)

3.29 3.02

3 .22

3-9

7.67

I1.0

Ohs.

6.0

63:,”

2.0

2-9 3-r 3-r

4-o

7.6

II.2

Theory

Compared thaiIt leizglk (No. of C units)

_i GPC GEL

4-96

5-36) 4.77! 7.63

N atom. * See text. b Buried OH group. c Buried __

Diethyl carbonate

n-Propyl sulfide Carbon disulfide

( 68.1 ( 61.8

Dipropylene glycol Tripropylene glycol Tetraethylene glycol Hydrogenated bisphenol “A” 86.7 96.1

roo.4 r02.S

Hexanonitrile Adiponitrile 7-16)

5.04) 4-75 4.20 4.23

( 54.5 ’

s7 Propionitrile “j-4

4.75

rr6.0

Acetonitrile

2

79-3 79-5

Nitroethane Nitrobenzene Nitrobenzene

( (

5-47)

--I

5.79)

( 60.4 l 65.8

At base

To peak

EMiotr (rd)

INBENZENEUSINGa#O

Propylene carbonate

Triphenylcarbinol

2

3

IOUdl

I

Run conditions: see Table III.

THEELUTIONOFMISCELLANEOUSCOBIPOUNDS

TABLE XIII

4.98 4.55

0.5

2.17 2-34

0.63 4.60

CT-CB

Bd

Dd Bd

Ad Ad Bd Bd

Bd

Ad w Bd

BC

AC

AC

A

Ab A

COh6?2

Cominents aid

E u1

.=:

.

5 g

2

z

g $

$

ul

o.

2 w CL

126.2

Quinoline

Methyl methacrylate

2

3 gs.2

139.8

p-Methoxyphenol

To peak

Elutim (ml)

I

Run conditions: see Table III.

4-95

6.45

5.60

4-Q

3.1s

I.79

Theory

3.81

I-59

0.30

Obs.

Compared chaifr lerrgth (No. of C wits)

GE’C GEL

At base

THE ELUTION OF OTHER COMPOUNDS IN BENZENE USING 40 A

TABLE XIV

-0.84

-r-59

-1.49

A

Ewor (Ob. -

d Ll

T)

t

CT --cB

C

B

B

Cokaar1

BASLCGELPERMEATLON

CHROMATOGRAPHY

STUDIES.I.1.

557

the size of the corresponding saturated compounds, but in benzene solvent the unsaturated structure appeared to change the molecular size consistantly by about 0.35 carbon atom units. Thus heptene-1 appeared only 6.65 carbon atoms long in benzene. There was some scatter to the data, but there was general agreement with this value, even among the more polar compounds. An interesting case is +-N=N+, which has an effective chain size of 4.06 carbon atoms long US. 4.05 calculated. Acetylenic unsaturation also briefly studied. The aliphatic acetylenic compounds studied appeared to be about x.5 C units smaller than their saturated analogues, and the correction factor was 0.75 C units smaller with phenylacetylene, a single example of an aromatic acetylenic compound. The relatively few adjusting factors in this report did not allow quite all the data to give high quality correlations, although further correction factors may be needed. In almost all of these deviating results, one observes changes in sorption due to the presence of additional polar or semi-polar sites, These examples are given in Tables XIII and XIV. However, an attempt is made to interpret Table XIII and some of the deviations in light of other observations in this report, or in most cases an attempt was made to determine the basic process that lead to changes in anticipated elution volumes in the current work. Ionol (z,6-di-tert.-butyl-+cresol) gave the expected elution, as did triphenylcarbinol and propylene carbonate. The nitrogen atoms of nitroethane and nitrobenzene were buried, in position, but they were effective in determing the effective GPC chain lengths in benzene. The nitrogen atoms of the nitriles studied did not contribute to the effective GPC chain lengths, which parallels results with the non-hindered amines studied, In the propylene glycol series, no correction of 0.15 times the lengths of the longest chain needed to be made; probably this reflected that the small propylene glycol homologues are not gel swellers and they are probably repelled from the gel surface in benzene. Similarly, no correction for sorption was needed with tetraethylene glycol in benzene in these GPC studies. In contrast, hydrogenated bisphenol “A” was smaller than one might initially anticipate, and the correction of 0.30 times the longest chain appeared applicable, as one might presume necessary for a bidentate compound. Sulfur atoms did not commonly contribute to the effective chain length of the compounds in the current study, as judged for example by comparing ut-propyl sulfide with rt-hexane; both eluted at close to the same elution volume. Further evaluation of the esters studied herein lead to the assumption that the ethereal oxygen of this group may be as much as 50 o/oeffective in determining effective GPC size in benzene. Further evaluations appear desirable for these compounds. A few compounds did not elute in benzene as early as one might predict from their revised chain lengths. In general, this was due to the presence of more than one polar and/or aromatic sites close together in the same molecule. The examples observed were methyl methacrylate, quinoline, $-methoxyphenol, and ðoxyethanol. While the GPC size of these compounds was predicted with an error of 20 % , the elution volume was in general predicted with an error of IO oA or less. One should also be on guard for a shift in calibration curve with benzene solvent and the GPC gels such as those used in this study. The calibration of columns I3 and C represent different curves with the same columns. Column D is also physically the same column as I3 and C, where only a few of the desired calibration points were available due to not realizing the likelihood of the change. In this case, the V$ term ,J.Chromatog., 32 (x968)543-558

55s

J. G. HENDRICKSON

for column D appeared to be 3.16 ml less.than for column C, and even a bigger difference was observed between columns B and C. The contributing factors to calibration change include aging, major changes in the solvent or samples reaching to install the columns in the GPC the columns, and connecting and reconnecting system. SUMMARY Elution volumes in gel permeation chromatography have been examined for go compounds using benzene as the elution solvent. Sorption to a degree was observed with almost all the aromatic and oxygenated compounds studied. Thus, elution volumes increased in the order ?z-heptane c gz-heptaldehyde < G-heptyl alcohol < g-heptanol < gz-heptanoic acid. After rc-heptanoic acid, toluene and then +cresol eluted, indicating the separation of aromatic from non-aromatic compounds having the same number of carbon atoms. Rules were proposed for predicting elution volumes for solutes in benzene with the 40 A gel permeation chromatography gel. There was a good correlation between observed and theoretical values. REFERENCES I J. G. HENDRICICSON AND J, C. hloon~~,J, Poly~zer Sci., A-I-4 (I~GG) 167. 2 J. G. HENDRICKSON, paper presented at the Fourlh IWwnational GPC Symposdunz, FZa.,May zrsl, rg.57. 3 l3. CORTXS-JONES, Nature, rgr (1961) 272. 4 L. E. MALEY,J. Po@mw sci,,8C (IgG5) 243. 5 I,.R. SNYDER, J, Gkrowcatog.,23 (1966) 388. J. Chromatog.,

32 (1968) 543-558

Miami

Reach.