Separation and Determination of Mono-, Di-, and Tripentaerythritol by

Mark A. Tapsak, Guo Hong Wang, Krist Azizian, and William P. Weber. Analytical ... William N. Hunter. 2000, ... Alexander B. Susàn , William P. Dunca...
3 downloads 0 Views 306KB Size
Separation and Determination of Mono-, Di-, and Tripentaerythritol by Programmed Temperature Gas Chromatography DONALD S. WIERSMA, ROSS

E. HOYLE, and HANS REMPIS Research Department, Canadian Chemical Co., ltd., Edmonton, A h ., Canada )A new method for analyzing mixtures of mono-, di-, and tripentaerythritol is presented. Polyhydric alcohols such as pentaerythritol can b e converted quantitatively to the acetate esters b y direct treatment with acetic anhydride. The monopentaerythritol tetraacetate, dipentaerythritol hexaacetate, and tripentaerythritol octaacetate are sufficiently stable and volatile to permit their separation by gas liquid chromatography. The acetylation reaction product is resolved using temperature programming techniques on a parallel dual column system. The acetate esters of pentaerythritol formals are also resolved and identified. A measure of the quantitative accuracy is given and agreement between hydroxyl value b y calculation and chemical procedure is illustrated.

P

(I) is prepared by the aldol condensation reaction of 3 moles of formaldehyde and 1 mole of acetaldehyde to yield the intermediate pentaerythrose (11). A Cannizzaro reaction of the intermediate with formaldehyde in the presence of a strong alkali base yields pentaerythritol.

reaction products found in commercial grades of pentaerythritol are dipentaerythritol (111),tripentaerythritol (IV), bispentaerythritol monoformal (V), and pentaerythritol cyclic monoformal (VI). CHzOH

determined by chemical procedures where the formal is hydrolyzed and the liberated formaldehyde is measured by volumetric or colorimetric means. This paper describes a gas chromatographic

CHzOH

I I

HOCH,-C-CH~OCH~-~-CH*OH

I

CHzOH

CHzOH I11

CH2OH

,

CH&H

CHzOH

I

I

HOCH~-C-CH~OCH,-C-CH~OCH~-C-CH~OH

I

I

CH2OI-I

I

CH&H IV

CHZOH

CHzOH CH2OH HOCH~-C-CHaOCH~OCH~-C---CH~OH I I

I

I

CHzOH

CHiOH V

ENTdERYTHRITOL

3HCHO

OH+ CH3CHO +

CHzOH HOCH,-C-CHO I

(1)

CHQOH I1 CHzOH HOCHz-C-CHO

+ HCHO + MOH

-+

CHzOH CHZOH 1

HOCH,-~-CH,OH

(2)

I

CHzOH I (where 11 represents an alkali metal) I n actual practice, the reaction is more complex in that polypentaerythritols and pentaerythritol formals are produced in side reactions. The main side

VI

Commercial grades of pentaerythritol are usually assessed on the basis of hydroxyl content and melting point. Relatively small errors in hydroxyl assay lead to erroneous interpretation as to di- and tripentaerythritol. The oldest and most common method for analyzing pentaerythritol mixtures is a gravimetric method reported b y Kraft ( 3 ) . The dibenzylidene acetal of pentaerythritol is made and the determination is not interfered with by polypentaerythritol in concentrations up to 20%. TJ7yler (4) reported a method for analyzing mono- and dipentaerythritol mixtures based on their respective water solubilities. Higher molecular weight pentaerythritols and pentaerythritol formals interfere. i l n infrared method reported b y Jaffe and Pinchas (2) based on the absorption due to the ether linkage of the polypentaerythritols is not specific for di- and tripentaerythritol. Esposito and Swann (1) used gas chromatography for identification of monopentaerythritol in synthetic resins. The pentaerythritol formals have been

procedure for the quantitative analysis of a mixture of pentaerythritols, which applies to all the components in the mixture and removes the limitations of the published methods. Commercial grades of pentaerythritol can he analyzed with greater certainty. EXPERIMENTAL

Apparatus. A Podbielniak S o . 9580 gas chromatograph modified for temperature programming techniques was used. This instrument was equipped with a thermal conductivity detector, Minneapolis-Honeywellprecorder, and a disk-type integrator. The analytical columns were constructed from 3/16-inch o.d., Yo. 316 stainless steel. Procedure. One gram of t h e pentaerythritol mixture was transferred to a 125-ml. iodine flask. Five milliliters of reagent grade acetic anhydride were added and t h e contents were refluxed for 2.5 hours. T h e reaction mixture was allowed t o cool and a sample was introduced onto t h e chromatographic column. T h e startVOL. 34, NO. 12, NOVEMBER 1962

* 1533

20

260

240

TENPERATURE 'C. 280 I

I

I1

Figure 1.

Separation and identification of polyol acetates 1. Solvent, acetic acid-acetic anhydride 2. Pentaerythritol cyclic monoformal diacetate 3. Pentaerythritol tetraacetate 4. Dipentaerythritol hexaacetate 5. Bispenlaerythritol monoformal hexaacetate 6. Tripentaerythritol octaacetate

Table I.

Analysis

of Pentaerythritol Mixtures Found, averAdded, Found, age,

Sample

Component

Synthetic blend A Pentaergthritol

Dipentaerythritol

Tripentaerythritol

Synthetic blend B Pentaerythritol

%

%

88.12

88.09 88.00 87.80 87.91

87.95

9.90 10.06 10.21 9.96

10.03

2.01 1.94 2.00 2.13

2.02

65.09 64.38 63.68 64.62

64.44

23.98 25.33 25.53 24.39

24.81

IO. 93 10.29 10.77 10.99

10.74

89.76 90.38 7.78 7.62 2.07 1.62 0.39 0.38

90.07

9.88

2.00

64.24

Dipentaergt,hritol

25.43

Tripentaerythritol

10.33

Typical sample of Pentaerythritol commercial Technical Grade Dipentaerythritol Tripentaerythritol Bispentaerythritol monoformal

1534 *

%

ANALYTICAL CHEMISTRY

7.70 1.85

0.38

Error,

o/o

Error, average, %b

-0.05 -0.13 -0.36 -0.23

-0.19

+0.20 +1.82 +3.34 $0.80

+1.50

$0.56 -3.00 0.00 $6.50

+1.00

+1.32 +o. 21 -0.87 $0.59

$0.31

-5.70 -0.39 $0.40 -4.08 +5.80 -0.38 $4.52 i-6.38

-2.40

f3.96

ing column temperature was 220" C and the column temperature was programmed a t a rate of 5" C. per minute u p t o a maximum temperature of 320' C. The peak areas were calculated and calibration factors were applied in order to obtain a weight per cent analysis. Analytical operating conditions are: column length, 4 feet; diameter, 3/16-in~h a d . ; solid support, Chromosorb W, 3Cto 60-mesh; liquid phase, SE-30, 15 wt. % concentration; temperature programmed a t 5" per minute (220" to 320' C.). Detector temperature, 325' C.; injection port temperature, 350' C.; helium flow, 140 ml.,/min.; detector current, 240 ma.; sample size, 5 pl.; recorder, 0.1 to 1.0 m.v. span, 1 see. pen speed. Calibration and Calculation. The quantitative aspects of the method were studied by preparing synthetic blends of t h e polyols. T h e pure mono-, di-, and tripentnerythritols were prepared by a series of evaporative crystallizations of impure mixtures. Pentaerythritol cyclic monoformal was synthesized by reacting pentaerythritol and formaldehyde under acid conditions and isolating the product b y extraction techniques. The bispentaerythritol monoformal mas isolated from pentaerythritol mixtures by extraction techniques. The thermal conductivity correcting factors were determined by analyzing the acetylated synthetic blends. The factors for mono-, di-, and tripentaerythritol were 0.89, 1.11, and 2.22, respectively, for synthetic blend B. Some differences were found in these factors a t significantly different ratios of mono-, di-, and tripentaerythritol. Using these factors, peak area was converted to weight per cent. When significant amounts of unknown material are present in a sample, the internal standard technique should be used for quantitative results. Because the formals were generally present in low concentrations, calibration factors were not determined but were assigned as one. DISCUSSION

The pentaerythritols are white crystalline compounds which readily sublime and decompose when heated above their melting points, but the acetate derivatives have sufficiently high vapor pressures and thermal stability to allow separation by gas chromatographic techniques. The acetylation reaction time can be accelerated by the use of an esterification catalyst. However, use of a catalyst such as p-toluene sulfonic acid resulted in the hydrolysis of the pentaerythritol formals. The completion of the acetylation reaction was determined by infrared and gas chromatographic techniques. The partially acetylated pentaerythritols were readily resolved from the fully acetylated members. A study was made of extracting the acetates from the reaction mixture with chloroform and analyzing the con-

centrated chloroform extract, b u t no advantages were derived from this procedure. =1 parallel dual column system was used to minimize the effect of substrate bleeding during the high temperature programming of the column. SE-30, a methyl silicone, gave satisfactory performance at the prescribed conditions. KO decrease in column performance was observed after four to fire months of operation. Other column substrates evaluated, Apiezon L and M, 1,4-butanediol succinate, and silicone 550, were unsatisfactory. Figure 1 shows a chromatogram of the polyol acetate mixture. The components are well resolved and good peak symmetry was obtained. T h e components are eluted in an increasing order of molecular weight. Column efficiency expressed as HETP was in the order of 0.08 to 0.10 cm. for the dipentaerythritol hexaacetate peak. The base-line drift was within acceptable limits to enable automatic integration of peak areas. The large concentration of acetic acidacetic anhydride solvent was not taken into account for the quantitative calculation. Peak identification was based on the respective retention times of the pure polyol acetates. Figure 2 shows a chromatogram of a typical commercial grade of pentaerythritol. Table I shows quantitative data obtained for synthetic blends and a typical commercial grade of pentaerythritol. The per cent error is greatest at the higher level of di- and tripentaerythritol. Commercial samples of pentaerythrito1 are generally assayed on the basis of hydroxyl value and melting point. Dipentaerythritol is then calculated from the hydroxyl value obtained. This basis for determining dipentaerythritol content is misleading q’ more dipentaerythritol is indicated than actually found directly by gas chromatographic procedure. The commonly used acetic anhydride-pyridine acetylation procedure for hydroxyl content was found to yield low results. High purity monopentaerythritol was analyzed by three different laboratories and found to be 0.2 to 0.3 wt. % lower than the theoretical value. Table I1 gives data on per cent h\ droxyl by acetylation procedure us. calculation from chromatographic analysis. The data show that better hydroxyl content agreement exists for synthetic blend A when the corrected hydro\;?-1 value by the acetylation procedure is used. The correction used was -0.25 wt. yo,a n average value obtained b y the three laboratories. I n the example of the typical commercial grade pentaerythritol, the hydroxyl value by acetylation procedure was also

260

240

Figure 2.

‘C. 280

320

Analysis of typical commercial grade pentaerythritol 1.

2. 3. 4.

5.

Table II.

TEMPERATURE

Solvent, acetic acid-acetic anhydride Pentaerythritol tetraacelate Dipentaeryihritol hexaacetate Bispentaerythritol monoformal hexaocetate Tripentaerythritol octaacetate

Per Cent Hydroxyl b y Acetylation Procedure vs. Calculation from Gas Chromatographic Analysis

Pure mono pentaerythritol,

5%

Added, mono di tri Found by gas chromat.ography, mono di tri formal Monopentaerythritol by dibenzylidene method Hydroxyl theoretical Hydroxyl by calculation from gas chromatographic analysis Hydroxyl by acetylation method Hydroxyl difference from theoretical by acetylation method Corrected hydroxyl by acetylation method

lowered by the presence of small amounts of impurities giving low or no hydroxyl contribution. Monopentaerythritol content by the independent dibenzylidene method was in good agreemerit with gas chromatographic analysis. ACKNOWLEDGMENT

The authors thank the personnel a t Celanese Chemical Co. Research Laboratory, Clarkwood, Tes., and Canadian Chemical c0.,Development and control Laboratories, Edmonton, Alta., for assisting in the hydroxyl analysis.

Synthetic blend A, % 88.12 9.88

Typical commercial grade pentaerythritol,

2.00

7c

87.95 10.03 2.02

90.07

50.0

88.22 48.75

89.09

50.00

48.72 48.42

48.92 48.41

-0.33 48.67

4s 66

99-100

49.75 -0.25

50.00

7.70

1.85 0.38

:

LITERATURE CITED

(1) Esposito, G. G., Swann, CHEhf.

pvl.

33, 1854-8

H,

(2) JafTe, H , Pinchas, S., Ibid., 23, 1164-5 (1951). ( 3 ) Kraft, M. Ya., J . Chem. ~ n d u s t r y (MOSCOW),8, 507 (1931), C . A . 25, 51148 (1931). (4) Wyler, J. A , , IXD. ENG.CHEY., ED. 777-8 (lg46). RECEIVED for review July 3, 1962. -kcepted August 27, 1962. Presented at the 45th Canadian Chemical Conference and Exhibition, The Chemical Institute of Canada, Edmonton, May 28-30, 1962.

VOL. 34, NO. 12, NOVEMBER 1962

0

1535