UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Degradation of dea treating solutions Choy, Edward Takwah 1978

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
UBC_1978_A7 C45.pdf [ 6.88MB ]
[if-you-see-this-DO-NOT-CLICK]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0058692.json
JSON-LD: 1.0058692+ld.json
RDF/XML (Pretty): 1.0058692.xml
RDF/JSON: 1.0058692+rdf.json
Turtle: 1.0058692+rdf-turtle.txt
N-Triples: 1.0058692+rdf-ntriples.txt
Original Record: 1.0058692 +original-record.json
Full Text
1.0058692.txt
Citation
1.0058692.ris

Full Text

DEGRADATION OF DEA TREATING SOLUTIONS by EDWARD TAKWAH CHOY A THESIS SUBMITTED IN PARTIAL FULFITJJVENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Chemical Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA"' MAY 1978 Edward Takwah Choy, 1978 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Chemical Engineering The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date 1st June. 1978 ABSTRACT The objectives of this research were (i) to develop an analytical method capable of measuring trace quantities of DEA degradation products and (ii) to study DEA degradation under carefully controlled experimental conditions. The analytical method consisted essentially of removing water by air stripping, silylating the residue with N,0-bis(trimethylsilyl) acetamide (or "BSA"), and analyzing the silylated compounds by gas chromatography. An OV-17 stainless steel column and ^ flame ionization detector were used. The effectiveness of this method was subsequently confirmed by its ability to analyze numerous DEA samples from laboratory experiments and industrial plants. DEA degradation was first studied at atmospheric pressure using an all glass and teflon apparatus. When a 30% DEA solution was contacted with pure C02 for up to 23 days, no known DEA degradation products were detected. However, the DEA concentration was observed to decrease with time. This may be caused by the formation of heat-stable salts which are undetectable by gas chromatography. DEA degradation experiments were then conducted at elevated pressures (up to 4238 kPa (600 psig)) and temperatures (165 to 185° C) using a stainless steel reactor. The results showed that DEA degradation proceeded rapidly in the presence of CC^ and N,N-bis(2-hydroxyethyl) piperazine (HEP) was one of the major degradation products. Four other i degradation products called T, X, Y and Z were also observed when the column temperature of the chromatograph was set to 142° C. The concentration of HEP was found to increase uniformly with time whereas concentrations of T and Y increased only slightly. The concentration versus time plots of compounds X and Z exhibited a maximum which suggests that they are degradation intermediates. The overall DEA loss was governed by a first order rate equation of the following form: d [PEA] _ dt " Aexp - EA /RT J^DEA J where A, Ea and R are 2.03xl010 hour"1, 2.17xl04 cal. g-mole-1, 1.99 cal g-mole 1 °K 1 respectively. A degradation mechanism is proposed which resembles that suggested by Polderman eX at. (16): U DEA + GO^ VSss>,HEOD + H20 >h HEP + CC>2 where U denotes one, or more, unidentified compounds. The theoretical concentration values of DEA and HEP based on the above mechanism fitted the experimental data fairly well. However, the predicted concentrations of HEOD did not exhibit a sharp maximum and therefore failed to match the behavior of X or Z. ii The DEA samples obtained from the high pressure tests were also analyzed using a higher G.C. column temperature of 174° C. In this case, seven additional peaks were observed indicating other degradation products. One of the seven peaks could be caused by THEED, which was previously reported by Hakka eX at. (17). It is therefore clear that the degradation of DEA is more complex than suggested by the above mechanism. Special high pressure tests were conducted to discover the effects of acid gas composition, pressure, initial DEA concentration and metals on DEA degradation. It was found that CO2 and temperature were the most important parameters affecting DEA degradation but purely thermal degradation was insignificant. (X^ pressure appeared to be unimportant, between 2170 and 4238 kPa (300 to 600 psig), probably due to a large excess of (X^. When the initial DEA con centration was 10% (as opposed to 30%), the degradation of DEA proceeded more slowly than expected from the aforementioned rate expression. The reason for this behavior is presently unknown. When the DEA solution was saturated with I^S at 1480 kPa (200 psig) and room temperature prior to contact with C02 at 4238 kPa (600 psig), it was observed that the degradation at 175° C decreased substantially. This was probably due to the fact that approximately half of the DEA combined with H9S and was hence protected against C09 attack. iii Initial results from preliminary paper chromatographic and mass spectroscopic studies are also reported. These studies provide guidance for future work.. iv TABLE OF CONTENTS Page ABSTRACT i LIST OF TABLES vii LIST OF FIGURES ix ACKNOWIEDGEMENTS xiI. TJSITRODUCTION 1 1. Specific Objectives of the Present Study 9 II. LITERATURE REVIEW 10 1. DEA Degradation2. Analytical Techniques for Amine Analysis 13 III. CHROMATOGRAPHIC ANALYSIS OF DEGRADED DEA SOLUTIONS 27 1. Introduction '22. Selection of a Suitable Analytical Technique 27 3. Adverse Properties of DEA and Its Degradation '35 Products for G. C. Analysis 4. Development of the Present Analytical Method 36 5. Choice of Silylation Reagent 39 6. Choice of Chromatographic Column ^2 7. Column Conditioning 44 8. Choice of Solvent 49. Drying of Aqueous DEA Samples 45J 10. - Internal Standard 46 11. Choice of Septum 48 12. Choice of Operating Conditions 49 13. Equipment Specifications 50 14. Sample Preparation Procedure : 50 15. Calibration Method for Quantitative Analysis 52 16. Maintenance of Analytical Equipment 55 IV. DEA DEGRADATION UNDER CONTROLLED CONDITIONS 57 1. Atmospheric Pressure Equipment 52. High Pressure Equipment 60 v Page V. DATA ANALYSIS 63 1. Theory of DEA Decomposition 62. Theory of Degradation Product Formation 65 VI. RESULTS AND DISCUSSION 67 1. Interpretation of Chromatograms 61.1 M&a&uAzmznt o{> VEA Fzaks 7 1.2 MtcUsufiQjnznt o& Dtg^adcution Vtwdu.eZ Peats 71 2. High Molecular Weight Degradation Products 74 3. Analytical Technique 76 4. Results of Atmospheric Pressure Tests 77 5. Results of High Pressure Tests 78 6. Results of the Kinetic Model Study 90 7. DEA Degradation Under Special Conditions 94 8. Analysis of Industrial DEA Sample 100 VII. CONCLUSIONS 101 VIII. RECOMMENDATION FOR FURTHER RESEARCH 105 IX. NOiMMCLATURE 108 X. BIBLIOGRAPHY 112 APPENDICES A GENERAL INFORT^ATION ON SIX G. C. ACTIVE PHASES RELEVANT 118 TO DEA ANALYSIS B IR AND NMR SPECTRA OF DEA AND RELATED DEGRADATION PRODUCTS 121 C PAPER CHROMATOGRAPHIC SEPARATION OF SOME DEA DEGRADATION 127 PRODUCTS D LISTING OF THE COMPUTER PROGRAM USED IN PREDICTING THE 129 THEORETICAL CONCENTRATIONS OF DEA, HEP AND HEOD AS WELL AS THE REACTION RATE CONSTANTS OF THE PROPOSED DEA DEGRADATION MODEL E INDUSTRIAL DEA SAMPLES 132 vi LIST OF TABLES Table Page 1 General information on three Western Canadian 2 gas fields. 2 Some possible degradation products from DEA gas- 7 treating solutions. 3 Analytical equipment and operating conditions used 18 by Brydia et al. (23). 4 Analytical equipment and its operating conditions 21 used by Piekos et at. (25). 5 Analytical equipment and operating conditions used 23 by Saha et al. (26) . 6 values and solvent systems related to the separation 24 or some hydroxyamines using paper chromatography (27). 7 Comparison of analytical techniques suitable for the 28 analysis of degraded DEA solutions. 8 Selected physical properties of ethanolamines and HEP. 37 9 Some comrron silylation reagents capable of forming 40 trimethylsilyl derivatives. 10 Analytical equipment and its operating conditions 51 for G. C. Analysis. 11 Operating conditions for the atmospheric pressure 77 equipment. 12 Results of high pressure tests. 81 13 Overall reaction rate constants obtained with the 88 high pressure equipment. (Operating conditions: pressure - 4238 kPa (600 psig); acid gas - pure (X^; initial DEA strength - 30 wt.%). 14 Reaction rate constants giving the best fit between 90 the theoretical and experimental HEP concentrations. vn Page 15 Experimental and theoretical concentrations of DEA, 391 HEP and HEOD. 16 Special experimental conditions and the results '97 obtained with the high pressure equipment. A-l General information on active phases relevant to 119 G. C. analysis of DEA solutions. E-l Analysis of industrial DEA gas-treating samples. 135 viii LIST CF FIGURES Figure Page 1 Flow sheet of a typical industrial amine unit. 4 2 Mass spectrum of DEA. 32 3 Mass spectrum of HEP. 3 4 Mass spectrum of TEHEED. 34 5 Calibration plot of the peak area ratios of DEA/DECA 53 as a function of the amount of DEA present. 6 Calibration plot of the peak area ratios of HEP/DECA 54 as a function of the amount of HEP present. 7 Schematic diagram of the atmospheric pressure apparatus 58 for DEA degradation reaction studies. 8 Schematic diagram of the high pressure apparatus for 62 DEA degradation raction studies. 9 Chromatogram of ethanolamines and some previously 68 reported degradation products. (Sensitivity: l-512xl0~10 AFS; 2-128xl0~11 AFS; 3-18x10-11 AFS; 4-128xl0_11 AFS; 5-16xl0_11 AFS; 6-32xl0_11 AFS; -12 7-4x10 AFS. Conditions as given in Table 10.) 10 Typical chromatogram of a lean DEA sample obtained with 69 the high pressure equipment. (Sensitivity: l-512xl0~10 AFS; 2-16xl0_1° AFS; 3-32xl0_11 AFS; -12 4-32x10 AFS. Conditions as given in Table 10.) 11 Chromatogram of parent, 30% DEA solution. (Sensitivity: 70 l-512xl0~10 AFS; 2-32xl0_1° AFS; 3-32xl0-11 AFS; -12 4-32x10 AFS. Conditions as given in Table 10.) 12 Illustration to show how triangulation method was used 73 to evaluate distorted G. C. peak areas. 13 A typical high temperature chromatogram of a lean DEA 75 ix page sample obtained with the high pressure equipment. 75 (Sensitivity: ^5UkLQ-10 ^ 2-16X10"10 AFS; 3-32xl0-11 AFS; 4-64xl0~12 AFS; 5-128xl0~12 AFS; 6-64xl0~12 AFS; 7-32x1 Cf12 AFS. Conditions as given in Table 10 with the column temperature at 174° C.) 14 DEA loss as a function of time obtained with the 79 atmospheric pressure equipment. (Operating conditions: pressure—101 kPa (14.7 psi); temperature—60° C; acid gas—pure G02; initial DEA strength—30%.) 15 DEA concentration as a function of time obtained 83 with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure C02; initial DEA strength—30 wt.%; the symbols are based on experimental measurements whereas the lines represent the model predictions.) 16 HEP concentration as a function of time obtained 84 with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure <X>2; initial DEA strength—30 wt.%; the symbols are based on experimental measurements whereas the lines represent the model predictions.) 17 Formation of compound X as a function of time obtained 85 with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure C02; initial DEA strength—30 wt.%; the symbols are based on experimental measurements whereas the lines were drawn to pass through these data points.) 18 Formation of compound Z as a function of time obtained 86 with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure (X>2; initial DEA strength—30 wt.%; the symbols are based on experimental measurements and the lines were drawn to pass through these data points.) 19 Overall DEA reaction rate constants as a function of 87 absolute temperature as listed in Table 13. 20 Formation of HEOD as a function of time predicted by the 95 kinetic model. x Page B-l IR spectra of (a) DEA and (b) HEOD. 122 B-2 IR spectra of (a) HEP and (b) TEHEED. 123 B-3 NMR spectrum of partially deuterated DEA. 124 B-4 NMR spectrum of HEP. 125 B-5 NMR spectrum of HEOD. 6 E-l Typical chrcmatogram of a DEA gas-treating sample 133 obtained from industrial source. (Sensitivity: l-512xl0"10 AFS; 2-32xl0~10 AFS; 3-8xl0~10 AFS; -11 -12 4-32x10 AFS; 5-32x10 AFS. Condition as given in Table 10.) E-2 Typical chromatogram of a DEA gas-treating sample 134 obtained from an industrial source. (Sensitivity: l-512xl0_10 AFS; 2-4xl0~10 AFS; 3-32xl0-11 AFS; 4-64x10~12 AFS; 5-32xl0~12 AFS. Condition as given in Table 10 with the column terrperature at 174° C.) xi ACKNOWLEDGEMENTS The author gratefully acknowledges the advice, guidance, encouragement and assistance of Dr. Axel Meisen during this work. I would also like to express my appreciation to the following persons: Drs. J.B. Farmer and J.B. Scheffer for their advice and permission to use their laboratory during the initial phases of the study; Dr. G.K. Eigendorf and other members of the staff in the mass spectaxametry laboratory for their advice and generous help in the operation of the mass spectrograph; Drs. J.G. Hooley and R.W. Lockhart for their assistance in lending apparatus for general testing; Ms. Beatrix Krizsan for the drafting of the figures and, finally, to numerous members of the Departments of Chemical Engineering and Chemistry for their advice, guidance and assistance. The author is also extremely grateful for the financial support provided by The Canadian Natural Gas Processing Association. xii 1 I. JJSITRODUCTICN The pressure and carposition of raw natural gas, i.e. gas as it leaves the geological formation, can vary greatly, but Table 1 gives some typical values for three Western Canadian gas fields. Most of the acid gases (H2S and <X>2) as well as water have to be removed before the gas can be sold in order to minimize pipeline corrosion:,^pumping costs, health hazards and pollution when the gas is finally burnt or converted. For Western Canada and the United States, the product natural gas should usually contain less than 0.57 grams/100 cubic meter (0.25 grains/100 SCF) of H2S and 6.4 grams/100 cubic meter (4 Lbs/SCFxlO ) of water (1). The removal of H2S and C02, or "sweetening", is therefore a major process used by the natural gas processing industry. Sweetening by absorption with aqueous ethanolamine solutions has found greatest acceptance (2, 3), although other processes (4, 5) are used as well. Monoethanolamine (MEA) is often preferred to diethanolamine (DEA) due to its lower cost and greater absorptive capacity on a unit weight basis, but DEA has the following advantages over MEA: greater resistance to COS and CS2 attack, smaller vapour pressure, lower affinity for hydrocarbons, less energy requirements (up to 35%) for regeneration (6), reduced corrosiveness, increased filter life 2 TABLE 1: General information on three Western Canadian gas fields Property Field I Field II Field III Pressure 6996 kPa 6996 kPa 6996 kPa (1000 psig) (1000 psig) (1000 psig) Temperature 32° C 41° C 43° C (90° F) (105° F) (110° F) Gas composition (mole %) N2 0.4 0.2 0.9 co2 2.2 6.4 5.2 H2S 9.2 34.5 27.7 Cl 82.4 58.2 66.0 c2 3.4 0.5 0.3 C3 1.2 • ^4 0.3 n-C4 0.4 i-c5 0.2 n-C5 0.2 Other hydrocarbons 0.1 0.1 0.2 >C3 " Water Saturated Saturated Saturated 3 by a factor of 5 to 10 (6) and reduced solution foaming. As a result, more and more gas treating plants are being converted to the DEA process (7). A flowsheet of a typical industrial DEA sweetening unit is shown in Figure 1. Sour natural gas containing CO2 and H2S is contacted with an aqueous DEA solution (15 to 30 wt.% of DEA) by passing it counter-currently through an absorber. This unit usually contains 10 to 20 plates, but packed absorbers are also in use. The gas leaving the absorber is sweet, i.e. essentially free of acid gases/, and it is passed through a knock-out drum to remove any entrained DEA solution. The sweet gas is then usually dehydrated before delivery to the user. The pressure of the rich DEA solution leaving the absorber is let down in a vent tank and the majority of absorbed hydro carbons are thus vaporized. After heating, the rich DEA solution is sent to a stripper where the absorbed H2S and C02 are released. The overhead acid gases are flared or, in case of high H2S con centrations, are sent to a Claus plant (8) for recovery of elemental sulphur. The regenerated lean DEA solution from the bottom of the stripper is cooled and recycled to the absorber. In order to prevent the build-up of impurities (such as iron sulphide and other 41 °C ABSORBER 6169 kPa SOUR ^ j GAS LEAN AMINE SOLUTION SWEET GAS 99 °C HYDROCARBONS 4 KNOCK-OUT DRUM (929 kPa) 127 °C STEAM f~^k EXCHANGER 91 °C/ REBOILER D u 198 kPa H20 CONDENSER ^ACiD GAS REFLUX. PUMP STRIPPER ACCUMULATOR I27°C 124°C HIGH PRESSURE PUMP BOOSTER SURGE PUMP TANK FIGURE 1: Flow sheet of a typical industrial amine unit. 5 non-regenerable conpounds), a small slip stream of the DEA solution may be passed through a filter system (7, 9). The principal reactions of DEA with i^S and C02 can be represented by the following equations (10): HO-CEL-CH, 38° C HO-O^-CH^^ + H2S 116P C H0-CH2-CH2N^ H0-CH2-CH2 HS (1) HO-CH -CH 38° C . \ (- / E0-CH2-Cil2^ m + H2S 116° C H0-CH2-CH2 NH H0-CH2-CH2/ 2 (2) H0-CH_-CH„ 2. 2> 38° C + C0„+H90 i H0-CH2-CH2'/ 1 149° C HO-CH^^ H0-CH2-CH2. NH, HC03 (3) 2 1 NH +C0o+Ho0 2 2* 38° C H0-CH2-CH2 149° C H0-CH2-CH2^ NH, CO. H0-CH2-CH2 (4) 6 The reactions are reversible with the forward (absorption step) being exothermic and the backward (desorption step) being endothermic. As explained more fully in the next Chapter, Eqs. 1 to 4 are accompanied by undesirable side reactions which result in amine degradation. The degradation products, some of which are summarized in Table 2, are either corrosive or give rise to corrosive compounds (1, 7). It is not possible.to purify DEA solutions by distillation alone and a majority of the gas treating plants control the build-up of degradation products by purging or activated carbon absorption (7). The effectiveness of absorption has however not yet been fully established. The DEA loss and increased corrosion can result in major costs and the incentive for suppressing the side reaction is therefore high. Plant experience has indicated that amine losses and corrosion rates are affected by the following operating variables (1, 7, 10): temperature, pressure, raw gas composition, amine concentration, pH of the amine solution and the presence of metal ions. The qaantttatlv£ relationship between these variables is however unknown. Although DEA samples from sweetening processes are routinely analyzed to monitor the purity of the solutions and to detect TABLE 2: Sore possible degradation products from DEA gas-treating solutions. NAME ABBREVIATION STRUCTURAL FORMULA MOLECULAR WEIGHT Hydroxyethyl imidazolidone HEI CH2—CH2 CH C = 0 2 N CH2-—CH2--OH 113 3-(2-hydroxyethyl)-2-oxazolidinone HEOD CH^N—CH2—CH2—OH CH2 C = 0 N0 131 N,N-bis(2-hydroxyethyl) piperazine HEP .CH2—CH2. HO-CH5—CH^—N^ ^N-CH^—CH5—OH \CH2-CH2/ 174 TABLE 2: Some possible degradation products from DEA gas-treating solutions, (continued) NAME ABBREVIATION STRUCTURAL FORMULA MOLECULAR WEIGHT N,N,N-tris(2-hydroxyethyl) ethylenediamine THEED HO-OLr—CH2. H H0-CH2—CH2>N-OT—ffl2-^2—CH2—°H 236 N-MetJiyldietJianolamine MDEA HO-CH^O^ HO-CH^CH2 N-OL. 119 Oxazolidone OZD CH—N-H t 2 CH2 OO 87 Arrinoethylethanolairiine AEEA H m~CH2—CH2"—^-CH^—CH2~0H 104 co 9 corrosion related materials (7, 11, 12), very little has been published in the open literature on the analytical precedures used and their accuracy. The analysis is difficult because DEA and its degradation products have fairly low vapour pressures, readily decompose at elevated temperatures, are highly polar, occur in aqueous solution and at small concentrations. Furthermore, any relationships between the rates of degradation and process conditions are generally unknown. 1. Specific Objectives of the Present Study The main objective of the present study was to develop a reliable, quantitative technique for determining DEA and its degradation products. A further objective was to relate the formation of degradation products to some processing parameters such as temperature, pressure and DEA concentration. ' II. LITERATURE REVIEW The DEA sweetening process has been widely accepted for removing CO2 and H^S from natural gas streams and numerous studies have been conducted using DEA as a sweetening reagent (7, 8, 9, 13). In addition to general handbooks (14, 15) on natural gas engineering, the "Gas Conditioning Fact Book" (5) published by The Dow Chemical Company describes various gas-treating methods commonly used. It also presents general analytical methods for routine analysis of industrial gas-treating solutions. However, there are surprisingly few references dealing specifically with DEA degradation. Furthermore, the open literature contains no information on fast, reliable and quantitative analytical methods suitable for measuring small concentrations of DEA and its degradation products in aqueous solutions. Such an analytical technique is of course a pre-requisite for quantitative studies on DEA degradation. The review of the literature is divided into two major parts: DEA Degradation and Analytical Techniques. 1. DEA Degradation Polderman and Steele (16) were the first to publish a comprehensive investigation on DEA degradation in 1955. The authors used a stainless steel pressure autoclave as a reaction vessel in 11 which a 25 wt. % solution of DEA was first saturated with C02 at 25° C. The vessel was then sealed and heated to the desired temperature (between 100 and 175° C). The pressure inside the autoclave ranged from 1342 (180 psig) to 4238 (600 psig) kPa. After 8 hours, the autoclave was cooled to 25° C and the contents were discharged and fractionally distilled. It was thus possible to determine the amount of DEA remaining and the amount converted to higher-boiling, nitrogenous compounds. The DEA conversion was reported to range from 0% (at 100° C and 1342 kPa (180 psig)) to 97% (at 175° C and 4238 kPa (600 psig)). Using fractional distillation and crystallization, the authors attempted to analyze the nature of the degradation products. A white crystalline solid was isolated from the DEA residue and was identified as N,N'—bis (2-hydroxyethy1) piperazine ("HEP"). The authors also noted the presence of other degradation products, but did not characterize them further owing to the lack of suitable analytical and separation techniques at that time. For the same reason, the concentrations of HEP in the reacted DEA solutions could not be determined accurately. The authors postulated the following reaction scheme for the formation of HEP: CH0—CH„ , 2 | 2 » H0-CTIo-CH~-N 0 + Ho0 (5) 2 2 v / 2 C II 0 3-[2-kydAoxytthyl)-l-oxazoLldcmt, "HE0V" H0-CH2-CH2s^ NH + C0~-H0-CH2-CH2/ DEA CH„-CH„ CHr-CH„ ,2,2 / 2 \ 2 H0-CH„-CHo-N 0 »HO-CH_-CH„-N N-CH^CH^-OH + 2C0o (6) II 0 N,N'-.bZi [l-hydn.0xy2X.kyt) pZpeAazlne, "HEP" Hakka zt aZ. (17) continued the work of Polderman &£ aZ. (16). Using modern techniques such as infrared (IR) spectroscopy, mass spectroscopy (M.S.), nuclear magnetic resonance (NMR), gas chroma tography (G.C.) and thin layer chromatography (TLC), the authors were able to isolate another degradation product, N,N,N'-tris (2-hydro-xyethyl)-ethylenediamine ("THEED"). Its chemical structure is shown below: H0-CH~-CH„V H 2 2V I J;N-CH5-CH„-N-CH9-CH9-OH H0-CH2-CH2^ THEEV According to Hakka aZ. (17), this degradation product occurred frequently at the level of 0.5 to 2 wt.% of the gas treating solution and was by far the major unidentified degradation product. The authors found that both HEP and THEED are able to combine with 13 CC>2 and ^S and that their basicity is equivalent to that of txiethanolamine (TEA). However, under gas-treating conditions, only one of the nitrogen atoms in the HEP or THEED molecule is likely to react with the acid gas. Hence, on a weight basis, the capacity of the gas-treating solution falls with increasing DEA degradation. In addition to the work cited above, Smith and Younger (7) discussed DEA degradation and mentioned several other degradation products reported by gas plant operators. Among them was a com pound which had the same retention time as triethanolarnine (TEA) in gas chromatographic analysis. Corrosion studies have been published by various researchers (S, 9, 16). DEA itself is not considered to be very corrosive, but degraded DEA solutions attack mild steel extremely readily (11, 12). So far, no quantitative studies that relate DEA degradation to processing conditions such as ^S/CC^ ratios,' solution loadings, temperature and pressure have been published. 2. Analytical Techniques for Amine Analysis Henry and Grennert (11, 12) in 1955 were amongst the first researchers interested in the detection and measurement of stable DEA salts present in refinery samples. The authors pointed out that acid gases other than I^S would also .react with DEA. If these materials were more strongly acidic in aqueous solution than hydrosulfuric acid, they would form heat-stable salts with DEA from which the latter would not be liberated during the regeneration process. These acidic constituents, although, usually present in refinery gases in very small concentrations relative to I^S, would still tend to accumulate in a closed system. This would result in a gradual decrease of the available (free) DEA, until finally its replenishment became necessary in order to maintain efficient absorption of I^S. Henry eX at.- (11, 12) investigated four types of acidic materials namely: (a) organic acids, (b) chloride, (c) cyanide and thiocyanate and (d) sulfite, sulfate and thiosulfate. The investigation of organic acids was performed by potentiometric titration. The authors observed two inflection points in the titration curves of lean DEA samples at pH 6.3 and 3.5 respectively. The second one (at pH = 3.5) was not found in the titration curve of fresh, unused DEA but would become more pronounced with the accumulation of heat-stable salts. Further investigation using different synthetic solutions of DEA with acetic acid additions led to the discovery that the difference between the first and second inflection points in the titration curve was found to be a good measure of the added acetic acid. Also, the first 15 inflection point in the potentiometric titration of refinery samples was a direct measure of the uncombined DEA. For pure DEA solutions, this value was obtained by titration to the methyl red end point, but for refinery DEA samples, the difference between the potentiometric value and the end point value increased with organic acid content. Hence this method might give a rather large discrepancy in the obtained results. Chlorides were found to have very little effect on the formation of heat-stable salt. On the other hand, cyanide was found to be able to react with constituents of H2S-enriched DEA solution to form thiocyanate, which reduced the available amine content of the solution. The principal reaction appears to be: (H0C2H4)2NH2 S2 + 2 CN (H0C2H4)2NH2 SCN (7) Further studies led to the conclusion that in solutions with polysulfides, conversion of cyanide to thiocyanate occurs almost inmediately. The formation of thiosulfate salt was reported by the authors to be analogous to the formation of thiocyanate: (H0C2H4)2NH2 S2 + 302- •^2 (H0C2H4)2NH2 S203 (8) It was found that the above reaction takes place in H2S-enriched DEA solutions in the presence of even trace amounts of oxygen. The rate of reaction depends on the degree of saturation of the DEA with H2S. For given H2S and 02 concentrations, the rate of formation of thiosulfate was found to increase with temperature. 16 Henry oX at. (11, 12) also discussed conventional wet cheirdstry methods such as titration and Kjeldahl total nitrogen determination, as well as other methods for the determination of total sulfur, sulfide, mercaptide, sulfate, thiocyanate, cyanide, chloride, carbonate, alkalinity and sodium. The above two publications are by far the most complete investigation on heat-stable DEA salts ever performed and published up to the present time. Most of these methods have since been-used extensively by industrial laboratories for the analysis of DEA gas-treating solutions. However, they fail to identify the presence of degraded, organic compounds even in a qualitative manner. Furthermore, in 1955 Polderman oX at. (16) noted some inconsistencies in the results obtained,-with the above analytical methods. For example, chemical analyses by conventional acid titration may show a decline of DEA concentration in an industrial sample whereas the Kjeldahl total nitrogen determination may indicate that the sample contains more organic nitrogen than was possible in the form of DEA. In 1957, the Dow Chemical Company published the "Gas Conditioning Fact Book" (5) which listed all conventional, wet chemistry methods for the testing of DEA gas-treating samples. However, these procedures are unsuitable for identifying DEA degradation products. Also, since the book was written in 1957, it does not cover modern techniques such as gas chromatography (18, 19), Mass spectroscopy (20, 21) and thin layer chromatography (18, 22). 17 Brydia et at. (23) discussed the inability of direct gas chromatographic methods to analyze ethanolamines quantitatively. This is because of the following adverse properties associated with these types of compounds: low vapour pressure (< 1.3 Pa (0.01 mm Hg) @ 20° C for DEA) , high polarity and instability at elevated temperatures, especially in the presence of metals (24). The authors showed however that G.C. analysis became possible once the ethanolamines are reacted with trifluoroacetyl("TFA") anhydride. The latter compound attacks the amino and hydroxyl groups of DEA and forms a DEA—trifluoroacetyl derivative: S 0 H0-CH9-CH« CF^-C' F-,C-C-0-CH--CH9 v II \ju + o >> > J ^ ^N-C-CF^ + 3CF.,CCOH (9) H0-CHo-CHo/ CF0-C, FoC-C-0-CHo-CH„/ 2 2 0 II 0 PEA TFA anhydride PEA-TFA derivative TFA acid The analytical equipment and operating conditions are summarized in Table 3. This method is fairly simple, rapid (complete elution of MEA, DEA and TEA takes less than 14 minutes) and capable of detecting some impurities which are unobtainable by chemical means. However, TABLE 3: Analytical equipment and operating conditions used by Brydia 2Jt al. (23). Gas Ckk.omatoQK.apk Manufacturer — F & M Scientific Corp. Model — 810 with dual column Detector — Thermal conductivity detector with W-2 filaments. ChA.omatogn.apklc Column Material — Aluminum Dimensions — V O.D., 5' long * Packing — 5% neopentylglycol succinate coated on 60/100 mesh chromosorb G. Op&iating Conditions Injector port temperature Column temperature Detector temperature Detector current Carrier gas Carrier gas flow Injected sample size 250° C 172° C 250° C 185 mA Helium 30 ml/min 5 pi Refer to Appendix A for a more detailed description of this material. when water is present, TFA anhydride reacts with it to give TFA acid (CF3CCOH) which results in severe G.C. peak tailing. The authors also used a flame ionization detector and found an increase in the sensitivity and decrease in TFA acid peak tailing. However, an unexplainable decrease in precision of the G.C. results was also observed. by Brydia tt at. (23) by converting the ethanolamine mixtures to their trimethylsilyl ("TMS") derivatives instead. The silylation reagent, N,Obis (triitethylsilyl) acetamide ("BSA") reacts with both the amino and hydroxyl groups of ethanolamines as shown in Eg. 10: Piekos zt at. (25) eliminated the shortcomings discussed H0-CH„-CH. 2 .O-Si (CH,) 3 3 3 ^N-Si(CH,) 3 3 PEA BSA 'NH-Si (CH ) J3 ,0 (10) PEA - TMS Vnnlvativd- ' MSA The abbreviation MSA dmotoA U-t/vumeXhytJ>iZytM.aet^ldt. 20 The analytical equipment together with the operating conditions are summarized in Table 4. The authors also reported the conditions under which ethanolamines reacted with BSA quantitatively (25). The primary advantages of this method are simplicity and rapidity. Complete elution of the MEA, DEA and TEA derivatives requires less than 10 minutes. MSA, excess BSA and trimethylsilanol (a by-product from the water - BSA reaction) are eluted ahead of the other com pounds. Separate experiments showed that water concentrations up to 5% could be tolerated provided a large excess of BSA was employed. Just recently, Saha oX al. (26) renewed the interest in direct G.C. analysis of ethanolamines. The authors indicated that several disadvantages arise by converting the amines to G.C. stable derivatives prior to analysis. For example, derivative preparation is time consuming, derivative reactions may be incomplete and the derivatives may not be stable for long periods. These problems are of particular significance to laboratories where large numbers of samples have to be analyzed. On the other hand, hydroxy1 and amino groups have a strong affinity for most column support materials. This leads to long elution times, significant peak broadening and peak asymmetry. Saha oX al. (26) consequently investigated the use of organic TABLE 4: Analytical equipment and its operating conditions used by Piekos et at. (25). Gas Chromatograph Manufacturer Model Detector W.G. Pye & Co. Ltd. 104 H2 flame ionization detector. Chromatographic. Column Material Dimensions Packing Glass - 0.4 cm. I.D., 5' long 3% OV-1 coated on 100/120 mesh Diatomite CQ.* Operating Conditions Injector port temperature Column temperature Detector temperature Carrier gas Carrier gas flow Injected sample size - 190° C - Isothermal at 130 C for 2 min. and 15 sec. then temperature programmed to 180° C at 49°/min. - 210° C - Argon or Nitrogen - 19 ml/min - 0.4 - 0.7 ul * Refer to Appendix A for a more detailed description of this material. polymer beads with varying pore geometries and weak interacting surfaces as potential G.C. column packing materials. These compounds, being in the solid state at the operating temperature, were expected to permit rapid mass transfer and hence fast elution and sharp peaks. Saha oX at. (26) discovered that a new column packing, * Tenax-GC , which is a porous polymer based on 2, 6-diphenyl-P-phenylene oxide was able to separate ethanolamines with excellent results. The analytical equipment and operating conditions are summarized in Table 5. The authors reported that by using the above operating conditions, it was possible to separate a mixture of MEA, DEA and TEA in less than 8 minutes with excellent resolution, reprodu cibility and accuracy. Furthermore, aqueous solutions of amines could be injected into the chromatograph directly without difficulty. Zimmerman oX at. (27) investigated the separation of amines # and related compounds by paper chromatography. The values obtained for three different solvents are summarized in Table 6. * Refer to Appendix A for a more detailed description of this material. # Rf value is defined as the distance of sample spot from the origin divided by the distance travelled by the solvent front from the same origin. TABLE 5: Analytical equipment and operating conditions used by Sana e£ al. (26). Gas Chn.omatoQK.aph Manufacturer Model Detector Perkin - Elmer Corp. 900 flame ionization detector. ChKomatogKaphlc Column Material Dimensions Packing Stainless steel 1/8" O.D., 4' long Tenax - G C porous polymer column packing, 60/80 mesh. Qp&iatlng Conditions Injector port temperature Column temperature Detector temperature Carrier gas Carrier gas flow 350° C Isothermal at 135 C for 1 min. then temperature programmed to 350° at 16° C/min. 325° C Nitrogen 25 ml/min TABLE 6: values and solvent systems related to the separation of some hydroxyamines using paper chromatography (27). Name Solvent System A B C Etiianolamine 0.19 0.32 0.54 Diethanolamine 0.21 0.47 0.62 Triethanolamine 0.22 0.45 0.62 HEP 0.13 0.43 0.55 N,N,N1,N'-Tetrakis- 0.29 0.48 0.66 (hydroxyethyl)-ethylene-* diamine , ("TEHEED") * This compound is very similar to THEFT) but is cjuartenary substituted by the hydroxyethyl group instead of tertiary substituted. The solvent systems employed were: Solvent A : n-butanol, 77%; acetic acid, 6%; water, 17% Solvent B : n-butanol saturated with 0-1% aqueous NH^ Solvent C : pyridine, 42%; ethyl acetate, 25%; water, 25% With solvent system A and B, the migration was allowed to proceed for 4 hours. With sorveaafe". C, which moved more rapidly, the migration was stopped after 3 hours. Upon removal of the paper chromatograms, the solvent front was marked while still wet. After air-drying, the chromatograms were sprayed with 0.25% ninhydrin in 1:1 pyridine-ethanol and maintained at 105° C for 5 minutes to bring out the colour of the spots. The authors observed that solvent systems A and B gave extensive tailing of the spots. In many cases, this made visual detection very difficult. Qn the other hand, so'lvent C gave good, coherent spots but the migration was comparatively rapid, leading to poorer separations. Amines and amine related compounds are always of interest to scientists in biological fields. Extensive studies have been published in the analysis of amines in the past which include salting out chromatography (28), fractional distillation, potentio metric titration (11, 12, 29), Kjeldahl nitrogen (5, 30) and Van Slyke analyses (5, 30). Modern publications using gas chromatographic (31, 32, 33, 34, 35) and thin layer chromatographic (27, 36) techniques are also common. III. CHROMATOGRAPHIC ANALYSIS OF DEGRADED DEA SOLUTIONS 1. Introduction Before a study of DEA degradation could be undertaken, a reliable, quantitative method had to be available for the analysis of both residue DEA and trace quantities of degradation products. This method should be reliable, rapid, highly sensitive and require minimal sample size and preparation time. In addition, it should be applicable to DEA solutions degraded under laboratory as well as industrial conditions. 2. Selection of a Suitable Analytical Technique As mentioned in the previous chapter, virtually no information has been published on an analytical technique with the aforementioned attributes. Consequently the development of such a technique was the primary objective of the present study. Since conventional wet chemistry methods lack the desired reliability, accuracy and specificity, attention had to be focused on modern analytical methods such as infrared (37) and ultraviolet (38, 39) spectroscopy, mass spectroscopy (20, 21) as well as paper (18, 22, 27), high pressure liquid (18, 40) and gas (18, 19) chromatography. A brief comparison of these techniques is given in Table 7. TABLE 7: Comparison of analytical techni Technique Principle Infrared spectroscopy Characteristic IR absorption of functional groups - OH and - NH in 3600-3700 and 3400 crn-1 regions respectively. Ultraviolet Characteristic UV spectroscopy absorption bands in the region 190-400 mu. Optical spectroscopy Characteristic optical absorption bands in the region 400-800 mu. s suitable for the analysis of degraded DEA solutions. Advantages High sensitivity, rapid, very small sample requirement, simple operation and low cost; unique identification of organic compounds by its functional groups and "finger print" region (1600-600cm_1) IR absorption characteristics, very good documentation on IR spectra available from the literature. Disadvantages Applicability extends to single component, qualitative identification only; inability to analyze samples in aqueous and most organic solvents; lack of standards or information on degradation products; instrument not readily available. High sensitivity, rapid, very small sample requirement, simple operation and low cost; ability to analyze simple mixtures in aqueous or organic solvents quantitatively; instrument readily available; good documentation on UV spectra. No distinct UV absorption bands for DEA and its known degradation products, inability to convert the above compounds to UV absorptive species by chemical reaction. High sensitivity, rapid, very small sample requirement and simple operation; ability to analyze simple mixtures in aqueous or organic solvents quantitatively; instrument readily available. No distinct optical absorption bands for DEA and its known degradation products; inability to convert the above compounds to optically absorptive species by chemical reaction. TABLE 7: Comparison of analytical techniques suitable for the analysis of degraded DEA solutions. (Continued) Technique Principle Advantages Disadvantages Mass spectroscopy Characteristic ion fragmentation pattern and molecular ion peak detection. Extreme sensitivity and small sample requirement; ability to analyze simple mixtures quantitatively; unique characterization and identification of compounds by molecular weights and ion fragmentation patterns; good documentation on M.S. spectra. Inability to analyze DEA samples due to low volatility and fragile compounds present; lack of standards or information on degraded compounds relevant to M.S. analysis; difficult operation and high cost; instrument not readily available. Papersf.c;;--. vp' v chromatography Separation and identifi cation by differences in partition coefficients between two solvents. Combination of separation and identification technique with high sensitivity, small sample size, simple and low operation cost; ability to analyze mixtures in aqueous and organic solvents; simultaneous multiple analyses possible at room temperature; suitable for the analysis of high molecular weight fragile compounds with very simple detection devices; good documentation on paper or thin layer chromatograms. Very slowand useable mainly as a qualitative identification technique; inavailability of some degradation products as standards and incomplete knowledge of DEA degradation products. vo TABLE 7: Comparison of analytical techniqu Technique High pressure liquid chromatography Principle Separation and identifi cation by differences in partitation coefficients between two phases. Gas Separation and chromatography identification by differences in partitation coefficients between the two phases. suitable for the analysis of degraded DEA solutions, (continued) Advantages Combination of separation and identification technique with high sensitivity, rapidity, small sample size, simple and low operation cost; ability to analyze mixtures in aqueous and organic :. -,.. solvents quantitatively; analysis at room temperature is particularly suitable for high molecular weight, fragile compounds; large variety of detection devices available and good documentation of this technique is available. Disadvantages UV detection is impossible; instrument is unavailable for long term usage; inavailability of some degradation products as standards and incomplete knowledge of DEA degradation. Combination of separation and identification technique with high sensitivity, rapidity, small sample size, simple and low operation cost; ability to analyze mixtures in aqueous and organic solvents quanti tatively; capability to operate at room and elevated temperatures for effective separation of high molecular weight, low volatility com pounds; large variety of detection devices available and good documentation of this technique is available; the instrument is readily available Heavy, fragile compounds might decompose in G.C. columns, high polarity compounds are difficult to separate; inavailability of some degradation products as standards and incomplete retention data of DEA degradation products. OJ o 31 * Infrared and ultraviolet spectroscopic techniques were quickly eliminated due to the disadvantages listed in Table 7. However, once individual components of DEA degradation products can be isolated, IR spectroscopy will become an invaluable characterization technique due to its ability to detect functional groups (37, 41). Initially, mass spectroscopy was thought to be the ideal analytical technique for the present study because of its outstanding advantages listed in Table 7. Mass spectra of HEP and TEHEED, which were not available from the literature, were determined in this study. Figure 2 to 4 show the mass spectra of DEA, HEP and TEHEED. However, this technique was finally abandoned because of the disadvantages also listed in Table 7. Should all these .undesirable: factors be overcome, M.S. would once more be used extensively because of its unique characterization abilities. From the drawbacks associated with the above three analytical techniques, it was concluded that a combined separation and identi fication technique was necessary for the present study. Paper chromatography was first investigated. Initial results, which are discussed in Appendix C, showed * Refer to Appendix B for IR spectra of DEA and some degradation products. 100 801 g 601 DEA HO-CHg-CH. HO-CH-CH 2 ^ "2 parent mass 105 40 20 0 >0 40 60 80 100 MASS 120 140 160 180 200 FIGURE 2: Mass Spectrum of DEA. co INTENSITY ee 100 80 •z 40 10] 0 T— 20 40 60 il H0-CH2-CH2V TEHEED :N-CH2-CH2-N CH,-CHo-~OH HO-CHrCHf Dore;moss23S CH2-CHg-OH Ml -r I 1 80 100 120 MASS 140 160 180 200 220 240 FIGURE 4: Mass Spectrum of TEHEED. great premise for this method as a qualitative separation technique. Ultimately, paper chromatography may be coupled with M.S. to give quantitative results. High pressure liquid chromatography was also tested briefly using equipment available in the Chemistry Department. Although the initial experimental results were encouraging, this method had to be abandoned due to lack of access to the apparatus. Hence, after several months of trials with the above six analytical techniques, it was decided to develop a gas chromatographic method for the present DEA study. 3. Adverse Properties of DEA and Its Degradation Products for  G.C. Analysis Until recently, the usage of G.C. for the analysis of hydroxyamines and related compounds has proved^to^be .very difficult. This is due to four main adverse factors: low vapour pressure (< 1.3 Pa (0.01mm Hg) @ 20° C), high polarity, instability at elevated temperatures (especially in the presence of metals (24)) and simultaneous occurrence with water. The low vapour pressure requires DEA to be subjected to high injector and column temperatures for effective G.C. analysis. However, the poor thermal stability of DEA enhances the probability of amine decomposition and therefore leads to poor cnjantitative measurements and reproducibility. The polar hydroxyl and amino groups have a high affinity for most column packings and this results in long elution times, large peak broadening and peak asymmetry. Water is a further complicating factor since only a few column packings (18, 19, 31) can tolerate aqueous amine samples. If it is desired to extract DEA and its degradation products with an organic solvent, difficulties arise with selecting a suitable solvent. These various shortcomings had to be overcome before G.C. could be used for the effective analysis of aqueous, degraded DEA solutions. Table 8 lists some physical properties of ethanolamines and HEP relevant to the present study. 4. Development of the Present Analytical Method Derivative G.C. methods, as proposed by Brydia <it at. (23) and Piekos oX at. (25), indicated that the volatility and polarity characteristics of DEA solutions could be overcome. However, four main disadvantages arose with the derivative methods: (i) the inability to form amine-derivatives in the presence of excess water, due to the extreme reactivity of silylation reagents with moisture, (ii) the possibility of incomplete silylation resulting in extra G.C. peaks caused by partially silylated compounds, TABLE 8: Selected physical properties of ethanolamines and HEP. Property MEA DEA TEA HEP Formula H2NCH2CH2OH HN(CH2CH2OH)2 N(CH2CH20H)3 C8H18N2°2 Molecular weight 61.08 105.14 149.19 174.24 Sp. gr. @ 20° C 1.0179 1.0919 1.1258 a Sp.gr./at, @ 20 to 30° C 0.00078 0.00055 30 to 40° C 0.00065 Vapour pressure, @ 20° C <133 Pa <1.3 Pa <1.3 Pa (1 mm Hg) (0.01 mm Hg) (0.01 mm Hg) Boiling point, @ 101 kPa (760 mm Hg) 170.4° C 268.4° C (dee.) 335.4° C (dec.) 215-220° C 6.67 kPa( 50 mm Hg) 245 4.00 kPa( 30 mm Hg) 143 237 13.3 Pa(0.10 mm Hg) 71 152 205 Flash point 93°.C 138° C 193° C Viscosity @ 20° C, Cp 24.1 867 921 TABLE 8: Selected physical properties of ethanolamines and HEP. (continued), Property MEA DEA TEA HEP Melting point, C 10.3 28.0 21.6 133.5-136 Refractive index, n^ @ 20° C 30° C 1.4549 1.4747 1.4852 20 o Anv-jj/flt @ 20 to 40 C 0.00034 0.00028 0.00020 Solubility (§20 C in water water in Complete Complete 95.4 wt.% Complete Complete Very soluble Solubility in organic liquids @ 25 C acetone Complete benzene 0.6 wt.% carbon tetrachloride 0.1 ethyl ether 0.7 heptane 0.1 methanol Complete dimethyl formamide Complete Complete 0.03 wt.% 0.01 0.5 0.03 Complete Complete Complete 2 wt.% Complete 2 wt.% 0.03 Complete Complete Insoluble Insoluble Slightly soluble Insoluble Insoluble Very soluble Soluble u> oo (iii) the long-term instability of the derivatives and (iv) the limited choice of silylation solvents due to the high reactivity of the reagent and the low solubility of various DEA degradation products. These drawbacks made the application of derivative G.C. to the^present work quite difficult. 5. Choice of Silylation Reagent Since its discovery, silylation reagents have been widely used to react with the active hydrogen in hydrogen-bonded materials such as carboxylic acids, alcohols, airdnes and amides to produce fairly stable, easily separated and identifiable derivatives (33, 42, 43, 44). Even very delicate and complex biochemicals such as carbohydrates, amino acids, peptides and steroids can be analyzed by G.C. after stabilization with suitable silylation reagents (32, 33, 43) Table 9 lists some commonly used silylation reagents capable of forming trimethylsilyl (TMS) derivatives with a large variety of important chemicals having active hydrogen atoms (42, 44). However, all these reagents have one common undesirable property, viz. their extreme sensitivity to moisture. A comprehensive study of »BSA silylating properties was made by Klebe QJt aJL. (45). Their investigation showed BSA to be a highly effective silylation reagent. Fiirthermore, it is rather TABLE 9: Some common silylation reagents capable of forming trimethylsilyl derivatives. Abbre Chemical Molecular Boiling Name viation formula weight point Usage Special features N, O-bis-(trimethylsilyl) acetamide BSA Si(CHQ) 6 3 3 CH3-C=N-Si(CH3) 203.43 71-73°C @ 35 mm General silylation N,O-bis-(trimethylsilyl) trifluoroacetamide BSTFA Si(CH ) 1 J J 0 1 CFQ-C=N-Si(CH,) 3 3 3 .:257.40 40° C @ 12 mm General silylation Greater volatility than BSA. N-methyl-N-TMS trifluoroacetamide MSTFA CF3-C0NCH3 Si(CHQ) 3 3 199.25 132° C General silylation Most volatile derivative with extremely low retention time of its by-products. Hexamethyldisilazane HMDS (CH„)-SiNHSi(CHQ) 3 3  3 . :. 161.41 126° C Carbo- ' ~:. hydrates Evolves NH3 during silylation reaction. Trimethylchlorosilane TMCS (CHo)-SiCl 3 ... .108.7 57° C Biochemicals Often used as catalyst for silylation reaction} evolves Cl2 during silylation reaction. o inexpensive and does not produce noxious, gaseous by-products. Hence, it was chosen to be the silylation reagent for the present study. The silylation of DEA and HEP can be represented by the following equations: H0-CH„-CH„. ^O-SifCH.,) 2 2\ / .33 >N-H + 3 CH_,-C' H0-CH2-CH2^ 3 "^N-Si(CH3)3 VEA BSA (CH0) Si-0-CHO-CHZ NSsN-Si(CH3) + 3 CH3-C" (11) (CH3) Si-0-CH2-CH2^ 3 \N-SKCH3) 3 2 2> c i * <• 1 H VEA - TMS VeMvaMvz MSA The abbreviation MSA denotes N-trimethylsilylacetamide. Similarly the reaction between HEP and BSA is given by: .CH0-CH0V O-Si (CH,) / 2 2\ / 3 3 H0-CH9-CH„-Ns. ^N-CH^-CH^-OH + 2 CH,-CV Z Z N^-CH-/ Z A J ^N-Si(CH,) A Z -5-3 HE? BSA 42 ^CH2-CH (CH-)-Si-0-CHO-CH„-N ^N-CHO-CHo-0-Si-(CH,) + 2 CR-.-C (12) XCH2-CH/ N-Si(CH3)3 H HEP-TMS PeUvotu/e. MSA 6. Choice of Chromatographic Column The G.C. column is the heart of the Chromatograph where actual separation of sample components is achieved. For direct G.C. analysis, the choice of columns is often the most critical factor, due to the strong physico-chemical interactions and slow diffusion of the compounds to be separated. For derivative G.C. analysis, however, the choice of silylation method is more important, since this method alleviates factors that prevent direct G.C. analysis in the first place. High molecular weight, silicone stationary phases often prove to be quite satisfactory for the separation of silylated compounds (32, 33, 40, 41, 44). During the initial development of a suitable G.C. method for analyzing DEA gas-treating solutions, a V O.D., 6' long glass * column containing 4% Carbowax 20 M on 60/80 mesh Carbopack B deactivated with 0.8% KOH was purchased from the Supelco, Inc. Carbopack is a highly inert graphitized carbon stationary phase Refer to Appendix A for a more detailed description of this material. 43 and the above column was reported to be very effective in the separation of trace amounts of aliphatic amine mixtures in aqueous solutions (31). Initial direct G.C. analysis of methyl and t-butyl amines confirmed this. However, when aqueous DEA. solutions were subjected to the same"procedure, extreme sensitivity settings were required to produce a small DEA peak. This was due to the low volatility of DEA up to 140° C. At higher temperatures, the DEA started to decompose in the injector and/or column and deactivated the column. This showed that direct G.C. analysis using the Carbopack column was not suitable for the analysis of ethanolamines. Attention was therefore switched to G.C. derivative analysis. In the course of the development of DEA silylation techniques, two OV columns have been tested extensively. The first one was a * 1/8" O.D., 6' long glass column containing 10% OV-11 on 100/120 mesh chromosorb WHP purchased from Supelco Chromatography Supplies. The second was a 1/8" O.D., 6' long stainless steel column containing * 8% OV-17 on 80/100 mesh Chromosorb WHP supplied by Western Chromatography Supplies. Tests showed that both columns separated DEA and its degradation compounds effectively even at very low concentrations. Although the OV-11 glass column gave somewhat better resolution and less baseline drifting, the OV-17 stainless steel column was chosen because it was cheaper,; more rugged and had somewhat shorter elution times at the same column temperatures. * Refer to Appendix A for a more detailed description of this material. 7. Column Conditdoning Columns used for the analysis of silylation derivatives must be thoroughly conditioned before use. A powerful column conditioner, Silyl-8, has been developed for such purpose (44). This reagent is a silylating reagent which reacts with the active sites in the column and renders' them inert. 10 to 25 ul samples of Silyl-8 are injected directly into the chromatograph with the column at a temperature of 150° C and the flame ionization detector in operating mode. Such treatment should be repeated until a stable baseline is attained. Thereafter, samples containing BSA should keep the column conditioned. However, if the column has been used for non-silylated samples, it might be necessary to repeat the above conditioning treatment to remove non-volatile residues which can be silylated. 8. Choice of Solvent The selection of a solvent to dissolve degraded DEA samples for silylation proved to be quite challenging since the solvent must not react with the silylation reagent. Thus many comnDn organic solvents such as alcohols cannot be used because they contain active hydrogen atoms. The following solvents which are commonly used for silylation reactions (44) , were therefore evaluated for the present study: acetonitrile (CH,CN), dimethyl sulfoxide ((CHJ SO), furan (C^H^O) and dinethyl fomiamide (HCON(CH3) 2). Acetonitrile, tetxahydrofuran, pyridine, Chloroform and furan had to be eliminated because they did not dissolve HEP and other DEA degradation products. Dimethyl sulfoxide (DMSO) and dimethyl formamide (DMF) did not suffer from this disadvantage but the former is expensive and hazardous. Consequently, DMF was finally chosen as the solvent. A further advantage of DMF is its high boiling point (153° C @ S.T.P.) which made high temperature silylation possible without appreciable loss of solvent. DMF samples containing reacted DEA residues were also found to be stable for months if the samples were kept at room temperature. 9. Drying of Aqueous DEA Samples As mentioned previously, BSA and other silylation reagents react quickly with water (42, 44): O-Si(CH-) a* 3 + H„0 >(CHJ— Si-O-Si-(CH J +•• CH~-C-NH. (13) N-Si(CH-) 3 3 BSA 46 Mthough some attempts have been made to perform silylation in the presence of water (42, 43, 46), this approach was not adopted here since the effect of acidic gases in the DEA samples was unknown. Instead, the aqueous DEA samples were dried by placing them in a sand bath at 80° C and passing bone-dry air over them for 40 minutes. To remove the last trace of water, a few millilitres of methylene chloride were added to the DEA residue and the evaporation repeated to ensure the complete azeotropic removal of water (32). This procedure also stripped l^S and CG^ from the samples. 10. Internal Standard In quantitative G.C. analysis, the peak area is usually used to determine concentration and all operating parameters, which influence the peak size, must therefore be held constant. Since it is very difficult to keep parameters such as carrier flow rate, column temperature, injected sample size etc. absolutely constant, an "internal standard" is frequently employed. In this method a substance, is added to each sample in fixed amounts. The peak size of the sample relative to that of the internal standard is then used for all concentration determinations. Moderate variations in operating parameters do not affect the quantitative determination of concentrations since both the sample and the internal standard are affected to a similar extent. The requirements for an internal standard are: 47 (i) Elution times similar to but not identical to other peaks. (ii) Availability in high purity and at reasonable cost. (iii) Concentrations similar to those of unknown compounds. For derivative G.C. analysis involving BSA as a silylation reagent, various internal standards have been proposed depending on the types of compounds to be separated. Brittain oX at. (42) in their work on sugar analysis suggested 2, 3'-difluorobenzo-phenone and tetraphenyl silane for separations at high column temperature (175 to 252° C). Gehrke eX at. (32) in their analysis of amino acids reported phenanthrene, decanolic acid and fluorene to be suitable internal standards for column temperature between 154 and 265° C. After some initial trials, the column temperature was chosen to be 142° C for optimum separations of DEA, HEP and other degradation products. Based on the above literature sources, one would therefore suspected a member of the fatty acid family to be an appropriate internal standard. This fatty acid should produce a trimethylsilyl derivative with an elution time slightly longer than DEA-TMS (3.5 min. @ 142° C column temperature). This would ensure maximum accuracy in the determination of DEA concentrations since any deviation" the operating conditions would affect the DEA and the internal standard in a very similar manner: Furthermore, the elution of the internal 48 standard right after DEA would eliminate possible interference with other DEA degradation products. After some trials with dodecyl (Cg) , decanolic (C^Q) and tetradecyl (C14) fatty acids, decanolic acid (CH3-(CH2)g-COOH, "DECA") was chosen. Its trimethylsilylated product (DECA-TMS) has a retention time of 5 minutes. For the present study, a large batch of DECA in DMF solvent was prepared with a concentration of 7.46x10 mole/ml. 11. Choice of Septum The choice of septum depends on how it is to be used. For high injector and column temperatures, ordinary rubber septa tend to bleed and deteriorate rapidly. Furthermore, small broken pieces of septum materials (dirt) tend to accumulate just inside the injector port and may result in high background noise, peak tailing and sample loss. Hence, in the present work where the chromatograph had to be operated at extremely high sensitivities, only high quality septa could be used. After considerable testings, it was found that Septum #2015 developed by the Unimetric Corporation was suitable. This septum is made of high quality silicone rubber reinforced by fiber glass. An inert Teflon layer is also bonded to the side of the septum facing the column. 49 A G.C. needle guide was used in conjunction with the above septum. This allows the needle to be guided into the "cool" needle entrance thus eliminating premature flashing of volatile components. In addition, it serves as a spacer for needle penetration depth and it lengthens septum life by using only a single pierced hole for repeated injections. Needle guides are indispensable for high precision, delicate analytical analyses where the sample from the syringe is to be delivered inside a glass chromatographic column without metal contact. 12. Choice of Operating Conditions Attempts were made to optimize all operating variables in order to achieve the best separation of DEA and its degradation products. It may be noted that temperature pix^gramming equipment (18, 19) was unavailable. Initially, the column temperature (130° C) and nitrogen carrier gas flow rate (19 ml/min.) were adapted from Piekos £t at. (25). After some trials, the column temperature was adjusted to 142° C which was high enough to give good separation with low elution times and the carrier flow rate was increased to 20 ml/min. At this carrier flow rate, the optimum H2 and air flow rates to the flame ionization detector were determined to be 35 and 350 ml/min, respectively. 50 The injector port temperature was set to 195° C which was sufficient to vaporize the sample instantaneously without thermal decomposition. The temperature of the flame ionization detector is not very critical and was chosen to be 220° C to avoid condensation of samples. 13. Equipment Specifications A summary of the final analytical equipment used and its operating conditions are given in Table 10. 14. Sample Preparation Procedure The sample preparation procedure consists of 3 main steps, namely, drying o-f, sample,, sZlylatlon w-Ltk BSA and G.C. analyst. After drying 1 ml of aqueous DEA sample by the method discussed previously, DMF was used as a solvent for the re-dissolution of the DEA residue. 2 ml DECA spiked DMF solution were.added to each DEA -4 . • sample so that each contained 1.49x10 mole of decariolic acid . Pure DMF was then added until the final solution was exactly 25.0 ml. Exactly 4.0 ml of the above solution were transferred to a 12 ml sampling bottle equipped with a Teflon-lined screw cap. A large excess of 2 ml BSA (8.31x10""3 mole) was then added to the sampling bottle and the cap was replaced and screwed tight..Special 51 TABLE 10: Analytical equipment and its operating conditions for G.C. analysis. Gas Chromatograph Manufacturer Model Detector Chromatographic Column Supplier Material Dimensions Packing Varian Aerograph 1440 Series, single column model H2 flame ionization detector. Western Chromatography Supplies, New Westminster B.C. 316 Stainless Steel 1/8" O.D., 6* long 8% OV-17 on 80/100 mesh chromosorb WHP Operating Conditions Injector Port Temperature Column temperature Detector temperature Carrier gas Carrier gas flow • Air flow 195° C 142° C 220° C - N, H2 flow 20 ml/min 350 ml/min 35 ml/min Recorder Supplier Model Response Chart speed Syringe Supplier Model Injected sample size Coming ; 840 Series with mechanical disc integrator 1 mv, 1 sec. full scale Variable Precision Sampling Corporation Pressure-Lok liquid syringe, series D 2 pi volume 0.3 pi 52 oil-resistant, electro-plating adhesive tape developed by the 3M Conpany was then wrapped around the sampling cap to ensure a perfect seal. The bottle was then well shaken and inmersed in a 60° C oil bath for 1 hour which was sufficient to complete the silylation reactions. Silylation at 70° C gave essentially the same results. When the silylation reaction was complete, the sampling bottle was removed from the oil bath and the Teflon lined stopper cap replaced by a silicon rubber septum which penriitted syringe withdrawal of a sample. This procedure was adopted to minimize the contact between the sample and atmospheric moisture. Finally, sample volumes of 0.3 ul were injected into the chromatograph pre-set to the previously described conditions. 15. Calibration Method for Quantitative Analysis The calibration plot for DEA was prepared by determining the ratio of DEA to DECA peak areas as a function of the amount of DEA injected into the chromatograph. The calibration plot for HEP was obtained in.an analogous manner. The plots are shown in Figs. 5 and 6. Whenever possible, the peak areas were measured by the disc integrator included in the Corning recorder. 0 2.0 4.0 6.0 8.0 AMOUNT OF DEA x 10"4 (MOLE) FIGURE 5: Calibration plot of the peak area ratios of DEA/DECA as a function the amount of DEA present. 12 <10 Iii cr < 8 ATTENUATION SETTINGS HEP 8 xlO"12 DECA 32 xlO"11 2.0 4.0 60 8.0 AMOUNT OF HEP x IO"6 (MOLE) 10.0 FIGURE 6: Calibration plot of the peak area ratios of HEP/DECA as a function of the amount of HEP present. 55 16. Maintenance of Analytical Equipment The gas chromatograph and microsyringes must be kept extremely clean at all times. This was most important when operating in high sensitivity ranges since contaminants might produce extra or "ghost" peaks. Sometimes a rapid loss of sample peak size with repeated injection resulted, due to the absorption of sample on dirty surfaces. Routine maintenance and cleaning methods are described in the "Varian Oiromatograph Operation Manual" (47). In using gas chromatography for the analysis of trimethylsilyl derivatives, Si02 deposits tend to accumulate slowly in the flame ionization detector. Bleeding of the silicone column phases also results in silicone oxide formation. The deposition of such material on electrodes and Teflon insulators results in increased noise and decreased detector sensitivity. These deposits can be removed effectively from the detector cap and collector cylinder by immersing them in acetone or chloroform followed by gentle scrubbing. The detector tower, ignitor coil etc. can be similarily cleaned by scrubbing with wetted tissue paper. However, the flame tip assembly is hard to reach and the complete removal of the probes (which are very fragile)(47) is not recraranended unless absolutely necessary. An easier method is to inject 10 to 35 pi of Freon 113 flame ionization detector cleaner directly into the chromatograph with the equipment operating and the G.C. column temperature at 150 C. The Freon elutes from the column within a few seconds and, when burnt in the hydrogen flame, produces HF as the cleaning agent. Usually a few injections are needed to restore the detector to its normal behavior. 57 IV. DEA DEGRADATION UNDER COslTROLLED CONDITIONS As pointed out in the Introduction, very little information has been published on DEA degradation. Plant experience indicates that DEA losses are affected by the following variables: temperature, pressure, raw gas composition, DEA concentration and pH of the DEA solutions. Due to the complexity of the problem, little effort has however been made to correlate these variables with DEA losses. One of the objectives of the present study was therefore to detenriine the degradation of DEA and the formation rates of degradation products in a series of well controlled experiments. Two types of equipment were constructed. The first one was suitable for operating at atmospheric pressures and moderate temperatures, whereas the second equipment was suitable for operating at elevated pressures (up to 5617 kPa (800 psig)) and temperatures (up to 190° C) . 1. Atmospheric Pressure Equipment As discussed in Chapter III in connection with the chromatographic analysis of degraded DEA solutions, metallic surfaces tend to accelerate DEA degradation (24). Hence, the equipment built for atmospheric pressure operation was constructed so that only Pyrex glass or Teflon came in contact with the DEA solution. As seen from Fig. 7, the DEA GROUND GLASS STOPPER REFLUX CONDENSER-PARALLEL SIDE ARM GLASS -CONNECTOR FLASK COVER WITH 4 GROUND GLASS OPENINGS 1000 ml FLASK EQUIPPED WITH SPECIAL HEATING JACKET TO FEED GAS OR PRESSURE RELIEF VALVE iff STOPCOCK I pa. DEA SAMP LING PORT r—. -VARIABLE SPEED MOTOR •TYGON TUBING •GLASS SHAFT TO TEMPERATURE CONTROLLER AND DIGITAL VOLTMETER —COVER CLAMP TEFLON PADDLE IRON CONSTANTAN THERMO COUPLE GLASS THERMOCOUPLE WELL HIGH BOILING POINTOIL TEFLON PADDLE FIGURE 7: Schematic diagram of the atmospheric pressure apparatus for DEA degradation reaction studies. " 59 solution was contained in a IQO 0 ml glass flask. A cover plate containing four ground glass openings and lined with a silicon rubber gasket could be attached to the flask by means by 3 cover clamps, thus providing a gas-tight seal. A commercial all glass stirrer assembly was installed in one of the four cover plate openings. It was designed for gas-tight operation and was powered by a variable speed motor. Two Teflon paddles were attached to the glass stirrer, one for agitating the DEA solution and the other for mixing the gases above the solution. A reflux condenser and sampling valve were attached to the reaction flask by means of a parallel side arm connector. The condenser was used to prevent the,build-up of pressure and return any vaporized water. Using the sampling valve adjacent to the condenser, in conjunction with a two-way flow valve, acid gas could either be introduced into the reaction flask or the flask could be connected to a pressure relief valve. The solution temperature was maintained constant by placing the reaction flask into an electric heating jacket. The power input to the heating jacket was regulated by a Thermal Electric proportional controller. (Model 232), which was capable of keeping the temperature within + 3° C of the desired value. The solution temperature was measured and fed back to the temperature controller by an Iron-Constantan thermocouple immersed in a thermocouple well filled with 60 high boiling point oil. A digital millivoltmeter was used in conjunction with the thermocouple to monitor the solution temperature. Chemically pure CC^, CH4 and H^S gases were metered by three precision rotameters before mixing and introducing them into the reaction vessel. The acid gas composition was determined from previous, carefully prepared calibration plots of the three rotameters. Due to the hazardous nature of E^S (48), the entire apparatus was placed in a fume-cupboard. A typical run consisted of introducing 350 ml of aqueous DEA solution of the desired strength into the reaction flask. Acid gas was passed over the solution for about 90 minutes, which was sufficient to saturate the solution. Once a day, a 3 ml liquid sample was withdrawn from the DEA sampling port before resaturating the solution with an acid gas of the same composition. Runs lastings up to 23 days were conducted at a maximum temperature of 90° C. 2. High Pressure Equipment The primary purpose of this equipment was to check whether it was possible to reproduce the high rates of HEP formation reported by Polderman it at. (16) at elevated temperatures and pressures. The high pressure reaction vessel was a modified 316 stainless steel miniature sample cylinder manufactured by Nupro. It has a maximum service pressure of 6996 kPa (1000 psig) and a capacity of 61 25 ml. The approximate internal dimension of the reactor are 22 mm dia. x 80 mm high. Figure 8 shows the schematic diagram of the high pressure apparatus. All other parts used in conjunction with the reaction vessel were also made from 316 stainless steel. The reaction vessel was submerged in a high temperature bath filled with glycerine and equipped with a mechanical stirrer. The power input to the temperature bath was regulated by a Thermal Electric proportional controller (Model 232) which was capable of keeping the temperature within + 3° C of the desired value. The solution temperature was measured and fed back to the temperature controller by an Iron-Constantan thermocouple immersed in a thermocouple well (filled with high boiling point oil) which was placed inside the temperature bath. A digital millivoltmeter was used in con junction with the thermocouple to monitor the solution temperature. No attempt was made to withdraw samples from the high pressure apparatus during a run. Instead, 15 ml of aqueous DEA solution of the desired strength was allowed to be in contact with the acid gas at the desired pressure and temperatures for a specific period. The reaction vessel was then removed from the oil bath, cooled to room temperature and brought down to atmospheric pressure by means of a purge valve. Only then was the DEA solution withdrawn from the reaction vessel and analyzed. Experiments lasting as long as 11 hours were conducted at temperatures and pressures of up to 185° C and 4238' kPa (600 psig) respectively. C02INLET -FROM HIGH PRESSURE CYLINDER T JOINT PRESSURE GAUGE • REACTION VESSEL (25 ml CAPACITY) BLEED VALVE FIGURE 8: Schematic diagram of the high pressure apparatus for DEA degradation reaction studies. V. DATA ANALYSIS According to Polderman and Steele (16) , DEA degradation by CO2 can be represented by DEA + C02 *HEOD + H20 (14) 2 HEOD * HEP + 2C02 (15) These equations suggest that only two degradation products (HEOD and HEP) should be formed. However, both Hakka zt al. (17) and Smith oX. al. (7) reported the presence of other degradation products in industrial gas-treating solutions. Hence, Eqs. 14 and 15 cannot explain DEA degradation fully. Nevertheless, the equations may be used to form at least a partial basis for a mathematicai model describing DEA degradation. 1. Theory of DEA Decomposition Initial analyses of a series of lean DEA samples obtained from high pressure operations showed that the DEA concentration decreased uniformly with time and could be explained in terms of a first order reaction. Hence, (16) where j^DEA J= Concentration of DEA t = Time V„ = Overall reaction rate constant DEA The integrated form of the above equation is log 10 DEA = log 10 DEA ' - w 2.303 (17) If the reaction is first order, a semi-logarithmic plot of the experimentally determined DEA concentrations versus time should fall on a straight line. Systematic departure from the straight line would indicate a reaction mechanism of order other than unity. It should be pointed out that Eejs. 16 and 17 are based on the assumption that OO2 i-s present in excess, which was the case in the present study. Once the overall reaction order and the reaction rate constant are obtained, the activation energy can be found from the Arrhenius equation (49, 50): (18) or ^IOW = logioA " E (19) 2.303 RT 65 where A = Freguency factor Ea = Activation energy R = Universal-gas constant T = Absolute temperature The valves of A and E& can therefore be found by plotting kj-^ versus 1/T on send-lc>garithmic graph paper. 2. Theory of Degradation Product Formation Since DEA degradation is more complex than suggested by Eqs. 14 and 15, the reaction scheme could be modified as follows: k2 — > h HEP (20) U denotes an unspecified degradation product which may decompose further, but tests have shown that HEP is stable at the high pressure and temperatures used in this study. The rate equations corresponding to the more general reaction scheme (Eq. 20) are: ^pL^na] -^[HBOD]2 (22) (23) with initial conditions (at t = 0): Exponents of HEOD ranging from 0.5 to 4 were considered for Eqs. 22 and 23. An exponent of 2 gave the best agreement between the predicted and measured HEP concentrations. Since Eqs. 22 and 23 are non-linear, they had to be solved numerically and a standard Runge-Kutta method, which was available from the UBC computer library, was chosen. Since kp^ was known from the DEA measurements, only k^ and k^ needed to be specified to determine the theoretical values of HEOD and HEP. The specifications were made using a Simplex search routine which minimize the difference between the measured and theoretical HEP values. A listing of the computer program is given in Appendix D. VT RESULTS AND DISCUSSION 1. Interpretation of Chromatograins After considerable preliminary testing, the important parameters discussed in Chapter III were optimized for the analysis of DEA solutions (as shown in Table 10). Figure 9 shows a chromatogram for a mixtures of various commercially available ethanolamines and degradation products after BSA silylation. Figure 10 shows a typical chromatogram of a lean DEA sample obtained from a high pressure experiment. In this particular case, a 30% DEA solution was contacted with pure C02 at 4238 kPa (600 psig) and 185° C for 4 hours. For comparison, a chromatogram of the parent, 30% DEA sample used for the high pressure experiments is given in Fig. 11. 1.1 MtcUuAejmiint VEA PeakA The DEA peaks (Figs. 9 to 11) are symmetrical and sharp, indicating proper selection of the G.C. column, operating conditions and sample injection technique. The reproducibility of DEA measurements was good (within + 4%) and the lower limit of measuring DEA was found to be about -6 1.44 x 10 moles (for a 30% DEA parent solution this corresponds to -2 1.38 x 10 % DEA) . The detection of DEA peaks was possible at an even lower level. However, at such low concentrations, the peaks ]VTINUTES FIGURE 9: oo MINUTES FIGURE 10- Typical chronatogram of a lean DEA sample obtained with the high pressure equipment. (Sensitivity: 1-512x10-10 ^ps; 2-16x10-10 AFS; 3-32x10-11 AFS; 4-32x10-12 AFS. Conditions as given in Table 10.) ^> o 10 15 JVOMUTES FIGURE 11: Chrcmatogram of parent, 30% DEA solution. (Sensitivity: l-512xl0-10 AFS; 2-32xl0_1° AFS; 3-32xl0-11 AFS; 4-32xlO~"L/ AFS. Conditions as given in Table 10.) o suffer some distortion which is probably due to the interference from the solvent and residual BSA. 1.2 MzaAuA.ejme.nt ofa Vegftadation Product Peaks As shown in Fig. 10, five other peaks labelled "HEP", "T", "X", "Y" and "Z" were observed, when samples obtained with the high pressure equipment were analyzed. HEP and T are degradation compounds previously reported by other researchers (7, 17). The compound Y might be a degradation product or a trace impurity since this compound was also present in the parent chromatogram (Fig. 11). The compounds X and Z are also thought to be caused by degradation products, although the parent DEA chromatogram (Fig. 11) also contains compound X but at a very much smaller concentration. Since all degradation products were present at very low concentrations, the chromatograph had to be set to extremely high -12 . . . sensitivity ranges (32 x 10 AFS or less). At such sensitivity levels, very minor contaminents and high line-voltage fluctuations can induce disturbances and distort the peaks somewhat. The former often resulted in baseline drift and the latter in minor peaks at regular intervals. If a mechanical disc integrator is used to estimate peak areas, the baseline drift can introduce significant errors. This problem was overcome by losing a triangulation method (19) for measuring the peak areas. As shown in Fig. 12, a level baseline was first estimated and drawn on the chromatogram. The width of the peak, W was regarded as the distance between the start and end points of the peak. The peak height, H, was taken to be the perpendicular distance between the baseline and the apex of the peak. The area was calculated from Area = Js;W-H (24) The concentration of HEP in the experimental samples was measured by this method and evaluated from the previously prepared calibration plot (Fig. 6). However, the concentrations of other products could not be determined quantitatively since these compounds have not yet been identified. Hence, only the area ratios of these peaks to the decanolic acid (DECA) peaks are reported at present. Although the triangulation method was time consuming, the precision was found to be good provided that the peaks were large and not too different from Gaussian shapes (19). The reproducibility was also good (within + 7%) and the lower limit for quantitative HEP measurements was found to be approximately 6.25 x 10 mole/ml. 2. High Molecular Weight Degradation Products Apart from HEP, Hakka eX at. (17) also reported that the compound THEED is often present in industrial DEA solutions. This compound is unavailable commercially and the authors reported that it decomposes during G.C. analysis. However, the compound TEHEED, which has a structure very similar to THEED could be purchased. The present silylation procedure made G.C. analysis and detection -5 of TKHKHII possible even at concentrations as low as 2.95 x 10 mole/ml provided the column temperature was raised to about 174° C. The TEHEED peak was Gaussian in shape and gave no indication of sample decomposition. Hence the ability to separate fragile compounds like THEED does exist. o * When the column temperature is raised to 174 C, HEP and TEHEED eluted after 5.6 and 31.3 minutes, respectively. A high temperature chromatogram for a typical lean DEA sample obtained with the high pressure equipment is shown in Fig. 13. Two significant peaks were observed at 10.7 and 13 minutes after injection and were followed by 5 other smaller peaks at 17.2, 23.4, 29.0, 33.0 and 42.4 minutes. These latter seven peaks are caused by unknown compounds and indicate the presence of species of much higher molecular weight than HEP. Consequently they cannot be the other degradation products listed in Table 2. The possibility that one of -the seven compounds * This column tenperature did not cause DEA degradation in the gas chromatograph as confirmed by injecting a sample of the initial DEA solution. MINUTES FIGURE 13: A typical high temperature chromatogram of a lean DEA sample obtained with the high'pressure ecruipment. (Sensitivity: l-512xirr10 AFS; 2-16x10-10 AFS; 3-32x10-11 AFS;~4-64xlO"12 AFS; 5-128xl0"l2 pps; 6-64xl0~l2 AFS; 7-32xl0~l2 AFS. Conditions as given in Table 10 with the column temperature at 174° C.) -J is THEED does exist but further work is required to confirm this as well as to establish the formation rates of such compounds. A second internal standard should be added to the existing system for effective, quantitative determinations. Based on the existence of high molecular weight degradation products, which have not been previously reported, the DEA degradation mechanism must be more complex than postulated by Polderman &t aZ. (16). 3. Analytical Technique The reliability of the newly developed G.C. technique is confirmed by the numerous tests of samples which contain DEA and its degradation products in a wide range of concentrations. This method is fast, has good reproducibility, has high sensitivity and is capable of analyzing fragile compounds. Relatively inexpensive G.C. columns of the CV type could be used for the analyses of high boiling point, fragile compounds with long column life and good separation characteristics. Furthermore, more than 36 samples from Canadian gas-processing plants and oil refineries have been received and analyzed in this laboratory. The results are tabulated in Appendix E for future reference. The capability of handling industrial samples, which contain a wide variety of contaminants (11, 12) demonstrates the versatility and effectiveness of the newly developed G.C. method. 4. Results of Atmospheric Pressure Tests The all glass apparatus discussed in Chapter III, section 1 was used under the following operating conditions listed in Table 11. Table II: Operating conditions for the atmospheric pressure equipment. TmpeAatuAe.: 30 - 90° C VEA Concentration: 10 - 30% Gcu> Composition: H2S CO. 0 - 50% 0 - 50? CH, 0 - 100% It was observed that the DEA solutions changed to a light yellow colour after about 1 day and that the colour darkened with time. After about 7 days of operation, the solution changed to a deep yellow colour. 78 However, in spite of the colour change, none of the degradation products previously reported (7, 17) were detected after conducting the experiments for up to 23 days. These observations indicate that the conversion of DEA into these compounds either did not occur or take place at an extremely slow rate. However, as shown in Fig. 14 the DEA concentration decreased steadily with time. It is quite unlikely that the DEA loss resulted from volatilization since water has a higher vapour pressure than DEA. A more probable explanation is that heat-stable salts were formed which are undetectable by G.C. analysis. As discussed in greater detail in Chapter II, the formation of these heat-stable salts is quite common in industrial plants where acidic gases (COS, HCN), organic acids and pblysulfide compounds are frequently present. In the present case, the formation of these heat-stable salts might be caused by the impurities present in the commercial CO2 gas supply and the large concentration of 02 present. The latter was introduced by leakage as well as during sample withdrawl. 5. Results of High Pressure Tests Preliminary work using the high pressure equipment gave essentially the same results as those obtained with the atmospheric pressure equipment for temperatures up to 140° C. However, by increasing the DEA solution temperature to 175° C, complete conversion 0 10 15 20 TIME (DAYS) FIGURE 14: DEA loss as a function of time obtained with the atmospheric pressure equipment. (Operating conditions: pressure—101 kPa (14.7 psi); temperature—60° C; acid gas—pure C09; initial DEA strength--30%.) 80 of DEA to its degradation products including HEP was observed after 8 hours of contact with CC>2 at 4238 kPa (600 psig) . These results are in good agreement with the findings of Polderman oX at (16) . Furthermore, degradation product peaks were observed in the chromatograms. The peak areas of those degradation products also changed with time. It was concluded that further experiments with the high pressure equipment were desirable especially because the reaction times were short (a few hours as compared to several days for the atmospheric pressure equipments). Table 12 lists the results obtained with the high pressure equipment conducted at 165, 175 and 185° C. In all cases, 15 ml of 30% DEA aqueous solutions were subjected to pure C02 at a pressure of 4238 kPa (600 psig). The analysis of the initial 30% DEA solution is also included for reference. The concentrations of DEA, HEP, X and Z are shown as a function of time in Figs. 15 to 18. The DEA degradation is governed by a first order reaction as demonstrated by the serni-lcigarithmic plot shown in Fig. 15. The solid lines represent the best fit using a least square method with the intercept fixed at 3.12 x 10 mole/ml (the parent, 30% DEA concentration). The overall reaction rate constants were then estimated from the slopes of the lines. Since the high pressure experiments were conducted at three different temperatures, the activation energy and the frequency factor could also be determined. The overall rate constants are summarized in Table 13 and Fig. 19 is Table 12: Results of high pressure tests. Concentrations Temperature, Time, DEA, HEP, T X Y Z ° C hours g-mole/ml g-mole/ml 185 1 8.7xl0"4 6.25xl0~7 0.0 6.84 5.55xl0~2 3.40 » 2 3.2xl0~4 1.19xl0-5 0.0 10.7 1.79xl0~2 ,4.51 3 1.88xl0~4 2.16xl0"5 0.0 8.99 0.0 5.38 4 9.38xl0~5 3.26xl0"5 0.0 6.35 7.44xl0~2 5.59 5 0.0 4.75xl0"5 3.75xl0"2 3.23 2.89x10"2 6.45 6 0.0 6.06xl0~5 6.2xl0"2 2.41 5.43xl0"2 5.76 7 0.0 8.44xl0-5 1.14xl0_1 1.30 5.61xl0~2 5.07 8 0.0 1.16xl0"4 9.0xl0~2 0.24 4.09xl0"2 4.07 175 1 2.0xl0-3 0.0 0.0 6.08 0.0 3.46 2 l.lxlO"3 l.OxlO-6 0.0 8.60 0.0 4.93 3 8.2xl0~4 6.88xl0"6 0.0 10.55 4.74xl0"2 6.64 4 5.4xl0-4 1.05xl0"5 0.0 10.1 4.36xl0~2 6.99 5 2.94xl0"4 2.11xl0"5 0.0 8.23 2.48xl0"2 8.12 » 6 2.1xl0-4 2.81xl0"5 0.0 6.33 4.41xl0"2 6.69 Table 12: Results of high pressure tests (continued). Temperature, Time, Concentrations DEA, HEP, T X Y Z ° c Hours g-mole/ml g-mole/ml 175 7 1.63xl0~4 3.56xl0"5 1.47xl0"2 5.18 4.43xl0"2 6.21 ii 8 9.0xl0-6 5.31xl0~5 8.01xl0-2 4.4 4.77xl0"2 5.26 165 3 1.2xl0"3 0.0 0.0 8.88 4.64xl0"2 6.46 4 1.03xl0"3 1.56xl0"6 0.0 9.85 0.0 7.93 5 6.0xl0~4 4.06xl0~6 0.0 11.65 0.0 8.78 6 4.0xl0~4 6.56xl0"6 0.0 9.25 0.0 9.28 -7 2.8xl0~4 l.llxlO-5 0.0 6.97 7.07xl0"2 8.51 8 2.19xl0~4 1.53xl0~5 0.0 5.46 0.0 8.76 9 1.44xl0"4 2.11xl0~5 0.0 4.88 3.51xl0"2 8.0 10 1.06xl0"4 3.17xl0~5 0.0 4.17 2.77xl0"2 7.12 11 8.75xl0~5 4.31xl0~5 0.0 3.27 4.2xl0"2 6.67 * 0 3.12xl0~3 0.0 6.93xl0"3 0.566 3.41x10"2 0.0 ™ . , „ „ ,T . „ , peak area of compound . Note. The concentrations of T, X, Y and Z are expressed as •*-—T. ——. _ . The sensitivity i T .c ^i. i ^ ^- c\i. j J ^T-^.A peak area of DEGA level for the detection of the compounds and DECA were set at 32xl0"12 AFS and 32xl0_11 AFS respectively. * Initial 30% DEA solution FIGURE 15: DEA concentration as a .function of time obtained with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure Ot^; initial DEA strength—30 wt.%; the symbols are based on experimental measurements whereas the lines represent the model predictions.) LU _J O m b 18.0 16.0 14.0 12.0 10.0 z o fe cr r-z LU o z o o CL Ld X LEGEND TEMPERATURE ,°C SYMBOL — I65 O I75 © 185 A _ 4 6 8 TIME (HOURS) 12 FIGURE 16: HEP concentration as a function of time obtained with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure C02; initial DEA strength—30 wt.%; the symbols are based on experimental measurements whereas the lines represent the model predictions.) TIME (HOURS) FIGURE 17: Formation of compound X as a function of time obtained with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig); acid gas—pure C02; initial DEA strength—30 wt.%; the symbols are based on experimental measurements and the lines were drawn to pass through these data points.) 16 T 1 1 1 1 1 1 1 1 1 r FIGURE 18: Formation of compound Z as a function of time obtained with the high pressure equipment. (Operating conditions: pressure—4238 kPa (600 psig) ; acid gas—pure CC^'" initial DEA strength—30 wt.%; the symbols are based on experimental measurements and the lines were drawn to pass through these data points.) FIGURE 19: Overall DEA reaction rate constants as a function of absolute temperature as listed in Table 13. 88 the corresponding Arrhenius plot. The activation energy and 4 -1 frequency factor were found to be 2.17 x 10 cal g-nole and 2.03 x 10"^ hour \ respectively. Table 13: Overall reaction rate constants obtained-with the high pressure equipment. (Operating conditions: pressure — 4238 kPa (600 psig); initial DEA. strength — 30 wt.%) acid gas — pure CO2; Temperature Overall rate constant, k^EA °c °K hour 165 438 0.331 175 448 0.450 185 458 0.987 89 The HEP concentration is a function of time and temperature as shown in Fig. 16. The solid lines represent the best fitted lines to the experimental HEP data obtained from the kinetic model discussed in Chapter V. Polderman oX at. (16) reported that the formation of HEP from DEA might not be a simple process and that an intermediate compound, HEOD, may be present. If this hypothesis is true, the concentration of HEOD should pass through a maximum. . Figures 17 and 18 show the concentration of X and Z respectively as a function of time. The concentrations of these two, as yet unidentified, compounds exhibit a maximum. It is therefore probable that both of them are reaction intermediates and one of them may be HEOD. Since no commercial HEOD standard was available, the identification of HEOD in the degraded samples was not possible. The concentrations of the other two unidentified compounds (T and Y) were not plotted but are only tabulated in Table 12 for future reference. This was done because the compounds gave very small peaks even at extreme detector sensitivity settings and they may therefore be regarded as very minor constitutents or DEA impurities. For example, the chromatogram of the parent DEA solution (Fig. 11) also contained the compounds T, X and Y. By comparing the concentration values of T and Y with other high pressure samples, it was found that they did not change substantially. Hence, the variation in these two compounds concentrations could be caused by measurement error 90 since the peak size was extremely small and the peaks were situated between the large peaks of compounds X and Z. .From the above results, it is clear that temperature plays an important role in DEA degradation. 6. Results of the Kinetic Model Study Some exploratory studies were undertaken to test the DEA degradation mechanism proposed in Chapter V. The reaction rate constants, which are sunmarized in Table 14, do not obey the usual Arrhenius equation. This implies that the kinetic model may be inadequate. Table 14: Reaction rate constants giving the best fit between the theoretical and experimental HEP concentrations. Temperature Rate Constants °C °K , hour V hour k2,ml g-mole "Siour 165 438 0. .329 0. 002 1.09 175 448 0. ,327 0. 123 4.02 185 458 0. ,324 0. 663 3.49 Table 15: Experimental and theoretical concentrations of DEA, HEP and HEOD. Temperature, Time, hours DEA-expr., g-mole/ml Concentration DEA ther;, g-mole/ml HEP -expr.j g-mole/ml HEP ,u-ther., g-mole/ml HEOD ... . - -ther., g-mole/ml 185 175 0 1 2 3 4 5 6 7 8 0 1 3.12x10 ,-4 -3 8.7x10 3.2x10 -4 1.88x10 9.38x10 0.0 0.0 0.0 0.0 3.12x10 2.00x10 -4 -5 -3 -3 3.12x10 -3 1.17x10 4.35x10 -4 1.62x10 6.05x10 -5 2.26x10 -5 8.42x10 -6 3.14x10 -6 1.17x10 -6 3.12x10 -3 2.00x10 -3 0.0 6.25x10 1.19x10 2.16x10 3.26x10 4.75x10 6.06x10 8.44x10 1.16x10 0.0 0.0 -7 -5 -5 -5 -5 -5 -5 -4 0.0 3.01x10 1.34x10 -5 2.74x10 -5 4.23x10 -5 5.69x10 -5 7.09x10 -5 8.41x10 -5 9.66x10 -5 0.0 5.09x10 -7 0.0 6.38x10 8.57x10 9.19x10 9.22x10 9.06x10 8.82x10 8.58x10 8.33x10 0.0 8.24x10 -4 -4 -4 -4 -4 -4 -4 -4 -4 Table 15: Experimental and theoretical concentrations of DEA, HEP and HEOD (continued). Temperature, Time, Concentration ° c hours DEA . expr., g-mole/ml DEA . ther., g-mole/ml HEP ; expr., g-mole/ml HEP ther., g-mole/ml HEOD ther., g-mole/ml 175 2 l.lOxlO"3 1.27xl0~3 l.OOxlO"6 3.01xl0"6 1.35xl0"3 II 3 8.20xl0"4 8.12xl0~4 6.88xl0"6 7.69xl0"6 1.67xl0"3 it 4 5.40xl0"4 5.18xl0"4 1.05xl0"5 1.41xl0~5 1.87xl0"3 II 5 2.94xl0"4 3.31xl0"4 2.11xl0~5 2.17xl0-5 2.00xl0~3 ti 6 2.10xl0~4 2.11xl0"4 2.81xl0~5 3.00xl0~5 2.07xl0"3 II 7 1.63xl0"4 1.35xl0-4 3.56xl0~5 3.87xl0"5 2.10xl0~3 II 8 9.00xl0~5 8.58xl0"5 5.31xl0~5 4.77xl0"5 2.12xl0-3 165 0 3.12xl0~3 3.12xl0~3 0.0 0.0 0.0 II 1 — 2.25xl0"3 — 1.51xl0~7 8.76xl0~4 II 2 — 1.61xl0"3 — 9.63xl0"7 1.50xl0"3 II 3 1.20xl0~3 1.16xl0"3 0.0 2.62xl0"6 1.95xl0-3 ii 4 1.03xl0~3 8.32xl0"4 1.56xl0-6 5.08xl0"6 2.27xl0"3 VO Table 15: Experimental and theoretical concentrations of DEA, HEP and HEOD (continued). Temperature Time, Concentration hours DEA , , expr., g-mole/ml DEA ther., g-mole/ml HEP expr., g-mole/ml HEP 7 ther., g-mole/ml HEOD ther., g-mole/ml 165 5 6 7 8 9 10 11 6.00x10 -4 4.00x10 -4 2.80x10 2.19x10 -4 1.44x10 1.06x10 -4 8.75x10 -5 5.98x10 -4 4.29x10 -4 3.08x10 -4 2.21x10 1.59x10 -4 -4 1.14x10 8.19x10" -4 4.06x10 6.56x10 1.11x10 -6 1.53x10 -5 2.11x10 -5 3.17x10 4.31x10 -5 8.19x10 -6 1.18x10 -5 1.59x10 -5 2.02x10 -5 2.47x10 2.93x10 -5 3.40x10 -5 2.50x10 -3 2.66x10 -3 2.77x10 -3 2.85x10 -3 2.90x10 -3 2.94x10 -3 2.96x10 -3 94 The experimental and theoretical concentrations of DEA and HEP are shown in Figs. 15 and 16. The theoretical HEOD concentrations predicted by the proposed kinetic model are shown in Fig. 20. The experimental and theoretical results for DEA and HEP (Figs. 15 and 16) were found to be in moderately good agreement. The discrepancies are largely due to experimental error, with inadequate temperature control being the most prominent one. The temperature controller was only capable of maintaining the oil bath temperature within + 3° C which is unsatisfactory for high reaction rates. Figure 20 shows that the HEOD concentration should rise rapidly before gradually decreasing again. This is in marked contrast to the behavior of compounds X and Z which decompose rapidly. Hence, the likelihood that either compound X or Z is HEOD is small. However, since this part of the study was exploratory, no efforts were made to improve the accuracy of the analysis of compounds X and Z. This could be done by using a larger sample size for G.C. injection, which would lower the detection sensitivity and give better base-line control. Alternatively, another internal standard with a retention time slightly greater than HEP could be used. 7. DEA Degradation Under Special Conditions In addition to the high pressure tests summarized in Table 12, a number of special tests were performed. The main objective was to find the effects of C0o pressure, DEA solution strength, the presence 40.0i 35.0 UJ 30.0 o l C9 b o cr r-z LU O z o (J o LU X 25.0 20.0 15.0 10.0 LEGEND TEMPERATURE,0 C SYMBOL 165 O 175 © 185 0 4 6 8 TIME (HOURS) 10 12 FIGURE 20: Formation of HEOD as a function of time predicted by the kinetic model. 96 of H^S and metal surfaces on DEA degradation. Table 16 lists the experimental conditions under which the special tests were performed and the corresponding results. In addition to the analysis for the initial 30% DEA solution, results obtained by contacting this DEA solution with 4238 kPa (600 psig) CC>2 for 4 hours at 175° C is also included for reference. The latter is subsequently called the "base run". The first special experiment involved contacting a 30% DEA solution with He at 1274 kPa (170 psig) and 185° C for 8 hours. The reason for the selection of this relatively low contacting pressure was due to the lack of suitable high pressure regulator. However, this He pressure was sufficient to prevent boiling of the DEA sample at 185° C and, since helium is an inert gas, the He pressure should have no effect on the experimental results. By comparing the results obtained from the He run and the "base run", it was found that CG^ is a major factor affecting DEA degradation since, in an inert _3 atmosphere, the resulting DEA concentration (3.02x10 mole/ml) is -3 very .close to that of the initial 30% DEA solution (3.12x10 mole/ml). However, even in an inert atmosphere, the concentration of degradation compounds T and X increased slightly. FHjrthermore, a trace of compound Z was also detected. The latter compound was not present in the initial DEA solution and this indicates that DEA degradation does occur to a small extent at high temperatures and in the presence Table 16: Special experimental conditions and the results obtained with the high pressure equipment. Experimental Conditions Concentration Initial DEA Pressure Temperature, Gas Reaction Special DEA, HEP, T X Y Z strength 0 C time feature g-mole/ml g-mole/ml 30% 1274 kPa (170 psig) 185 Pure He 8 3.02xl0"3 0.0 8. 09xl0~3 0.766 0.0122 0. 0188 30% 2170 (300 <kPa psig) 175 Pure co2 4 5.4xl0~4 1.05xl0~5 0. 0 10.81 0.0 3. 90 10% 4238 (600 kPa psig) 175 Pure co2 4 3.06xl0~4 0.0 0. 021 0.705 0.0398 2. 57 30% 4238 (600 kPa psig) 175 H2S + 4 co2 Presaturating with 1480 kPa (200 psig) H2S followed by contact .with" 4238 kPa (600 psig) C02 2.22xl0~3 0.0 0. 0 4.67 0.0316 0. 359 30% 4238 (600 kPa psig) 175 Pure co2 4 DEA solution contained in a glass vial 6.06xl0~4 1.03xl0~5 0. 0 6.58 0.0 6. 59 30% — — — — initial DEA 3.12xl0~3 0.0 6. 93xl0~3 0.566 0.034 0. 00 30% 4238 (600 kPa psig) 175 Pure co2 4 "Base run" 5.4xl0~4 1.05xl0~5 0. 0 10.1 0.0436 6. 99 »T *. m, _ c _ v „ , „ , peak area of compound _ .. .. Note. The concentration of T, X, Y and Z are expressed as -—: _ nTrr,^ . The sensitivity level for the detection of these compounds and DECA were set at 32xl0-12 AFS and 32xl0-''"^ AFS respectively. 98 of metal surfaces. The results obtained from the 2170 kPa (300 psig) CC>2 run are quite similar to those of the "base run" except for the concentration of compound Z. The C02 pressure therefore has little effect on the degradation of DEA and the formation of HEP. This is probably due to the fact that C02 is present in great excess in all runs performed. The results from the 10% DEA run showed that the concentration of degradation products were reduced. This is expected from the reaction mechanism postulated by Polderman oX at. (16). However, the measured DEA concentration is much greater than the predicted value of -4 1.84 x 10 mole/ml. The latter concentration was calculated on the basis of 10% DEA in the initial solution and an overall reaction rate constant, kp.^, of 0.45 hr The difference between the measured -4 -4 and predicted DEA concentration after 4 hours is 3.06x10 - 1.84x10 -4 or 1.22x10 mole/ml. The corresponding difference in the overall reaction rate constant is 0.128 hr Such deviations cannot be explained in terms of experimental errors. For example, a drop of 5° C is necessary to explain the decrease in the overall reaction rate constant and this is much greater than the precision of the temperature controller (+ 3° C). The reasons for the unexpected behavior of the 10% DEA solution are not known and merit further investigation in future. 99 Ths results frcm the H2S - CC^ run showed a higher DEA -3 concentration (2.22 x 10 g-mole/ml), absence of HEP and much lower concentrations of other degradation products than the "base run". This can be explained by assuming that most of the DEA is tied up by H2S and hence protected against C02 attack. Calculations based on the absorption of H2S in aqueous DEA solutions (5) confirmed such an interpretion. At 21° C (70° F) and an H2S pressure greater than 133 kPa (1000 mm Hg), 1.08 moles of H2S are associated with 1 mole of DEA at equilibrium. This means that pre-saturating a 30% DEA solution with H,>S before contacting it with C02, leaves only 46% _3 of the original DEA (1.43 x 10 mole/ml) free to react with (X>2. Based on the overall reaction rate constant at 175° C and an initial -3 DEA solution strength of 1.43 x 10 mole/ml, a calculated value of -4 2.48 x 10 mole/ml is obtained for the residue DEA after C02 contact. -4 This is in good agreement with the measured concentration of 2.22 x 10 mole/ml. -4 The slightly higher residue DEA concentration (6.06 x 10 g-mole/ml) obtained by reacting DEA in a glass vial with C02 demonstrates that metal surfaces accelerate DEA degradation somewhat. This finding is in agreement with the results published by S.G.E. (24). The rest of the degradation products were present in concentrations similar to those of the "base run". This shows that, even in the presence of glass surfaces, C09 causes DEA degradation at high temperatures. 100 8. Analysis of Industrial Samples Thirty six samples from Canadian gas-processing plants and oil refineries were analyzed. The results are tabulated in Appendix E. There are strong similarities between the composition of industrial samples and DEA solutions degraded under laboratory conditions. It may therefore be concluded that DEA undergoes essentially the same kind of degradation in both cases. The concentration ranges of DEA arid HEP were in very good agreement. This demonstrates that, when DEA samples of similar concentrations are in contact with acid gases, one of the end products is HEP. Industrial samples generally contain much greater amounts of T (up to 100 times). A probable reason is that the industrial DEA feed, being less pure than the laboratory reagent, contains triethanolamine (TEA) as well as other impurities. The TEA, which has the same retention time as compound T in G.C. analysis, would increase the T peak. Small deviations in the peaks of compounds X, Y and Z might be caused by the presence of other contaminants that were always present in industrial processes. VII CONCLUSIONS A reliable analytical method has been developed for the analysis of DEA gas-treating solutions. This method is based on the gas chromatographic separation of silylated derivatives. It can detect trace quantities of DEA, HEP and other degradation products in samples from laboratory experiments as well as industrial gas plants and petroleum refineries. Degradation experiments at atmospheric pressure showed that DEA solutions change colour after 1 day of contact with CO2 and ^S at temperatures between 30° C and 90° C. However, no known degradation products were detected even after 23 days of operation. However, the DEA concentration decreased with time and the formation of heat-stable salts, which was undetectable by G.C. analysis was suspected. Degradation tests at pressures and temperatures greater than 300 psig and 165° C, respectively, resulted in the formation of HEP, T, X, Y and Z plus seven other unidentified compounds. The first five degradation products could be detected by using a G.C. column temperature of 142° C whereas the latter seven compounds could only be readily observed by"using column temperature of 174° C. Compounds X or Z are suspected to be intermediates of DEA degradation since their concentration versus time plots exhibit maxima. Based on the results obtained from high temperature (174° C) G.C. analysis, degradation products with molecular weights similar to that of TEHEED (m.w. 236) are suspected. If this is true, then this is the first report of the separation of high molecular weight DEA degradation products by gas chromatography. The loss of DEA at elevated temperatures (data from high pressure experiments) was found to be governed by a first order reaction. The rate constants and activation energy of this reaction were also determined. A simple DEA degradation scheme has been proposed and is compared with DEA and HEP measurements. The reaction rate constants for the formation of HEOD and HEP have been obtained by fitting the model to the experimental data. The theoretical DEA and HEP concentrations were found to be quite similar to the experimental values. However, the theoretical HEOD concentrations was very different from experimental measurements of the compounds X and Z. The former concentration was found to increase rapidly and then level off instead of passing through a 103 maximum as observed for compounds X and Z. 9. DEA degradation is complex and leads to the foimations of many degradation products. The reaction mechanism proposed by Polderman nt aL. (16) is too simple to explain the experimental results fully. 10. Chromatograms obtained from samples of high pressure laboratory tests were very similar to those of industrial samples. The concentrations of DEA and HEP as well as other degradation products were present in similar concentrations in the laboratory and industrial samples. 11. Based on the present study, C02 and temperature appear to be the crucial factors in DEA degradation. 12. The' results obtained from high pressure tests conducted with a 10% DEA solution indicated lower degradation rates than expected. The reasons for this are not understood at this time. 13. H2S was found to inhibit DEA degradation. Saturating a 30% DEA solution with H2S at 1480 kPa (200 psig) prior to C02 contact, reduced the DEA loss by possibly limiting the amount of free DEA (up to 54 mole %) available to CC>2 attack. Stainless steel surfaces were found to accelerate DEA degradation somewhat. 105 VIII RECOMMENDATION FOR FURTHER RESEARCH 1. The direct G.C. technique proposed by Saha at. at. (26) should be investigated for the analysis of DEA and its degradation products. This method has proved to be capable of detecting ethanolamines in aqueous solution. However, it is uncertain whether it can also separate fragile compounds like HEP, TEHEED et. at the recommended high temperatures for injector and G.C. column (350° C and 135° C — 350° C temperature programming, respectively) recommended by Saha oX at. (26). Thermal decompositions of the above compounds is very likely. 2. A reliable analytical technique should be developed to analyze the amounts of DEA tied up in the form of heat-stable salts. These compounds cannot be detected by G.C. analysis and must be determined by wet chemistry methods (5, 11, 12, 29) which are time consuming. Perhaps physical properties such as the refractive index of DEA can be used to establish the DEA concentration. 3. The identification of all other compounds detected by the present chromatographic technique is also desirable. Normally, a separation method, such as paper chromatography or TLC has to be used to isolate sufficient quantities of material for other instrumental methods such as M.S. and I.R. High purity standards of HEOD and THEED should be 106 prepared or obtained from commercial suppliers which specialize in synthesising rare chemicals. 4. Atmospheric pressures tests should be resumed and carefully controlled experiments should be carried out to elucidate the DEA degradation process. It would be desirable to find out whether DEA undergoes the same degradation process with acid gases at low temperatures (but at a very much slower rate) or an entirely different DEA degradation process is taking place. The present study of DEA degradation using the all glass equipment is inconclusive but the inability to detect degradation products such as HEP, X, Z etc. might mean that the DEA degradation reaction at low temperature is entirely different than the high temperature mechanism. 5. More high pressure tests should also be performed using better temperature control (to + 0.5° C). Tests should also be performed using initial DEA concentrations ranging from about 5 to 40%. These experiments would be helpful in explaining the present results obtained with an initial DEA concentration of 10%. Isolation and identification of other degraded products present in the high pressure samples should be made so that the DEA. degradation mechanism at high temperatures can be elucidated. The effect of pH, other acidic gases (CS2 and COS) and metal surfaces on DEA degradation should be studied. Chemical additives, which may be used as degradation inhibitors, should also be examined. The effectiveness of activated carbon for cleaning degraded DEA solutions should be investigated. 108 IX NOMENCLATURE Symbol Explanation and typical units A Frequency factor in the Arrhenius Eq. 18, hour-1. AEEA AminoethylethanolaTtiine AFS Amps full scale BSA N ,0-Bis- (trimethysilyl) acetamide DEA Diethanolamine DECA Decanolic acid DMF Dimethyl formamide DMSO Dimethyl sulfoxide Ea Activation energy in the Arrhenius Eq. 18, Cal g-mole-1 G.C. Gas chromatography H Peak height used for the triangulation Eq. 19, mm. HEI Hydroxyethyl imidazolidone HEOD 3-(2-Hydroxyethyl)-2-oxazolidinone HEP N,N-Bis-(2-hydroxyethy1)piperazine 109 Symbol Explanation and typical units IR Infrared. >EA Overall reaction rate constant for Eq. 16, hour~l Reaction rate constant for Eq. 20 & 21, hour -1 Reaction rate constant for Eq. 21 & 22, hour -1 ^2 M.S. Reaction rate constant for Eq. 23, cm3 g-mole "Siour ^ Mass spectroscopy MDEA N-Methyldiethandamine MEA Monoethanolamine 20 Refractive index.at 20 C as measured by the Sodium D line The change in refractive index value with respect to a change in temperature, ° C~l NMR Nuclear magnetic resonance OV Trade name (Ohio Valley) OZD axazolidone R Universal gas constant, 1.99 Cal g-mole ^ °K ^ R, A term used frequently in paper chromatography Symbol Explanation and typical units which is defined as the distance of sample spot from the origin divided by the distance travelled by the solvent front from the same origin. SCF Standard cubic foot SCFD Standard cubic foot per day Sp. gr. ASp. gr./At Specific gravity The change in specific gravity value with respect to a change in temperature, 0C~1. S.T.P. Standard temperature and pressure, 0° C and 760 mm Hg. Absolute Temperature, °K Time, hour TEA Triethanolamine TKHKKI) N,N,N,N -Tetrakis-(hydroxyethyl)ethylenediamine TFA Trifluoroacetyl THEED N,N,N-Tris (2-hydroxyethyl) ethylenediamine TLC Thin layer chromatography TMS DV Trimethylsilyl Ultraviolet Explanation and typical units Peak width used for the triangulation Eq. 19, mm Denotes concentration term, g-mole/ml Denotes concentration term for the initial solution, g-mole/ml values obtained from experiments values obtained from computer data fittings BIBLIOGRAPHY Smith, R.F. and Younger, A.H. "Tips on DEA Treating", Hydroc. Proc, 51, 7, 1972. Kohl, A. and Riesenfeld, F., "Gas Purification", McGraw Hill, N.Y. 1960. Jones, V. Wayne and Perry, Charles, R., "Fundamentals of Gas Treating", Proc. Gas Cond. Conf., 23, E, 1973. Swain, C. Donald, "Gas Sweetening Processes of the 1960s", Hydroc. Proc., 49, 3, 1970. Dow Chemical Company, "Gas Conditioning Fact Book", The Dow Chemical Company, Midland, Michigan, 1962. Beck, J.E., "Diei±ianolamine—An Energy Conserver," Proc. Gas Cond. Conf., 25, A, 1975. Smith, R.F. and Younger, A.H., "Operating Experiences of Canadian Diethanolamine Plants", Proc. Gas Cond. Conf. 22, E, 1972. Gary, James, H. and Handwerk, Glenn, E., "Petroleum Refining— Technology and Economics", Chemical Processing and Engineering, volume 5, Marcel Dekker, Inc., N.Y. 1975. Heisler, Leopold and Weiss, Ing, Helmut, "Operating Experience at Aderklaa with Alkanolamine Gas Treating Plants for Sour Natural Gas Sweetening", Proc. Gas Cond. Conf., 25, H, 1975. Nonhebel, G., "Gas Purification Processes for Air Pollution Control", 2nd Edition, Newness - Butterworths, London, 1972. Henry, M.S. and Grennert, M., "How to Analyze DEA Solutions", Petroleum Refiner, 34, 6, 1955. Henry, M.S. and Grennert, M., "Detection and Measurement of Components in Diethanolamine used for Hydrogen Sulfide Removal from Petroleum Refinery Gases", Proc. California Natural Gasoline Assoc., 28th Annual Fall Meeting, 1955. Coldrey, P.W. and Harris, I.J., "Kinetics Of the Liquid Phase Reaction Between Carbon Dioxide and Diethanolamine", Can. J. Chem. Eng. 54, 1976. Katz, Donald, L., Cornell, David, Kobayashi, Riki, Poettmann, Fred, H., Vary, John, A., Elenbaas, Jack, R. and Weinang, Charles, F., "Handbook of Natural Gas Engineering", McGraw-Hill Book Company, Inc., N.Y. 1959. Campbell, John, M., "Gas Conditioning and Processing", 4th Edition, Campbell Petroleum Series, 1976. Polderman, L.D. and Steele, A.B., "Why Diethanolamine Breaks Down in Gas Treating Service", Oil & Gas J., 54, 65, 1956. Hakka, L.E., Singh, K.P., Bata, G.L., Testart, A.C, Andrejchyshyn, W.M., "Some Aspects of Diethanolamine Degradation in Gas Sweetening", Canadian Natural Gas Processing Meeting, May 1968. Zweig, Gunter and Sherma, Joseph, "CRC Handbook of Chromatography, Volume I and II", CRC Press, Cleveland, 1972. McNair, H.M. and Bonelli, E.J., "Basic Gas Chromatography", Varian Aerograph, Calif. 1968. Budzikiewicz, H., Djerassi, C. and Williams, D.H., "Mass Spectrometry of Organic Compounds!', Holden Day Inc., California, 1967. Brooks, C.J.W. and Middleditch, B.S., "Gas Chromatography— Mass Spectrometry", Mass Spectrom., 4_, 1977. 114 22. Pataki, Gyorgy, "Techniques of Thin Layer Chromatography in Amino Acid and Peptide Chemistry", Ann Arbor Science Publishers, Inc., Michigan, 1968. 23. Brydia, L.E. and Persinger, H.E., "Quantitative Gas Chromatographic Determination of Ethanolamines as Trifluoroacetyl Derivatives", Anal. Chem., 39, 11, 1967. 24. S.G.E., "Analysis of Ethanolamine", S.G.E. Technical Publication on Glass Lined Tubing, 1978. 25. Piekos, Ryszard, Kobyiczyk, Krzysztof and Grzybowski, Janusz, "Quantitative Gas Chromatographic Determination of Ethanolamines as Trimethylsilyl Derivatives", Anal. Chem., 47, 7, 1975. 26. Sana, N.C., Jain, S.K. and Dua, R.K., "A Rapid and Powerful Method for the Direct Gas Chromatographic Analysis of ALkanolamines; Application to Ethanolamines" (^hromatographia, 10, 7, 1977. 27. Zimmerman, Howard, K. and Cosmatos, Alexandros, "Migration of Aminoalcohols on Paper", J. Chromatog., 1_, 1962. 28. Sargent, Roger and Rieman, William, "Salting-Out Chromatography, II Amines", Anal. Chim. Acta, 17, 1957. 29. Union Carbide Chemical Company, "Procedure for Determining the Amount of Monoethanolamine or Diethanolamine Held in" Heat-Stable Salt Form in Acid Gas Absorbing Solutions", Union Carbide Chemical Company, Publication SPD-27. 30. Bonner, William, A. and Castro, Albert, J., "Essentials of Modern Organic Chemistry", Van Nostrand Reinhold Company, N.Y. 1965. 31. Di CorciavxAntonicr': and Samperi, Roberto, "Gas Chromatographic Determination at the Parts-Per-Million Level of Aliphatic Amines in Aqueous Solution", Anal. Chem., 46_, 8, 1974. Gehrke, Charles, W. and Leimer, Kenneth, "Trimethylsilylation of Amino Acids—Derivatization and Chromatography", J. Chromatog., 57, 1971. Gehrke, Charles, W. Kuo, Kenneth and Zumwalt Robert, W., "The Complete Gas-Liquid Chromatographic Separation of The Twenty Protein Amino Acids", J. Chromatog., 57, 1971. Smith, R. Lindsay and Waddington, D.J., "Gas Chromatographic Separation of Aliphatic Diamines", J. Chromatogr., 42, 1969. Preston, Seaton, T., "A Guide to the Analysis of Amines by Gas Chromatography", 2nd Edition, Polyscience Corporation, Illinois, 1973. Lynes, A., "The Separation and Identification of Some ALkanolamines and Their. Salts by Thin Layer Chromatography", J. Chromatog., 23, 1966. Nakanishi, K., "Infrared Spectroscopy—Practical", Holden-Day, San Francisco, 1962. .Fischer, Robert, B. and Peters, Dennis, G., "Basic Theory and Practice of Quantitative Chemical Analysis", Third Edition, W.B. Saunders Company, Philadelphia, 1968. Bauman, R.P., "Absorption Spectroscopy", John Wiley & Sons, Inc., N.Y. 1962. Parris, N.A., "Instrumental Liquid Chromatography", J. Chrom. Library 5, Elsevier Scientfic Publishing Company, 1978. Sadtler Research Laboratories Inc., "The Sadtler Standard Infrared Spectra", Sadtler Research Laboratories Inc., Philadelphia, 1968. 116 42. Brittain, George, D., Sullivan, James, E. and Schewe, Larry, R., "Silylation in the Presence of Water—The Development of a Commercial Reagent for Silylating Aqueous Solutions of Hydroxy and Polyhydroxy Compounds", "Recent Advances in Gas Chromatography" by Domsky, Irving, I. and Perry, John, A., Marcel Dekker, Inc. 1971. 43. Pierce, Alan, E., "Silylation of Organic Compounds", Pierce Chemical Company, Rockford, Illinois, 1968. 44. Pierce Chemical Company, "Handbook of Silylation", Pierce Chemical Company, Rockford, Illinois, 1972. 45. Klebe, J.F., Finkbeiner, H. and White, D.M., "Silylations with Bis(trimethylsilyl)acetamide, a High Reactive Silyl Donor", J.A.C.S., 88, 1966. 46. Weiss, Alvin, H. and Tambawala, Huseini, "TMS Derivatization in Aqueous Solutions", J. Chromatog. Sc., 10_, 1972. 47. Varian Aerograph, "1440 Series Basic Instrument Instruction Manual", Varian Aerograph, Publication No. 85-001122-01, California, 1971. 48. Archibald, Ralph, G., "Process Sour Gas Safely", Hydroc. Proc., 56, 3, 1977. 49. Daniels, Farrington and Alberty, Robert, A., "Physical Chemistry", 3rd Edition, John Wiley & Sons, Inc., N.Y. 1967. 50. Smith, J.M., "Chemical Engineering Kinetics", McGraw-Hill Book Company, Inc., N.Y. 1956. 51. Wijk, R. Van., "The Use of Poly-Para-2, 6-Diphenyl-Phenylene Oxide as a Porous Polymer in Gas Chromatography", J. Chromatog. Sc., 8, 1970. Rushworth, F.A. and Tunstall, D.P., "Nuclear Magnetic Resonance", Gordon and Breach Science Publishers, N.Y. Jackman, Lloyd, Miles and Stemhell, S., "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry", 2nd edition, Pergamon Press, N.Y. 1969. Sadtler Research Laboratories Inc., "TheSadtler Standard Nuclear Magnetic Resonance Spectra", Sadtler Research Laboratories Inc., Philadelphia, 1968. 118 APPENDIX A GENERAL INFORMATION ON SIX G.C. ACTIVE PHASES RELEVANT TO DEA ANALYSIS Once a choice has been made on whether a direct or derivative G.C. technique should be used, the selection of suitable G.C. column is the most critical factor to be considered. Usually packed columns contain an inert support, e.g. Chromosorb, with a thin coating of an active phase. Uniform particle sizes of the support (typically 40-120 mesh) give the highest efficiency and the active phase should have a low viscosity and good selectivity for the mixture components to be separated. A loading of 2 to 10% of the active phase is generally used. The lower coating content can enable a column to give faster separations at a lower temperature but also restrict the injected sample capacity' The maximum column temperature is limited by the onset of "bleeding" (i.e. volatilization of the active phase) or reaction with the column support. The characteristics of six active phases, which are relevant for DEA analysis, are summarized in Table A-l (18, 19, 31 51). Table A-l: General Information on active phases relevant to G.C. analysis of DEA solutions. Name Chemical Structure Type Maximum column temper ature limit Use Recommended solvent Suppliers CH, 0 II Oi-ll Neopentylglycol -)- 0-CH2-C-CH2-0-C-CH2-CH2-C+0-CH„ OV-1 OV-17 CH„ I 3 •Si—(H-CH0 f3 CHo-Si-0-3 I CH3 CH. I 3 Si-O' lb CH„ I 3 Si-CH„ I -CH„ Mildly polar Nonpolar dimethyl silicone gum Moderately polar 50% phenyl substituted phenyl methyl silicone 240° C 325-375 C 350-375 C Separation of silylated derivatives which cannot be separated by non-polar phases. General G.C. separations and in biomedical research., General G.C. separations particularly in biochemical research^ methylene chloride, chloroform toluene, chloroform toulene, chloroform Chromatographic Specialties Ltd. Chromatographic Specialities Ltd. Chromatographic Specialties Ltd. vo Table A-l: General information on activie phases relevant to G.C. analysis of DEA solutions (continued). Name Chemical Structure Type Maximum Use Recommended Suppliers column solvent temper ature limit OV-11 CH„ I 3 CH3-Si-0-|— Si-O-CH, CH, CHQ I 3 CH„ I 3 Si-(HSi-CH„ I „CH„ |CH„ Mildly polar 35% phenyl methyl dimethyl silicone 325-375 C Separation of TMS-amino compounds} toulene, chloroform, acetone Chromatographic Specialties Ltd. Carbowax HO--CH2-CH2-0- -H n Polar pplyglycol 225° C General G.C. separation. methylene chloride Supelco, Inc. Tenax-GC OH n Porous, high molecular weight linear polymer based on 2, 6-diphenyl-p-phenylene oxide 375-450 C Separation of high boiling point polar compounds. tetrahydro- Alltech furan, carbon Associates, disulphide, Inc. dioxane NJ o APPENDIX B IR AND NMR SPECTRA OF DEA AND RELATED DEGRADATION PRODUCTS Figures B-l (a) & (b), and B-2 (a) & (b) give IR spectra of DEA, HEOD, HEP and TEHEED respectively. Once pure components can be isolated from the degraded DEA solutions, their identification by infrared spectroscopy should be very straight forward. Furthermore, other isolated components may also be characterized according to their functional groups (37, 41). NMR spectroscopy has apparently not been considered for the analysis of DEA samples because of its complexity (52, 53) The main usage of NMR in organic chemistry is to provide information on the structural formula of compounds that have previously been analyzed by IR and MS and whose empirical formula and functional groups are therefore known. Figure B-3 through B-5 are NMR spectra of DEA, HEP and HEOD respectively obtained from the Sadtler NMR Spectra (54). Since this instrument is also available in this laboratory, it may be used to confirm the findings of other characterization techniques by comparing the NMR spectrum of an isolated component to the published spectra of a list of possible compounds. I.r-ltaSCDEraUHCt. OVO 0«Hu«Oj Bol. Ft. 105.14 HL P. M°D (lit.) CU' WOO 4000 1000 2SO0 JOOO 1S0O I4C0 U00 1200 1100 tool. SHCTKOGAAM _ «oo 7»5 8.'10 10 II Bttmil 3alW (a) (b) FIGURE B-l IR spectra of (a) DEA and (b) HEOD. i—1 to FIGURE B-2: IR spectra of (a) HEP and (b) TEHEED. h-1 NJ OJ f n l: i •• i-f: DO-CHJCHJ-HO-CWJCHJ-OD ; " j"'r~ 111 • • i • • • i • • • I : Mi [I::1 ill-r:: 2,2'-IMINCDIETKAKOL C4HUN02 Moi. wc. 105.14 M. Source: Jefferson Chemical Company, Inc. Houston, Texas 4 Filter bandwidth: Sweep time: Sweep width: Sweep offset: Spectrum cmp: Integra! orop: Cone. . 60mg/'0.5mi D20 i0 500 10 cps sec cps cps 80 (spec, amp. 3.2) 28°C 2.74 3.69 IR 5830; 10637 ASSIGNMENTS h 4.70 HDO j) FIGURE B-3: NMR spectrum of partially deuterated DEA. 1,4-PIPERAZINEDIETHANOL IR 18447 C8H18N2°2 Moi. Wt'. 174.24 Source: Aldrich Chemical Company, Inc., Milwaukee, Wisconsin Filter bandwidth: Sweep time: Sweep width: Sweep offset: Spectrum amp: Integral amp ASSIGNMENTS N.M.R. ©1972 250 500 32 80 (spec, amp. 12.5) a ca 2 40 h Hi b •s 48 i sec c 4 .25 i Hz d 2 .52 DMSO-dj k Hz e 3 ,30 H,0 1 f 2 m g n Solvent; Dimethyl sulfoxide-d^ FIGURE B-4: NMR spectrum of HEP. to 3-(2-HYDROXYETHYL)-2-OXAZOLIDINONE C5H9NO3 Moi. Wt. 131.13 Source: A. Kutner, Hercules Powder Co., Wilmington, Delaware 250 500 Filter Bandwidth: Sweep time: -Sweep width: Sweep offset: _ Spscfrum omp: _ Integral omp: 80 (spec, amp. 1.6) Cone. 80mg/0.5ml CDCI3 12.5 cps sec cps cps B. P. 150-155°C/0.25mm ASSIGNMENTS o 3.36 b c d e 3.66 3.75 3.99 4.37 IR 20793 FIGURE B-5: NMR spectrum of HEOD. to 127 APPENDIX C PAPER CHROMATOGRAPHIC SEPARATION OF SOME DEA DEGRADATION PRODUCTS As Zinimerman (27) has reported, amino compounds can be separated and identified by paper chromatography using very common chemical reagents and detection techniques. Once the separation has been accomplished, individual components can then be re-dissolved in suitable solvents for further analysis by other analytical techniques which are not suitable for mixtures. Initial experiments with paper chromatography have achieved some success with the separation and identification of DEA, HEP and TEHEED. The procedure was as follows: Whatman No. 1 filter paper was spotted with aqueous solutions of the above compounds about an inch from the bottom. Then the paper was lowered into a tank containing the solvent system: butanol, 4 parts/glacial acetic acid, 1 part/water, 5 parts/. After the end of a run, the paper was withdrawn from the solvent tank and the solvent front marked while still wet. Next, the paper was dried in air and the paper was sprayed with 0.25 wt % ninhydrin in acetone followed by heating at 105° C for 5 minutes to develop the chromatogram. The DEA, HEP and TEriwi) reacted with ninhydrin producing spots of white, reddish brown and dark brown colours, respectively. The separation of the three compounds was not too good but it was believed that successful and effective separations could be achieved by changing the polarity of the solvent and switching the stationary phase medium from paper to special TLC plates, thus minimizing the spreading of the spots (18, 36). The successful development of TLC will enable the separation of individual components in aqueous solution, at room temperatures thus minimizing the possibility of sample decxitposition and/br further degradation. Furthermore, the isolated compounds can then be subjected to IR and M.S. analyses without interferences from other components. These two techniques are often sufficient to identify and characterize unknown compounds. APPENDIX D A LISTING OF THE COMPUTER PROGRAM USED IN PREDICTING THE THEORETICAL CONCENTRATIONS OF DEA, HEP AND HEI AS WELL AS THE REACTION RATE CONSTANTS OF THE PROPOSED DEA. DEGRADATION MODEL 1 C PROGR. NAME: EC3 (A.M., MARCH 1970) 2 C PROGR. CALCULATES JL A , HECJO AND HEP CONCENTRATIONS FOR  3 C GIVEN VALUES OF Kl AND K2 4 C REACTIONS: DEA + C02 -> HECD + H20 (RATE CONST. IS Kl) 5 C 2 HEOD _-> HEP • 2C02 (RATE CONST.. IS K2 ) 6 C LET CTHE(I,l)iCE XP( I , i ) = THE OR . , EXP.CO NCOFDEAATI"-1HRS """ 7 C CTHElI,2>,CEXP<I,2)=THEOR., EXP. CONC OF HEOD AT 1-1 HRS 8 C CTHEI I ,3 ) , CEXP ( I , ~i )=THEOft. , EXP. CONC GF HEP AT 1-1 HRS ~ COMMON /RATE/NDATA,K0L:A,K1,K2 10 REAL KDEA,K1,K2 11 _ COMMON /CU7/CTHE(12,3),CEXP(12,3),EOEA(12),EHEP(12),SDEA,SHEP 12C""" READTEMP.A NO" EX P ERIMENT A LV A L" U ES0 F D E A , H EP" 13 READt 5 , 1000) NDATA, TEMP ,K.DEA ' \ 14 1000 FORMAT(12,2F10.U  T5 DO 100 1-1 , N DA T A : : ~ 16 REA0(5,2000)T,CEXP(I,1),CEXP<I,3) 17 2000 _ FORMAT (F7. 1, 2E 15 .5) '  "18 . " ~T6'0~ CONTINUE "~* 19 C RF AD VALUES OF Kl AND K2 20 READ(5,3000)K1,K2 21 3000 FORMAT12F10.2) 22 C ENTER SEARCH ROUTINE 23 CALL SEARCH "24 CWR I TIEN OUT °UT 25 WRITE(6,10000)TEMP,Kl,K2 26 10000 FORMAT ('1' ,'TEMP-=» ,Fb. 1 , ' OEGR.C ' , 5X,' K1=',1PE14.5 , ' 1/HR', 27 15X, 'KJ^.clt.S,1 CM**3/(GM0LE*HR) < ,//> 28 WRITE(6,llll) 29 _lll_L_ FORMAT (' MR DEA TH JJEA EXP DEA OIF HEOO',  30 14X, "HEPth HEPEXP"HEPDIE • j "" 31 00 400 1 = 1 tNDATA 32 IT=I-1  33 IF(CEXP(I,1).GE•0.0.AND.CEXP(I»3).GE.0.0)GO TU 410 34 IFCCEXPII,ll.LT.0.0.A\D.CEXP(I,3).LI.0.0)GO TO 420 35 IFICEXP(I, 1 ) .LT.O.O.U^.CEXPtI,3).LT.O.OIGO TO 430 36" 410 WR I IF 16,1200011 T,C THE ( 1,1) »CC.XP (1 , 1) ,EDEA(Ij,CTHE ( I . 2 ) , 37 ICTHEtI,3).CEXPII,3),EHEPII) 38 12000 FORMAT(' ' ,I2.1PL9 .2,fc£9.2)  130 39 GO TO 400 40 420 WRITE<6,13000)IT,(CTHE<I,J),J=l,3) Al 13000 FORMAT(' ',I2,1PE9.2,18X,2E9.2) 42 GO TO 400 43 430 IF (CCXPII,I) .GE.O.OJGU TO 440 44 WRITE(6, 14000) IT,CTHE( I,1 ),CTHE(I,2) »C THE t 1,3), 45 1CEXPJI,31,EHfcPlI) 46 14000 FORMAT(' • , 12.1PE9.2,18X.5E9.2) 47 GO TO 400 48 440 WRIILI 6, 1500 6) 1t,CTHE(1 ,1) ,CEXP (I, L),EDEAlI),C THE( 1,2), 49 1CTHEII,3) 50 15000 FORMAT!' •,I2,1PE9.2,4E9.2I 51 400 CONTINUE 52 WRITE(6, 17 000)SOFA , SHEP 53 17000 FORMAT!/, •SOEA=' , IP El 5. 5 , 5X , • SHEP = ' , E 1 5. 5 J 54 STOP 55 ENO 56 SUBROUTINE SUE INT 57 C . SU3R. 'SOLVES THE RATE EQUATIONS -58 EXTERNAL FUNC ; 59 COMMON /OJT/CTHE(12,3) ,CEXP< 1?,31,E DEA(12) ,EHE f*( 12),SDEA,SHEP 60 DIMENSION C(3) ,N3),&( 3),S(3),t(i) 61 COMMON /RATE/N DA TA,K0EA,K1,K2 62 REAL KDEA,Kl,K2 63 TI=0. 64 C( 1 l = CEXP( 1,1) 65 C(2)=0. 66 C<3)=CEXP(1,3) 67 CTHEI 1, U = CtXP ( 1 , 1 ) 68 CTHE(.1,2)=0. 69 CTHEI1,3l=CEXPl1 , 3) 70 H=0.02 71 HMIN=0.005 72 E=1.0E-05 73 ND1=NDATA-1 74 DO 100 I=1,ND1 75 TF=TI+1. 76 CALL RKC<3,TI,TF,C,F,H,HMIN,E,FUNC,G, S,T) 77 DO 120 J=l,3 78 120 CTHE(1+1,J)=C<J) 79 100 CONTINUE 80 RETURN 81 END 82 SUBROUTINE FUNC(T.CP) 83 DI MENS I ON C(3) , F(3 ) 84 COMMON /RATE/NDATA ,KDEA,rU,K2 85 REAL KDEA,K1,K2 86 K1=SQRT(K1*K1) 131 87 88 89 K2=SQRT(K2*N2> F(1 I=-KDEA*C(1 ) r{Pl = Kl*C( 1)-K2*C(2)**2 90 "I- I3>=(K2*Cl2)**2)/2. 91 RETURN 92 FND — 93 94 C SUBROUTINE SEARCH ROUTINE PERFORMS SIMPLX SEARCH 95 C OMMON / R AT E/ti 0 AT A ,K DE A , K 1 ,K2 ,-p 96 97 98 REAL KDEA,K1,K2 DIMENSION XK.! 3,2) , STEP (2) 99 EXTERNAL FUNCT 100 101 XM1,1! = K1 XK(1,2) = K2 . . .. . _ . _. 102 - CALL SIM PL X(F,XK ,1,3,STEP,1,50,25,l.E-6,FUNCT,4999) 103 1 OA 999 GO TO 100 WRITE!6,10 00) 105 1000 FORMAT!' SIMPLX SEARCH FAILED') 106 100 RETURN 107 FND ••• -— 108 FUNC1 I ON FUNCT(XK.NDIMX,IZ) 109 DIMENSION XKtNUIMX.l) 1 10 REAL KOEA.Kl, K.2 111 112 COMMON / R AT E/NDA T A , K.DE A , K 1, K2 CMCD COMMON /0UT/CTHE!12,3), CEXP(12,3),EDEA(12),EHEP(12),iDEA,SHEP 113 K1 = XKUZ,1) — 114 115 c K2=XK(IZ,2) CALL TO INTEGRATION SUBROUTINE 116 CALL SUB INT 117 c ERROR CALCULAT IONS t-OR DEA AND HEP CONCS 118 SDEA=0. 119 SHEP=0. , 120 121 122 DO 30 0 1 = 1 ,NDATA IF(CEXP(I,11.LT.O.IGO TC 250 EOEA!I)=CTHE(I,1l-CFXP(1, 1) 123 124 IFlCTHr(I,l).LE.0.0.UR.CEXPlI,l).LE.0.)GU IU SOEA=SOEAMCTHE< I, 1J-CEXP! I, 1) )**2 125 250 IF (CEXPl I, 3) .LT.O. IGO TO 300 _ 126 127 128 EHEP(I)=CTHt(I,3)-LtXPU,3) IF(CTHE(I,3).LE.0.0.0K.CEXP(I,3).LE.O.)GO TO 300 SHF.P=SHEP+(CTHE<I,3)-CEXP( 1,3) >**2 129 300 CONTINUE 130 SHEP=SHEP*!l.OE+10) 131 FUNCT= SHE? ^ . 132 •- " I"F"{ ABSTK 1 r.GT. 0:i2"ii91 j FUNC T-FUNCfflOO.*AOS ( Ki ) 133 RETURN 134 END _ 132 APPENDIX E INDUSTF<IAL DEA SAMPLES Thirty four DEA and two MEA gas-treating samples were received from Canadian natural gas processing plants and petroleum refineries. These samples were analyzed by the newly developed G.C. technique. Figure E-1 and E-2 show typical chromatogram of an industrial DEA sample analyzed by a G.C. column temperature of 142° C and 174° C respectively. In general, the chromatograms of industrial samples were found to be very similar to the high pressure DEA samples produced in laboratory tests, thus confirming that DEA and C02 react similarly under the two conditions. For the low column temperature analysis (Fig. E-1), only the peak size of degradation products varied and this might have been caused by different operating conditions. As for the high column temperature analysis (Fig. E-2), the unidentified compound with an elution time of 33 minutes is missing from the industrial sample. Furthermore, the peak with an elution time of 42.6 minutes is very broad and irregular, thus indicating that more than one compound is eluted. The results obtained for the 36 industrial samples are tabulated in Table E-1 together with a sample obtained from the UBC high pressure test. In the latter case (sample #37) , a 30% DEA solution was subjected Jto_pure CC^ for 4 hours at a temperature and pressure of 185° C and 600 psig, respectively. ]VLTOUTES FIGURE E-1: Typical chroiratogram of a DEA gas-treating sample obtained from industrial source. (Sensitivity: l-512xl0~ln AFS; 2-32x10-10 AFS; 3-8xl0~10 AFS; 4-32x10 11 AFS; 5-32xl0"12 AFS. Condition as given in Table 10.) j-1 MINUTES FIGURE E-2: Typical chromatogram of a DEA gas-treatang sample obtained from an industrial source. (Sensitivity: 1-512x10-10 AFS; 2-4x10 10 AFS; 3-32xl0-11 AFS; 4-64x10-12 AFS; 5-32xl0-1^ AFS6 Condition as given in Table 10 with the column temperature at 174 C.) 135 tt O »: tt tt * • » tt:« » » » tt » • t> «l •6 < »• X a a * * « » «l » «.» » « » X 3 i o » # » Ii> * »- • O — LU O, IU LU 5 JS » 4 • « * «| •n < aZ E l/l — -tf ) < a "J . • *>**#•»! * UJ 4> o • >~ u «• — » < -jf *>! i/> X » * » * tt • » a * « « I tt ^ '-/I < I •» c ft. < ! t 4 ft I « tt * * tt I 4 «. 4> «> * (t CC » » tt ' u * ft < . *» i/» tt ST *| - o O TS LO c -(MOO ..(cm O O O' NO O •» tt tt • tt tt'tt tt tt • tt « •»; 4> • - JC ft. x o a. < -J ft. * • 3C *;' »• . < » i •» ; . tt; • ^ < O - a _J > Z> o « n Cl : *t-l ' O ft.; C a o. 3i U. 3 «3' : - I ar: •*» Cf < »- w o — JI/I < o oi-I -i « tt tt «. » * «i tt * ttj * *> »h i> * » • •o • « ttj tt tt ••!•*• « » *:tt *> • Oi « tt tt! • *- a » i TABLE E-1: Analysis of industrial DEA gas-treating samples. (continued) ..•CLEAR »BPUWNISH» _»5..9*E-0*»1.6*E-05» 2,0200»..0.5583*_5.72*0«_*.1500«_ »17« _.*18» _ _»2.L»_ * * • 23* • CL EAR f . •YEllOh. •2.52E-03»1.38E-06» 3.*700* 0.39lO» 0.0299* 0.08*0* •CLEAR, •YELLOh • YFIIOw. » 110 910* 190 12 *STEAH •3.75E-05»*.17E-05* 1.9600* 0.8090* 0.1330* l.*300« f Fl 1TF »l.**E-0**?.58r-0S» 2.2600* 0.6B10* O.0BBB* 0.2083* •8L. PPI • LEAF FILTER** ACTIV4TE0 » CARBON...F ILT.« . • SALES •CLCUDY, FAC-«YFLL0W 'GAS 950*258-» 2b5 • 65 PSIG •STEAM  SULK F 1LTER tCAKnON »d.13E-05*!.91E-0** 2.2700* 0.8373* 0.0156» 0.1200* * 1LITY •STAGE 5 • FUEL GAS*CLEAK t • TREAT INC* BRGhN •Far. 11 I TY* • H2S tREODISH-• STRIPPERtYELLUii *PET.REF.*BL. PPT • H2S »RtDUISH-•STRIPPER*YELLOW _»PET, .REF.'BL. .PPT . * * •STAGE I!*CLCAR, »FX Fl I - »r,°.FUl; 175-* 225 210* TREATING PRECQAT • »1 JIA. EARTil* • SfjCK F 11 TFR » ,56E-05*8.00E-06« 1.7000* 0.3520* 0.021** 0.*3*0» *1.83E.-03*3.75E-07*.0.6ft2O*_J3J086 7*„.O^1.8B0»_0..6210*_ •2.21E-03*3.13E-07» 0.7790* 0.0900« 0.2750* 0.6070* • 0.00E-01»3.52E-05» 2.6700* 0.36*0* 0.19*0* 2.5*00* •STAGE II*CLEAR, » •EX SURGL*GPfCN-• TAfJK *YELLOw •SURGE *CLGUDY, *TANK *YELLGrt. « *BL.._PPT.. * * • PLANT «'i*CLEAR, • »YELLUW »PLANT_»2*CLCAR, i~" " •YELLOW. * *BL. PCT .• * . 38 E -05 * 3 . 2 1Er.0.5 » _3 , * 3_00• _0_1 73_*Q*_Q^221S*_11&500»_ »3.13E-05*2.59E-05* 1.7700* 0.*720* 0.1380* 1.0200* • 9.38E-0b»5.13E-06* 1.7800* 0.<>930» 0.0105* 0.0*15* •8.75E-05«1.63E-05* 1.2*00* 0.2830* 0.0165» I.0100* TABLE E-1: Analysis of industrial DEA gas-treating samples.. (continued) •25* < PLANT Jf 1 * CL 6 AH « *iie-< 240-*192- * 1* *STEAM * LEAF F ILTER *2.22E -03*2.75E- 060 1 .3400* 0.0905* 0.5910* 3.1200* ft * * 0D.ARK o 165* 1050* 251* * * 0 * * » 0 ft ft 4 •RROWN _* 4 . * * . _ • » ft ; •_ 0 ft *_ ft * 4 * * 4 ft 0 * ft ft ft * * • 26* < PLANT *2*CLEARf *llB-< 240-*192- ft 14 * STEAM 4 LEAF FILTER •2.29E -0305.81E- 060 3.1200* 0.1610* 0.4130* 3.25000 * » '* * *DAKK * 165* 1050* 251 * 0 • « 0 0 0 ft * * * < ' . *bRO'*N * 4 0 ft 0 * 0 0 ft ft ft 0 c 0 * 4 6 ft 0 * ft 0 0 ft • • 27* - < •CLEAR, » i ft ft *6.68E -05*5.19E- .05* ?».9_400* A • 29.00*. 0.0592* 0.4030*_ * • 4 * BR.-V ELL* * t ft * « * 0 0 0 ft 0 ft * * ft ' 4 * 0 ft ft 0 ft 0 0 « *?fl* < •ORECM- * 4 150-* 210 ft 14 * HOT OIL t LE AFKARBON •4.38E -03*2.45E- 05* 1.5100* 2.3600* 0.0000* 0.14600 • 0YE LLOh t • 4 950* ft FJL7ER ft • 0 * ft ft ft * 0 4 * BL. PPT ft 4 • ft ft ft 0 • 0 0 ft ft 0 * *-. _ . * « ft ft * ft * _ _ 0 ft ft 029* « •BROWN- * 99-4 123-*190- • 11.5 STEAM LEAF»CAR8QN**3*75E -05*7.25E- Ob* 2 .82000 T.~66co* 0.5680* 1.9500* * « < •YELLUW * 130< 930* 244* SOCK FILTER 0 ft 0 * 0 ft 0 9 0 <! - *ht . PPT » • * * 0 0 0 0 ft 0 * • 4 * * 4 * ft 0 ft 0 ft 0 ft ft • *30* * ME A UNlT*GKf.EN- t> 4 ft * *0.00E -01*0.OOE- 01* 0.0000* 0.0000* 0.0000* 0.0000* * • 1 PBl.Mf: .*YFLLUH». * 4 ft ft 0 * 0 ft 0 0 0 * * < *0L. PPT 9 * ft * ft * 0 ft ft * ft 4 4 ft < 0 • ft 0 fl 0 0 0 0 31 * < ME A RF- * CRF FN- 4> 4 * *O.O0E -01*0.OOE- 010 0.0000* 0.0000* 0.0000* 0.0000* CLAIMCR *HFAVY • 4 « ft 0 0 0 0 0 0 0 • * 1 PET.REF.*RL, PPT 0 • • * * 9 ft 0 0 0 0 • * 4 * ft * * * 0 0 0 0 « 0 9_ #32* * PE T • R Ef .* *CL £AR, ft* 4 ft ft* ** • ft F0 •2.38E -04*4.38E- 07* 0.2520* 0.26C0* 1.4400* 0.0790* • 0 ^ •GOLOEN- 9 4 • * • ft 0 # 0 0 0 0 # * 4 « *YFIInw * > * o ft 0 • ft 0 0 0 0 • 0 1 * * * * 4 » ' ft ft O ft * 0 0 ft ft • 33* « •REGENE- *CLEAR, ft • * 0 *2.11E -03*1.56E- 06* 1.9000* 0.03950 3.7500* 0.0372* 00 4 RA.TOR *.YLLLGW„. * > • 0 0 0 ft * 0 0 » • * • PET .REF.* ft ( r ft * 0 0 0 ft 0 ft 0 0 * * • * F ft 0 0 0 0 ft 0 ft * • 14* * * CL L-AR t * 100- » 21-4-731- * t=H20*GLY- * S3 CK FILTER * l.tiBE -0401.06E- 06* itUOO* 0.9430* 0.9500* 0.0653* * Y tLLUA * 106 > 500» 236* >CUL 0 0 * * ft • • • * * * > 9 ft ft 0 0 0 0 0 0 • 35* < • *CLEAK * 108 > _090*184- •10,5 •ST.EAM_._ *LCAF,C.ARBON«-*?.38E T05»6-31E- 0 6* 2... 5203* .0,6820* _Q._0314* _2..5200*_ t » * > * 202* •SOCK FILTER 0 0 0 0 0 0 0 ft * * F * * > « * 0 ft 0 _ £ 0 0 0 03h0 > *ci lllJOY * * « * •1.90E -0-3*1 .56F- 06* 0.5320* 0.5 5/00 0.0184* o. 11?nft f *YELLUK * > ft * 0 ft 0 0 0 & 4 * r 0 * * « * * 0 ft ft 0 ft 0 0370 • imr. t Aft.*YFLLUi. * 345. ».600* * •9.38E -05*3. 26ET 05* .0.. 0000 0 0..6.35Q* OtOZSO* * * ft ft * * 0 0 0 0 0 *. 0****0*00000 |>*#».«****9r***»«***ft*ft9*?*0ft**0*0*ftftftft0 «****«**« » * • * ft * * 4 * * ft ft NflT F : C Q Nf. F ti r RAT I DNS QF T.X.Y .Z ARE EXPRESSED AS RATIO GF THEIR PEAK AREAS TU AREA OF DECANOLIC ACID INTERNAL STANDARD 

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 16 3
Japan 5 0
Canada 3 0
France 3 0
China 3 11
Iran 2 0
Finland 2 0
Venezuela 2 1
Netherlands 2 1
Brazil 2 0
United Kingdom 1 0
Indonesia 1 1
Italy 1 0
City Views Downloads
Unknown 15 2
Tokyo 5 0
Washington 4 0
Ashburn 3 0
Beijing 3 0
Redmond 3 2
Kauniainen 2 0
Regina 2 0
La Porte 2 0
Waterford 1 0
Oklahoma City 1 0
Leiden 1 1
Leeds 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058692/manifest

Comment

Related Items