Analytical Solid-Phase. > cer Toeto o © ames S. Frit ® WILEY-VCHANALYTICAL SOLID-PHASE EXTRACTION‘This book is printed on acid-free pape. ‘Copyright © 1999 by John Wiley & Sons, Ic, All rights reserve Poblished luancously in Canada, 'No pat of tis publication may be reproduced, stored in retrieval system or transmitted in any form ‘orby any means electronic, mechanical, photocopying recording seanning or oherwise excep as permited under Sections 107 or 108 ofthe 1976 United States Copyright Act, without ether the Prior writen permission ofthe Publisher, or authorization through payment of the peop per: ‘copy fe tothe Copyright Clearance Cee, 222 Rosewood Drie, Danvers. MA 01923, (978) 750- 400, fax (978) 750.474, Requests to the Publisher fr permission should be addesed to the Permissions Department, John Wiley & Sons, Ie, 605 Thind Avenue, New York, NY 10138-0012, (212) 880-4011, fa 212) 850-4008, F-Mall PERMREQ @ WILEY.COM. Forerdring and customer service, call 1-800-CALL-WILEY. Library of Congress Catalog Card Number: 67-13943 Fritz, ames 8, James Sherwood), 1924- ‘Analytical soié phase extraction James S. Fiz pom Includes bibliographical erences and index. ISBN 0-471-24667-0 alk paper) 1. Extraction (Chemistry) 1 Title QD ESAFSS 1999 548.0862-de21__98.43758 Printed inthe United States of America. 1098765432CONTENTS PREFACE . . : xi ACKNOWLEDGMENTS ........---+ +000 xt 1 INTRODUCTION AND PRINCIPLES nt LI Introduction / 1 1.2. Principles / 4 1.3. Advantages of SPE / 13 2. SPE IN THE 1970s: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER . . 15 21 Introduction / 15 2.2. Comparison Between Activated Carbon and Porous Resin Adsorbents /.17 23. Purification of Resins / 17 2.4 Sampling and Sample Handling / 18 2.5. Preparation of Standards / 21 2.6 General Procedure for SPE / 22 2.7 Recovery Data / 26vill ‘CONTENTS 3 SOLID PARTICLES FOR SOLID-PHASE EXTRACTION OF ORGANIC COMPOUNDS FROMWATER ..... Introduction / 28 Desirable Properties of SPE Particles / 29 3.3. Bonded-Phase Silica Particles / 34 3.4 Organic Polymeric Adsorbents / 39 3.5 Carbons / 51 3.6 Comparison of Sorbents / 54 3.7. Molecular Sieves / 56 3.8. Sorbents for Normal-Phase SPE / 58 3.9. Other Sorbents / 60 bog FJ 4 PRACTICAL CONSIDERATIONS: EQUIPMENT AND TECHNIQUES 2... 1.0 eee eee eee ee 68 4.1. The Four Steps of Solid-Phase Extraction / 63 4.2 Apparatus / 67 4.3. Completeness of Extraction / 71 44 Elution / 77 4.5. Solid Samples and Slurries / 81 4.6 Automation / 83 5 ION-EXCHANGE SORBENTS 5.1 Introduction / 89 5.2. lon-Exchange SPE Processes / 92 5.3 Retention of Neutral Analytes by Sulfonated Resins / 96 5.4 Retention of Cations by Sulfonated Resins; Group Separations / 101 5.5 SPE with Anion-Exchange Resins / 109 6 RESIN-LOADED MEMBRANES Introduction / 118 Types of SPE Disks / 121 ‘Techniques for Membrane SPE 122 Semimicrodisk Extraction / 125 Recovery of Test Compounds Using Membrane Disks / 126contents ie 6.6 Applications / 130 6.7 Conclusions / 137 7 PRECONCENTRATION OF METALIONS ....... 140 7.1 Introduction / 140 7.2. lon-Exchange Methods / 141 7.3. Extraction of lon-Association Complexes / 144 s 1147 7.4 Extraction of Metal-Organic Comple) 7.5 Chelating Resins / 154 7.6 Some Conclusions / 157 8 MICROSCALE AND SEMIMICROSCALE. TECHNIQUES ... . 8.1 Solid-Phase Microextraction (SPME) / 161 8.2. Semimicro Solid-Phase Extraction (SM-SPE) / 172 8.3. Miniaturized Solid-Phase Extraction (M-SPE) / 179 84 Conclusions / 183 9 APPLICATIONS . . 9.1 Introduction / 185 9.2. Sample Types / 186 9.3. Techniques / 194 9.4 Problems with Practical Samples / 197 9.5 Applications According to Chemical Class | 199 beeen ee 203PREFACE In chemistry, as in other endeavors, the simplest ideas are often the best Extraction of solutes from an aqueous analytical sample by a liquid organic solvent has been in use for nearly a century (at the time of writing). But organic solvents pollute the air and water and can be messy to work with. ‘So why not use small particles of a porous organic polymer as an extractant in place of a liquid solvent? The feasibility of this approach was confirmed in the early 1970s; ‘numerous publications followed, but it has taken some additional years for its value to be fully appreciated. A symposium at the 1998 Pittsburgh Conference was entitled, “Solid-phase extraction—a neglected orphan.” But this symposium was really a celebration of the fact that solid-phase extraction (SPE) is no longer a neglected technique. On the contrary, SPE is undergoing a remarkable period of growth and increasing popularity. ‘This book was written for scientists who want to know more about SPE for analytical use. All major aspects of SPE are covered: basic principles, historical perspective, materials and equipment, extraction of organic sol- utes from aqueous samples, extraction of polar solutes from apolar organic solutions, ion-exchange SPE, extraction of metal ions, use of membrane disks, solid-phase extraction on microscale and semimicroscale, and s lected applications. The goal is to provide the reader with a real under- standing of how SPE works (principles and theory) as well as what one can do with SPE (scope and limitations). This knowledge will be useful in making intelligent choices among published SPE procedures and in modi- xixi PREFACE fying conditions to meet new situations. It is also hoped that researchers will be stimulated to develop entirely new and better methods for SPE. Janes S. FRITZ, Ames, lowaCHAPTER 1 INTRODUCTION AND PRINCIPLES 1.1 INTRODUCTION 1.1.1. The Importance of Sample Preparation Analytical laboratories are under great pressure to provide analyses more quickly and at lower cost. This burden falls most heavily on the sample preparation portion of the laboratory, which is asked to provide more reproducible results, accommodate lower technical skills, decrease the use of organic solvents, provide cleaner extracts for instrumental measurement, and do everything more quickly and at less cost. ‘The reason why there is such a need for improvement in sample prepa- ration techniques is that the majority of the sample analysis time is spent in preparing the sample. One study (1) showed that more than 60% of analysis, time was spent in sample preparation compared to only about 7% for the actual measurement of the sample constituents. The remainder of the time was taken up by sample collection and data handling In the past liquid-liquid extraction has played a major role in sample cleanup and concentration of the sample components to be measured. However, cecovery of sample components by liquid extraction is seldom complete. Liquid extraction tends to be slow and labor-intensive. More stringent environmental concems are making the use and disposal of large amounts of organic solvents more difficult2 INTRODUCTION AND PRINCIPLES ‘The popularity and use of solid-phase extraction (SPE) is growing at a fast rate. SPE is easily automated, faster, and in general more efficient than liquid-liquid extraction. The particles used in SPE are nonpolluting and the amount of liquid solvents used are tremendously lower than in liquid-liquid. extraction, 1.1.2 What is Solid-Phase Extraction? In solid-phase extraction solutes are extracted from a liquid phase into a solid phase. The solid phase typically consists of small, porous particles of silica with a bonded organic phase or of an organic polymer, such as crosslinked polystyrene. The extraction can take place in a batch mode in which the solid extractant is intimately mixed with the liquid sample solution. In chemical analysis tis more common to pack the solid extractant into a small tube and pass the liquid sample through the tube. A typical apparatus for SPE is shown in Figure 1.1. Solid-phase extraction is not limited tothe use of solid particles o extract solutes from a liquid sample. Air or other gaseous samples can also be passed through a packed tube to extract organic vapors or other substances. present in the sample. Substances that have been extracted by the solid particles can be removed by washing with an appropriate liquid solvent. For example, most organic analytes can be eluted from a SPE tube (column) with an organic solvent such as acetone, acetonitrile, or methanol. Usually, the volume of solvent needed for complete elution of the analytes is much smaller than the original sample volume. A concentration of the analytes is thus achieved. Extracted molecules can ofien be removed from solid particles by heating ina gentle stream of a nonreactive carrier gas. This can be a convenient way to transfer the molecules into a gas chromatograph for analysis. 1.1.3. History and Literature The history of solid-phase extraction dates back at least to the early 1970s, when columns packed with Rohm and Haas XAD resin particles were used to concentrate very low concentrations of organic pollutants from water samples (2). However, activated carbon had been used for several years prior to 1970 to accumulate organic solutes prior to analysis. The earlier work on SPE is discussed in some detail in Chapter 2. During the late 1980s and the 1990s, the development and use of analytical solid-phase extraction have expanded tremendously. A number1.1. INTRODUCTION 3 +— Clean Air Pressure =a T — Reservoir +— Adaptor +— SPE Column is — Absorbent agg 4 gg} —___4, Figure 1.1 Solid-phase extraction apparatus. of reviews and symposia on SPE have been published (3-5). In addition, at least two books on SPE have been published (6,7). At this writing, several companigs that sell equipment and supplies for SPE offer booklets on its, use and extensive bibliographies of SPE applications. These include the following: J.T. Baker, Inc., Phillipsburg, NJ; 3M Co., St. Paul, MN: Varian, Harbor City, CA; and Waters, Milford, MA.4 INTRODUCTION AND PRINCIPLES 1.2. PRINCIPLES ‘1 Comparison of SPE with Liquid-Liquid Extract n In liquid-liquid extraction the objective is to transfer the desired solutes from one liquid solution to another, nonmiscible liquid. Most commonly, the solutes are extracted from a predominately aqueous solution into an organic liquid. Analytical liquid-liquid extraction is usually performed in ‘ separatory funnel so that the two liquid phases can be separated after the extraction. The extractive liquid may be either heavier than water or lighter than water. To carry out an extraction, the extractive solvent is added to the aqueous sample solution, the vessel is tightly stoppered, and the vessel is shaken Yigorously to create a temporary emulsion. The emulsion consists of very small spherical droplets of the extractive liquid suspended in the aqueous Phase. The interfacial contact area between the two phases must be quite large in order to promote rapid mass transfer of the desired solutes from one Phase to another. Even so, itis sometimes necessary to continue shaking for Several minutes to attain a true equilibrium of the solutes between the two phases. When shaking is terminated, the emulsion should break up and the {wo liquid coalesce to form two continuous but nonmiscible liquid phases. A practical difficulty of liquid-liquid extraction is that emulsions some. times break up very slowly or incompletely. In classic analytical liquid extraction, when there is no longer an emul- sion the bottom liquid is carefully drained from the separatory funnel by ‘opening the stopcock very carefully until the last ofthe lower liquid has just been drained off. This can be a tedious operation. When chemical analysis of the extracted solutes cannot be carried out directly in the onganie liquid, itis necessary to do a backextraction. This can often be accomplished with ‘an aqueous solution under conditions (e.g... changed pH) causing the solutes to prefer the aqueous phase. Some of the concepts of liquid-liquid extraction apply to solid-phase extraction. Instead of a temporary liquid emulsion, the extractive material in SPE is a suspension of spherical solid particles in the aqueous sample. ‘The extractive particles typically are ~10-50 jum in diameter with a very large surface area, often ~200-800 m°/g. A large interfacial area between the particles and the sample solution is needed for rapid mass transfer of the extracted solutes from one phase to the other. The extractive particles ‘must be sufficiently dense and large enough in diameter to settle rapidly When agitation of the solid and liquid phases is terminated.12 PRINCIPLES 5 Although SPE can be done in a batch equilibration similar to that used in liquid-liquid extraction, it is much more common to use a small tube (minicolumn) or cartridge packed with the solid particles. The liquid sample is passed through the column, thus coming into intimate contact with the solid particles. With modem solid extractants, equilibrium is rapidly at- tained and the analytes tend to be extracted in a zone near the top of the SPE column. Unlike batch extraction, where there is only a single equilibration of solutes between the two phases, there are in effect multiple equilibrations when SPE is performed with a packed minicolumn, This is because the solutes continuously encounter fresh particles (containing little, if any, extracted solutes) as they pass down the column. A higher percentage of extraction is thus expected in column SPE compared to batch-type extrac- tions, It is usually necessary to transfer extracted analytes from the solid particles to a liquid phase for final measurement. The chemistry of this step will depend on the types of analytes and solid extractants that are used. Organic compounds can usually be cluted from a SPE minicolumn by a small volume ( 10-100 ml H,0;210-100ml__H,0-organic liquid o organic 5-50 fl Partition equilib Single Multiple Multiple ‘Separation of indi- No No Yes vidual analytes Easily automated No Yes No Elution fromex- Seldom needed: Organic liquid H,O-organic ‘active phase HO, pH control liquid mixture Concentration Moderate High Slight achieved 1.2.3 Completeness of Extraction Partitioning of a chemical solute between two phases follows the laws of chemical equilibrium. If a solute, A, exists in only a single chemical form, the distribution coefficient between two phases (K,) is given by Tal (lay Th where [A], and [A], are the concentrations of solute at equilibrium in the two phases. Usually phase 2 is organic and phase 1 is aqueous. The concentration of A in each phase is the mass (the amount) of A divided by the volume of that phase (Mass of A))/V> sy (Mass of A),/V, The ratio of the mass of A in the two phases is defined as the mass distribution ratio, D. Substituting into Equation 1-3 (1.3), we obtain aM aay12 PRINCIPLES 9 where V, and V; are the volumes of the two phases. Note that Ky will be a constant for each system (a particular solute, A, in phases I and 2) whereas the value of D,, will vary according to the volume ratio: (1.5) where it will be seen that the extraction of a solute from phase I to phase 2 will be greater when the relative volume of phase 2 to phase 1 is increased. The fraction of total solute remaining in phase 1 after an extraction (f) is, given by 1 (1.6) D+1 f ‘The fraction extracted (f,) is given by ay If a series of n extractions is performed in which the phases are separated after each extraction and the same volume as before of fresh phase 2 is added, the cumulative fraction of A not extracted is 1 (18) D+1h The total fraction of A extracted from phase 1 into phase 2, is of course, equal to (1 —f) and the percentage extraction is % extraction = 100 (1) (9) In liquid-liquid extraction, phase | is normally water and phase 2 is a liquid organic solvent. However, the same relationships apply equally well to the situation where phase | is water or a water-organic solvent mixture, and phase 2 consists of solid extractive particles of the type used in solid-phase extraction. Indeed, liquid-liquid extraction and liquid—solid extraction (SPE) are similar techniques when each is carried out on a batch basis. When agitated vigorously, two immiscible liquid phases form a temporary emulsion with a large interfacial surface area for rapid mass transfer of solutes from one phase to the other. When agitation is stopped, the emulsion preaks and two distinct liquid phases are re-formed.10 INTRODUCTION AND PRINCIPLES ‘When SPE is performed on a batch basis, small, porous particles of solid extractant are added to the liquid sample. The temporary emulsion formed in liquid-liquid extraction and the porous nature of the particles in solid- phase extraction both provide the large interfacial surface area needed for efficient mass transfer of the solutes. Some agitation is helpful in speeding up the extraction step when SPE is carried out on a batch basis. When equilibrium has been attained between the two phases, the solid particles, can be separated by filtration or simply by carefully pouring off the liquid phase. Although SPE is feasible in a batch-equilibration mode, it is advanta- _geous to use a short column packed with the solid extractant. Passing the liquid sample through such a column provides intimate contact with the solid particles, and there are multiple equilibria as the sample passes through the column and comes into contact with fresh solid extractant particles. In this mode the completeness of extraction will depend on the number of separate equilibrations with fresh solid extractant in the column as well as the mass distribution ratio. (In column operations the number of equilibrations is called the theoretical plate number.) Ina single equilibrium method such as simple liquid-liquid extraction, an “all or nothing” extraction is needed for a quantitative separation of a target solute from other substances in the sample. This means that extraction of the desired sample component should be nearly 100% while extraction of other components should be essentially zero. This situation is often difficult to achieve. SPE in a column provides multiple equilibrations and. therefore requires only a reasonable difference in extractability to separate two solutes. ‘The separation of two substances, A and B, by solid-phase extraction can be illustrated by an example, Suppose that there are eight separate equili- brations with fresh resin in a SPE minicolumn. This is the same as saying, that the column has eight theoretical plates. The fraction of a solute not extracted can be estimated from Equation (1.8). Toapply this equation to the SPE column, we will assume that the column contains equal volumes of solid-phase particles and aqueous sample. A solute, A, that is only 60% extracted by the first plate (D =60/40 = 1.50) will be quantitatively extracted after being in contract with all eight plates: —1_ a.s+D§ 0007 (0.07%) (extraction is 99.93% complete) A solute, B, that is only 1% extracted by the first plate willl have D 1/909 or approximately 0.01. After it has contracted all eight plates, then12 PRINCIPLES " 1 (oF f = 0.92 (92%) (extraction is 8% complete) Because of its very low D value, this 8% of solute B can be quickly rinsed from the column by a little water, while the A, which is mostly near the top of the column, will hardly move. In this way a quantitative separation of A and B can be obtained. 1.2.4 Sample Breakthrough ‘When concentrating dilute samples by solid-phase extraction with a short column, the capacity of the column should be well in excess of the total analyte content of the sample. This being the case, the next important question concerns the maximum sample volume that can be passed through, the SPE column before breakthrough occurs. This will depend on the retention volume of the analyte and the number of theoretical plates in the column. Applications involving concentration by SPE differ from ordinary ana- lytical chromatography in several respects; For instance, the sample enters the column as a front instead of a narrow plug. Since no separation of the analytes is intended, itis not necessary to have a large number of theoretical plates. Finally, the relative retention volume (Vg/V,) must be large in order to permit the use of large sample volumes. ‘When a sample is fed continuously into a column, itis usually assumed that the shape of the eluting front (the breakthrough curve) can be described as the integral of a Gaussian peak. Lovkist and Jénsson (8) examined five equations that have been proposed to describe the shape of the eluting front. Although the results differed for columns containing only one of two theoretical plates, the profiles calculated from the different equations were in reasonable agreement when the column plate number (n) was 5 and in ‘good agreement when n = 25, In Figure 1.4 the normalized flux J/Jp (where Jyand J are the concentrations of analyte in the entering and eluting samples, respectively) is plotted against the relative retention, t (T= W/V). In calculating the plots in Figure 1.4, a linear distribution isotherm was assumed. This is valid forall adsorbents at sufficiently low analyte concen- trations. The linear range in the isotherm is apt to be shorter for carbon than for resin- and bonded-phase silica particles. At 5% breakthrough, (HJ, ‘tis approximately 0.5 for m= 5 and approximately 0.7 for n = 23 (Fig. 1.4). This information can be used to estimate the breakthrough volume (Vp) for solid-phase extractions of given.2 INTRODUCTION AND PRINCIPLES 10 3 Fos oo ope ocr apmticmiang a ooimac mniomar cameo tau tau FIGURE 1.4 Comparison of different front equations as normalized flux (JLo) versus relative retention time 1 = dt, for 5 and 25 theoretical plats. (From Ref. 8 with permission.) dimensions. For example, consider a column L = 10 mm, inner diameter (ID) = 5 mm and a void fraction of the packed column = 0.33. The void volume (Vp) of this packed column is m (2.5/(10)(0.33) = 65 mm = 0.065 mi For an analyte with a capacity factor k= 1000, the retention volume Vg = 0.065 (1001) = 65 ml. If the column has five theoretical plates, the breakthrough volume (Va) at (Jp) = 0.05 is 0.5 x 65 ml = 32.5 ml. If the column has 25 theoretical plates, Vy = 0.7 x 65 = 45.5 ml If the analyte retained on the column can be subsequently eluted by a small volume of an organic solvent, a considerable degree of concentration is possible. In this example it would be reasonable to assume that three void volumes of an organic solvent could elute the analyte completely: 3 Vy = 3 X.0.065 = 0.2 ml. The maximum degree of concentration for n — 25 is 45.5/0.2, or 227-fold In SPE of trace organic analytes from predominately aqueous samples, the equilibria do strongly favor partition into the solid phase in most cases. ‘The capacity factor will, of course, vary with the structure of the analyte and the type of solid particles used, but k = 1000-10,000 or even higher is often possible. The number of theoretical plates in a SPE column is rarely measured as such, but in most cases it is probably small. Miller and Poole (9) found that. the number of plates in a typical SPE cartridge with a bed height of 6 mm varied from NV = 25 at a flow rate of 1 ml/min to N = 9 at a flow rate of 30 ml/min, Commercial cartridges at that time had low and variable packing densities, which sometimes led to poor performance with some channeling.13 ADVANTAGES OF SPE 18 ‘The number of plates in a given column length, and hence the efficiency of a SPE column, can be enhanced by careful packing and by use of adsorptive particles of small, uniform particle size. 1.3 ADVANTAGES OF SPE Solid-phase extraction has several important advantages over liquid-liquid extraction: 1. Faster: Easier Manipulation. A sample can be quickly passed through a SPE column or cartridge by means of a pump or with gentle pressure or suction. After a quick rinse, the extracted substances can be washed from the column by a small volume of an organic solvent or another appropriate eluent. These steps can be automated readily. By contrast, simple solvent extraction requires a considerable amount of manipulation in adding the extractive liquid, shaking, waiting for the emulsion to break, and carefully separating the two liquid phases. Often a washing step and a backextraction are required. Although robotics can sometimes be applied, the steps in- volved in liquid-liquid extraction are inherently more complicated than those in SPE. 2. Much Smaller Amounts of Liquid Organic Solvents Used. The large quantities of organic solvents used in analytical separations have become an important environmental concern. Aqueous samples become contami- nated with organic solvents and evaporative concentration of the extracts pollutes the air with organic vapors. Proper disposal of used organic solvents has become troublesome and expensive, The U.S. Environmental Protection ‘Agency (EPA) is working to replace liquid-liquid extraction with solid- phase extraction in analytical procedures. It is difficult to completely remove all organic and metal impurities from organic liquids. When rela- tively large volumes of organic solvents are used, the sample extract will be contaminated with these impurities. 3. Less Stringent Requirements for Separation. SPE is a multistage separation method and as such requires only a reasonable difference in extractability to separate two solutes 4, Higher Concentration Factors. The concentration factor is how many times more concentrated a substance is in the extract than itisin the original sample. In SPE, concentration factors of 1000 or more are possible. The concentration factor in liquid-liquid extraction depends in part on the volume ratio of the two liquids. Using vortex mixing it may be possible to4 IvTRODUCTION AND PRINCIPLES extract a 100-ml sample with perhaps I ml of an organic solvent. This could give a concentration factor of 100. However, the distribution ratio of the solute would need to be very high (~10*) to obtain complete extraction under these conditions. REFERENCES 1. W. Pipkin, Am. Lab. (Nov. 1990) 40D. 2. A.K. Burnham, G. V. Calder, J.S. Fritz, G, A. Junk, H. J, Svec, and R. Willis, Anal. Chem. 44 (1972) 139. 3. S.K. Poole, T. A. Dean, J. W, Oudsema, and C, F, Poole, Anal. Chim. Acta 236 (1990) 342. 4, M.Zief and R. Kiser, Am. Lab, (Jan, 1990) 70-82. 5. M.S. Mills and E, M. Thurman, J. Chromatogr. 629 (1993) 1-93, 6. N. Simpson, Solid Phase Extraction: Principles, Strategies, and Applica- tions, Marcel Dekker, New York, 1997. 7. E.M. Thurman and M. S. Mills, Solid-Phase Extraction, Principles and Practice, Wiley, New York, 1998, P. Livkist and J. A. Jonsson, Anal. Chem, $9 (1987) 818. 9. K.G. Miller and C.F. Poole, J. High Resol. Chromatogr. 17 (1994) 125.CHAPTER 2 SPE IN THE 1970s: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER 2.1 INTRODUCTION The early to mid-1970s was a period of rapid growth and development for solid-phase extraction. Actually, the technique was not given the name solid-phase extraction dusing that period. Solid particles such as activated carbon and porous polymeric resins were known as sorbents for trace organic constituents in aqueous samples. Columns filled witha solid sorbent ‘were sometimes called accumulation columns. Although the potable water supplies of cities and towns were commonly analyzed for inorganic constituents and biological organisms, very little was known in 1970 about the organic compounds naturally present in water or that were present because of some contamination. C. C, Johnson, on behalf of the American Public Health Association, stated in May 1, 1971 (1): We are considerably concerned about the amounts of chemical materials such as the wide spectrum of organics including pesticides and other chlorinated organics which may find their way into our water supplies and about the removal of these noxious agents from our water supplies. These similar concerns are held for those waters used in food processing and production. In 1972 (2) Bumham et al. published the results of an investigation of the pollutants in the water supply of Ames, Towa, which had an undesirable 1516 ‘SPE IN THE 1970s: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER taste and odor when water from certain wells was used. The offensive material had been tentatively identified as phenols and cresols. This con- clusion was based on a marginally positive 4-aminoantipyrene color test. A preconcentration of the offending chemicals was necessary to increase their concentration toa level at which their analytical detection and identification might be possible. The method developed was to pass 150 liters of water through acolumn (1.5 x7.0.cm) filled with a porous polymeric resin (Rohm. and Haas XAD-2). The flow rate through the column was 50 ml/min (4.0 bed volumes/min), The extracted organic pollutants were eluted from the column with 15 ml of ethyl ether. The ether effluent was carefully evapo- rated to a small volume. Injection of a portion of the ether effluent into a gas chromatograph gave a number of well-separated peaks, corresponding to the organic pollutants, that had been extracted from the water. However, identification of the peaks presented a real problem. Fortunately, the investigators had one of the first gas chromatography-mass spectroscopy (GC-MS) setups in existence and ‘were able to identify the compounds listed in Table 2.1 (2). ‘The compounds identified are consistent with the source of contamina- tion, which was believed to be residues from a coal gas plant operated in the city of Ames in the 1920s, The tar residues were buried in a pit that was connected hydrologically to an aquifer supplying the city water. ted Ames, Towa Well Concentration ppb Standard Deviation TABLE 2.1 Neutral Compounds in a ‘Name of Com Acenaphthylene 193 14 |-Methylnaphthalene 10 06 Methylindenes (Iwo isomers) 188 08 Indene 18.0 Ls Acenaphthene 7 02 2-2-Benzothiophene 037 ol Isopropylbenzene _ Ethylbenzene = Naphthalene - -Dihydroindene 158 Alkyl-2,3-dihydroindene Alkyl benzenes = Alkyl benzothiophenes ~ Alkyl naphthalenes e23 PURIFICATION OF RESINS "7 2.2. COMPARISON OF ACTIVATED CARBON AND POROUS RESIN ADSORBENTS Prior to 1972, the standard method for concentrating low concentrations of ‘organic compounds from water samples was the carbon adsorption method (CAM). The severe limitations of this method have been documented (3). One aspect of the (analytical) problem is that the carbon adsorption method recovers only a low percentage of the organic substances from water. It does not adsorb all the organics present; moreover a large portion of those adsorbed are not removed by the chloroform extraction, This indicates the need for a better method to collect and measure the organics in drinking water. ‘A second problem in the area of organic sampling is the recovery of the organics from water in unchanged form. The CAM can be used as a gross indication of the concentration of organics in water, but the residues recovered are not always in their original form, but may be altered by the use of solvents and heat. A need exists to recover the organics in an unchanged form if they are to be identified and used in determining their toxicity via animal experiments. Later studies presented extensive data comparing the recoveries of organic compounds from water by carbon adsorption and resin adsorption methods (4,5). The results of one study are summarized in Table 2.2 (5). ‘These results clearly suggest that organic polymeric resins are superior to activated carbon for isolating organic compounds from water. Low recoveries can result from either breakthrough (incomplete extrac- tion) or incomplete elution. It is possible that the latter may be the cause of some of the low recoveries in Table 2.2. 2.3 PURIFICATION OF RESINS It is vital that the resins used as sorbents be as free as possible from impurities that might be leached from the solid particles during the elution step. The Rohm and Haas resins (XAD-2 and XAD-4) that were used during the 1970s often had to be ground up and sieved to obtain a particle size sufficiently small for efficient column extraction. The grinding of spherical resin particles exposed new resin surfaces to the leaching action of the organic solvent used in the elution step. The recommended purification method was to purify the sized resin particles by sequential extraction with methanol, acetonitrile, and diethyl18 SSPE IN THE 10706: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER TABLE 2.2 Comparative Recovery Efficiency for Carbon and Resin Number of Compound Type Compounds Carbon XAD-2 Resin Alcohols 2 48 100 Aldehydes + ketones 3 4 4 Alkanes 5 B 5 Amines 3 4 14 Aromatics 7 6 68 Benzothiazoles 2 2 67 Esters + ethers 9 52 "4 Halogenated 9 22 7 ‘compounds PCBs 9 2 B Phenols 2 10 46 Toul 58 Average 2 59 percent recovery ether in a Soxhlet extractor for 8 h per solvent (6). Another effective method ‘was to purge a heated resin-packed tube with nitrogen, with several injec- tions of water to produce steam. The purified resins were stored in glass- stoppered bottles under methanol to maintain their high purity. In more recent years, polymeric resins for SPE are more pure and do not require the extensive treatment of the earlier materials. Still, blanks should always be run, especially for ultra trace work. 2.4 SAMPLING AND SAMPLE HANDLING Figure 2.1 is a scale drawing of a sample extraction device. Grab samples were taken of surface waters using clean 4-liter amber solvent bottles. After settling overnight, the clear water was decanted into the reservoir shown in the figure. The stopcock (E) was adjusted to deliver a flow rate of 25-50 m/min, When the water level reached the upper glass-wool plug, the sediment from the bottle was transferred to the reservoir using several rins with organic-free water. In this way organic solutes adhering loosely to the sediment would be washed into the column containing the extractive resin,24 SAMPLING AND SAMPLE HANDLING 19 FIGURE 2.1 Scale drawing of grab-sample extraction device: (A) 5-liter reser- voit, scaled down; (B) glass-wool plugs; (C) 24/40 ground-glass joint with PTFE sleeve; (D) 8 x 140-mm glass tube packed with ~2 g 40—60-1fesh resin; (£) PTFE stopcock, Proper handling of water samples is also vital to the success of the ‘method. Actual samples should be collected in a very clean glass bottle and analyzed without delay. During the passage of the sample through the resin column, the sample reservoir should be capped to prevent loss of some organics that tend to come to the water surface and be lost by evaporation. For example, 10 ppb (parts per billion) of cumene added to water samples showed only a 40% recovery with an uncapped reservoir but a 92% recovery ‘with a capped reservoir. Water samples containing even low concentrations, of saturated hydrocarbons present a special problem because some hydro- carbons cling tenaciously to the surfaces of the reservoir and column. It20 ‘SPE IN THE 19706: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER appears that this problem is alleviated by addition of either a detergent or a water-miscible organic solvent to the water sample. The need for clean containers for water samples must be emphasized, In ‘one experiment the recoveries of organic compounds in a standard sample were only about one half the usual recoveries. The difficulty was traced to a small grease spot inside the 4-liter sample bottle. On standing, some of the organic solutes were extracted from the water into the grease spot. Even a fingerprint can contain enough grease to take up some of the organic solutes, ‘The time at which a sample is taken and analyzed can affect the results. Ina study on the determination of haloforms in drinking water, the amounts, of CHCl; CHBrCl,, and CHBry were found to increase on storing the sample for several days. Continuing production of haloforms occurs while FIGURE 22 Scale drawing of composite sample extraction device: (A) standard garden hose coupling; (B) PTFE washer; (C) 12.7-mm-ID PTFE tubing; (D) zglass-wool plugs: (£) 12.7-mm-OD x 9-em-long glass tube packed with ~2-g 40-60-mesh resin.25 PREPARATION OF STANDARDS a chlorinated water is in the tion system, The increase in total halo- form concentration with time suggests that the water samples still contain certain organic matter that reacts with residual chlorine or bromine to produce additional halocarbons (8). ‘The apparatus shown in Figure 2.2 was used for composite sampling over a ~24-h period (7). The standard garden hose coupling was attached to a suitable faucet and the water flow adjusted to deliver ~150 ml/min, After 24 h the column was removed from the coupling, and PTFE sleeves were used to connect a 25-ml reservoir and a PTFE stopcock to appropriate ends of the column. The column was then eluted with diethyl ether, and the eluate was treated as just described for the grab samples. Bither sampling proce- dure may be employed dependent on whether one is interested in instanta- neous (grab) or average (composite) concentrations over an extended time period. The composite sampling device was initially connected to a faucet by a short length of garden hose. However, the use of even short lengths of polyvinylchloride tubing was found to introduce significant amounts of organic impurities. Some of these contaminants appeared to be plasticizers leached from the PVC tubing. This led to a study of the organic contami- nation from the different types of tubing used to contain or transfer various fluids (3), 2.5 PREPARATION OF STANDARDS Water samples containing known amounts of various organic analytes are needed to check out a concentration/analysis method and to determine the percentage recovery for the analytes. Since the solubility of many organic compounds is quite low, a technique must be used that will quickly dissolve the organic compound and give a homogeneous solution to sample. The best way to do this is first to prepare a standard solution of each organic test compound. This can be a solution of perhaps 150 ppm of the test compound in a solvent such as methanol, acetone, or ethyl acetate. Huge losses can ‘occur when hexane is the solvent, probably because of a floating layer and subsequent vaporization. A small but measured volume of this standard solution can be added to pure water just before adding the water sample to the SPE column, For example, 100 x! of standard solution (150 ppm) added to 15 mlof water produces a sample containing | ppm of the test compound. It is best to “squirt” the standard solution into the water as quickly as possible in order to provide some immediate mixing. With this technique22 ‘SPE IN THE 1970s: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER, any transient precipitate will consist of very small particles that will quickly dissolve as equilibration becomes more complete, The organic solvent for the 150-ppm standard solution should be selected with care, Any solvent that might undergo a chemical reaction with the analyte should be avoided, Fresh standard solutions should be prepared from time to time. 2.6 GENERAL PROCEDURE FOR SPE The procedure given below (6) is typical of those used for solid-phase extractions in the 1970s. Today, smaller, more efficient resins, small col- uumns, and different eluting solvents are used. However, the research of the 1970s provides a sound foundation for SPE as it is performed today. A comprehensive paper published in 1974 (6) has been identified as one of the 20 most frequently cited papers published in the Jounal of Chromatog- raphy (9). 2.6.1 Column Preparation ‘The apparatus used was similar to that in Figure 2.1 except for a 2-liter reservoir instead ofa 5-liter. With the upper 2-liter reservoir detached, insert a clean silanized glass-wool plug near the stopcock of the glass column. ‘Add the purified XAD resin as a methanol slurry until a resin bed approxi- mately 6 cm high is obtained (1.5~2.0 g of dry resin), then insert a second silanized glass—wool plug above the resin, Drain the methanol through the stopcock until the level just reaches the top of the resin bed, then wash the resin with three 20-ml portions of pure water. For each portion stop the flow when the liquid level reaches the top of the resin bed 2.6.2 Sample Preparation Attach the 2-liter reservoir to the column (Fig. 2.1) and add 1000 ml of Purified distilled or tap water. Add the organic compound(s) to be tested to the water by injecting a calibrated volume of a standard solution of the organic compound(s) in methanol. The volume injected and the concentra- tion of the standard is adjusted to achieve the desired amount of com- pound(s) in 1 liter of water.25 GENERAL PROCEDURE FOR SPE 23 2.6.3 Column Extraction Cap the reservoir with a one-hole stopper connected to a nitrogen source and allow the water to pass through the XAD resin column by gravity flow atarate of 30-50 ml/min, Ifthe flow rateis slower than this, apply a pressure ‘of about 1 psi (Ib/in.*) to the reservoir using organic-free nitrogen. When ‘most ofthe sample has passed through the column and the liquid level is at the top of the resin, wash the reservoir walls carefully with a 20-ml portion of pure water and drain through the column until the level reaches the top of the resin bed. Repeat this wash twice, letting the water drain completely only after the last wash. 2.6.4 Elution and Regeneration Wash the reservoir walls with two 10-mi portions of diethyl ether and allow each wash to drain into the XAD resin but not through the column, Remove the reservoir, cap the column with a 24/40 glass stopper, and allow the Giethyl ether to equilibrate with the resin for 10 min, Then remove the cap, open the stopcock, and allow the ether to flow through the column into a 30-mi test tube. Add an additional 5 ml of diethyl ether to the column and immediately allow it to flow through the resin into the receiver. The elution is complete when the gravity flow ceases even though the last traces of ether have not been removed from the resin. Regenerate the XAD resin column immediately after the ether elution, ‘Add methanol, shake to remove any air bubbles, then pass a total of30 ml ‘of methanol through the column. Close the stopcock, add an additional 15 ml of methanol, and cap the column with a 24/40 glass stopper. It is now ready for subsequent analyses without any further treatment beyond wetting the resin with pure water as outlined above in step 1 26.5 Drying Remove the residual water (0.5~2 ml) from the diethyl ether eluate by immersing the test tube receiver in liquid nitrogen for two 10-s intervals. Immediately decant the ether into the concentration vessel. Wash the ice in the test tube with I ml of diethyl ether, dip briefly in the liquid nitrogen to freeze any ice that may have melted and add this ether to the concentration vessel Alternatively, anhydrous sodium sulfate and petroleum ether (b.p. 30— 60) may be used to dry the ether eluate. Collect the ether eluate from the column in a 60-ml separatory funnel and reject the water layer. Add 15 ml24 ‘SPE INTHE 10706: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER of petroleum ether (30-60°) and 2 g of anhydrous sodium sulfate to the ether. Shake vigorously until a clear solution is obtained and transfer the ther to the concentration vessel, 2.6.6 Concentration of Eluate ‘Add a small boiling chip to the concentration vessel and attach a three-cavity Snyder column, Add about 2 ml of diethyl ether to the top of the Snyder column and tap gently to distribute the ether into the three cavities. Apply heat from a hot plate or steam bath so that the boiling action is vigorous enough to agitate the balls of the Snyder column continuously. A solvent ‘evaporation rate of 0.52.0 ml/min should be attained. When the volume of solution in the calibrated appendage of the concentration vessel is about 0.5 ml, remove the apparatus from the heat and immediately spray acetone over the outside walls of the concentration vessel. The condensation of the ether vapor causes an automatic sequential washing of the inside walls of the vessel with the ether held in the three cavities of the Snyder column. The volume of liquid in the calibrated section of the concentration vessel should now be 1.0 ml. Remove the Snyder column and add ether if necessary so that the solution volume is exactly 1,00 ml. Cap the vessel with a 14/20 stopper and swirl to mix the solution, Proceed with the GC analysis of the concentrate as soon as possible. The shape of the vessel used to concentrate the eluate is critical. Vessels of several shapes were tested (Fig. 2.3). The calibrated scale at the bottom is for measurement of the volume of concentrated eluate. A vessel of shape 0.20mi (—1.00m1 '—1.00m FIGURE 2.3 Drawings of the concentration vessels: A is recommended, B is unsatisfactory, and C is questionable, (From Ref. 6 with permission.)26 GENERAL PROCEDURE FOR SPE 25 ‘A was found to give the best results. Vessels of shape B or C were found to cause losses of 10% to 60% during the evaporation step. 2.6.7 Separation and Quantification Inject a suitable aliquot of the 1,00-ml concentrate into the gas chroma- tograph with a syringe. When the GC separation of the organic compound(s) inthe concentrate is completed, immediately repeat the GC separation using ‘a 2.0411 aliquot of a 1.00-ml standard ether solution containing the same ‘organic compound(s) at a concentration identical to that expected in the 1,00-ml concentrate assuming complete recovery of solute(s) from the water sample. The GC conditions are held rigidly constant for both sample and standard during these tests, and the percentage recovery of the organic is calculated directly from a comparison of the chromatogram peak 2.6.8 Identification Further concentrate the 1.00-ml ether solution to 0.1 ml after completion of step 7. Subject a 2.0-11l aliquot of this 0.1-ml solution to GC-MS analysis to confirm positively that no chemical transformation has occurred during any of the previous steps. TABLE 2.3 Average Recoveries of Test Compounds by Chemical Groups Using the “Porous Polymer” Extraction Method. ‘Number of Group Compounds Recovery, % ‘Acids s 101 Alcohols 8 95 Aldehydes 7 95 Alkylbenzenes 3 89 Esters 15 93 Ethers 5 90 Halogen compounds 10 87 Nitrogen compounds 10 89 Pesticides 6 83 Phenols 7 82 Polynuclear aromaties 8 88 Source: Data are summarized from Reference 6.26 ‘SPE IN THE 1970s: EXTRACTION OF ORGANIC POLLUTANTS FROM WATER. 2.7 RECOVERY DATA The recovery efficiency of the entire analytical scheme known as the porous polymer method was studied extensively and the results reported in a comprehensive paper published in 1974 (6). Organic test compounds were added to distilled water, and in some cases to tap water, in concentrations ranging from 10 to 100 ppb. The test compounds included acids, alcohols, aldehydes, alkylbenzenes, aromatic halides, esters, ethers, ketones, phenols. TABLE 2.4 Recoveries of Some Individual Compounds by Porous Polymer Extracti ‘Test Compound Acids Benzoie Oleic Alcohols, Halogen compounds Hexyl 93 Benzyl chloride 88 Decyl 91 Chlorobenzene 95 Benzyl an 1,2-Dichlorotoluene 96 Cinnamyl 85 _-2,4-Dichlorotoluene n Aldehydes Nitrogen compounds Benzaldehyde 101 Nitrobenzene 91 Salicylaldehyde 100 Indole 89 Quinoline 84 Alkyl benzenes Pesticides Ethylbenzene 81 Atrazine 3 Cumene 93 Lindan 95 Esters Phenols Dimethylphthalate 91 Phenol 40 Dibutyiphthalate 92 o-Cresol B Methylbenzoate 101 _-2-Chlorophenol 96 ‘Methyloctanoate 98 2,4.6-Trichlorophenol 9 ‘Methylpalmitate 70 Polynuclear aromatics Naphthalene 98 Anthracene 33 Acenaphthene 92. Source: This is an abbreviated lis fom Reference 6REFERENCES 2 and chlorinated phenols, pesticides and herbicides, polynuclear aromatics, and various compounds containing halogens, nitrogen, or sulfur. Aqueous samples containing acidic solutes were acidified with 5 mU/iter of concen- trated hydrochloric acid to ensure that the compounds were in their molecu- lar form. ‘A summary of the average recoveries obtained for 84 individual com- pounds are summarized according to compound class in Table 2.3. The average recovery for the 84 compounds studied was 90%. The recoveries of an abbreviated list of individual compounds are shown in Table 2.4. These results show the effect that chemical structure can have oon the recoveries. For example, phenol itself tends to have a significantly lower recovery than phenols that contain alkyl or halogen groups. REFERENCES 1. A.K, Bumham, G. V. Calder, J.S. Fritz, G. A. Junk, H.J. Svee, and R. Vick, JL Am. Waterworks Assoc. 65 (1973) 722. ‘A.K. Bumham, G. V. Calder, J.S. Fritz, G. A. Junk, HJ. Svec, and R. Willis, Anal. Chem, 44 (1972) 139. 3. G.A.Junk,H.J.Svec, R.D. Vick, andM. J. Avery, Em: Sci. Technol. 8 (1974) 100. 4. C.D. Chriswell, R. L, Ericson, G. A. Junk, K. W. Lee, J. S. Fritz, and H. J Svec, J. Am, Waterworks Assoc. 69 (1977) 669, 5. G.A.Junk, C.D. Chriswell, R. C. Chang, L. D. Kissinger, J.J. Richard, J. Fritz, and H. J. Svec, Z, Anal. Chem, 282 (1976) 331 6. GA, Junk, J. J. Richard, M. D. Grieser, D. Witiak, J. L. Witiak, M. D. Arguello, R. D. Vick, H. J. Svec, J.S. Fritz, and G. V. Calder, J. Chromatogr 199 (1974) 745. 7. G.A Junk, J.J, Richard, H. J. Svec, andJ. 8. Fritz, J. Am. Waterworks Assoc. 68 (1976) 216. 8. L.D, Kissinger and J. S. Fritz, J. Am. Watenvorks Assoc. 68 (1976) 435. 9. ILS. Fritz and G, A. Junk, J. Chromatogr. 625 (1992) 87.CHAPTER 3 SOLID PARTICLES FOR SOLID-PHASE EXTRACTION OF ORGANIC COMPOUNDS FROM WATER 3.1. INTRODUCTION A large number of solid particles have been used for SPE of organic compounds from predominately aqueous samples. Bonded-phase silica particles, especially ODS silica (octadecylsilane silica), are currently the most popullar type. However, a wide variety of porous polymeric resins have also been used. Activated carbon and more recently graphitized carbon materials constitute another class of solid particles for SPE. Before discussing the actual particles used in solid-phase extraction of organic compounds, let us review the various types of SPE available. Classification of the types of SPE available for organic analytes is the same as for HPLC, although the end goal of SPE is different from HPLC (see Chapter 1). 1. Reversed-Phase SPE. Here the goal is to isolate relatively nonpolar analytes from a polar sample such as water. This type of application requires the use of relatively hydrophobic adsorbent particles, such as silica with bonded octadecylsillane groups or an organic polymer82 DESIRABLE PROPERTIES OF SPE PARTICLES 29 with benzene rings. The extracted substances are eluted by a small volume of an organic solvent. 2. Normal-Phase SPE. This technique is used to isolate polar com- pounds from a nonpolar sample matrix. A vegetable oil dissolved in hexane is an example of such a sample matrix. Polar solid particles are then used to extract polar analytes from the sample. The extracted analytes are finally eluted by a polar solvent. Ion-Exchange SPE. Particles containing cation-exchange groups or anion-exchange groups are used to extract ionic analytes or analytes that can be converted to ionic form by adjusting the sample pH. After the SPE step, the extracted substances are eluted with an organic solvent after converting them back to the molecular form, Alterna- tively, the extracted ions may be removed from the ion-exchange sites by elution with a solvent containing a relatively high concen- tration of a displacing ion. SPE with ion-exchange particles is discussed in Chapter 5. In this chapter the desirable properties of SPE materials will be listed, followed by a discussion of the many types of particles that are used in solid-phase extraction. 3.2 DESIRABLE PROPERTIES OF SPE PARTICLES 3.2.1 Porous, Large Surface Area In SPE the uptake of a sample solute depends on an equilibrium between the sample solution and the solid SPE particle. This equilibrium is shifted ‘more strongly toward the solid as the surface area becomes larger. A surface area >100 m?/g is generally necessary for effective SPE. The particles most widely used today have a surface area between about 200 and 800 m*/g. ‘The pore size of a particle and its surface area are generally linked in an inverse relationship. The surface area decreases as the average pore size increases. This relationship, together with the ability of a series of polymeric adsorbents to extract a copper(Il)-dithiocarbamate complex from water is illustrated by the data in Table 3.1 (1). In both the polystyrene resins and the polyacrylate resins the extractability of the copper organic complex increases with surface area.30 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER TABLE 3.1 Physical Characteristics and Breakthrough Capacities of XAD Resins for CopperlI-Di(hydroxyethyl)dithiocarbamate Surface Area, Capacity, mm Adsorbent Type mig PoreSizeA _ Culg XAD-1 Polysiyrene/DVB 100 100 0.027 XAD2 Polystyrene/DVB 300 90 0132 XAD4 Polystyrene/DVB_ 725 40 oa XAD-7 _ Polyacrylate 450 90 0.239 XAD-8 _Polyacrylate 160 225 0.132 3.2.2. Reversible Adsorption Successful SPE has two major requirements: (1) a high, reproducible percentage of the analytical solutes must be taken up by the solid extractant; and (2) the solutes must then be easily and completely eluted from the solid particles. Carbon was the first medium to be used for the extraction of ‘organic compounds from water (2). Carbon has the advantage of very high surface area and consequently high uptake of organic solutes. However, the heterogeneous nature of the activated carbons used for SPE has caused problems such as irreversible sorption, widely differing affinities of differ- tent classes of compounds, and catalytic activity of the surface, which can lead to unwanted chemical reactions. Other widely used SPE materials such as bonded-phase silicas and porous polymers generally pose no problems with regard to reversible adsorption. A number of organic solvents are available that can effectively elute adsorbed solutes from the solid particles. 3.2.3. Pure, Low Leachable Impurities The solid particles need to be as free as possible of impurities that might be leached from the solid during elution of the retained sample constituents. If purification of the particles is needed, it often can be accomplished by ‘washing with an organic solvent or a succession of solvents, ‘The XAD resins used in the 1970s and 1980s posed a rather severe purification problem. Ethylbenzene, benzoic acid, and other impurities were trapped within the resin during the polymerization step. At that time, crushing and sieving was necessary to reduce the materials to a particle size suitable for solid-phase extraction, This process opened up additional surfaces from which the resin impurities continued to be leached. Fairly92 DESIRABLE PROPERTIES OF SPE PARTICLES a1 extensive Soxhlet extraction with several organic solvents was needed to reduce the leachable impurities to an acceptable level. Another purification ‘method was to heat a tube containing resin in the oven of a gas chroma tograph and to steam-clean by injecting several small portions of water. Impurities can come from sources other than the particles used in SPE. Junk, Avery, and Richard (3,4) found that various compounds can be leached from the polypropylene housing often used in commercial SPE cartridges and also from the frit that holds the particles in place, Table 3.2 lists compounds leached from the cartridge housing, frit, and the C18-bonded porous silica, The impurities leached from commercial cartridges varies TABLE 32 Compounds Identified in Extracts of Components of Solid-Phase Extraction Cartridges ‘Column Component Polypropylene Polyethylene C18-Bonded ‘Compound Frit Frit Porous Sil C8 alkene = = x 9 alkene = = x C10-C16 alkenes x - x C17-C28 alkenes x x = C8-C26 alkanes - = Naphthalene = _ Biphenyl x ‘Acenaphthene x 2,6-Di-tert-butyl-p-cresol x = 2.6-Di-tert-butyl-p-quinone x Phenol x Methyloctadecanoate x Dimethyloctadecylsilanol = - x Decamethylpentasiloxane - - x DodecamethyIhexasiloxane = x ‘Tetradecamethylheptasiloxane = — = x Diethylphthalate - = x Dibutylphthalate x x x Bis(2-ethylhexy)phthalate x x x Bis(2-ethylhexyladipate “The dash (—) means not positively identified in extract, Source: From Reference 3 with permission.32 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER with the eluting solvent and often with the lot number ofthe cartridge (Table 3.3). 3.2.4 Good Chemical Stability An acidic or basic sample solution or eluting solution is needed in some applications. Silica-base adsorbents are not very stable above pH 8 or in highly acidic solutions. However, most polymeric resins are quite stable in both basic and acidic media. Instead of eluting adsorbed sample analytes with an organic solvent, it is often convenient to use thermal desorption. The SPE minicolumn can be placed in a convenient heater that is attached to a gas chromatograph and the analytes thermally desorbed directly into the chromatograph. This mode of operation requires the use of solid particles with good thermal stability. Solid particles need to be stable in the organie solvents used in the elution step as well as in the aqueous sample. Polymeric resins should be suffi- TABLE 33 Compounds Fluted from Commercial C18 Cartridges Eluting Solvent Compounds Bihyl Acetate Benzene C10 alkane C19 alkane €20 alkane €22-C18 alkanes C15-C17 alkenes, C18-C26 alkenes Dimethyloctadecylsilanol Methylhexadecanoate Hexadecanoic acid Methyloctadecanoate = 2,6-Di-tert-butyl-p-cresol = Naphthalene Diethylphthalate = Dibutylphthalate - Bis(2-cthylhexyD)phthalate = Bis(2-ethylhexyDadipate - “The dash means not positively identified in eluant. Source: From Reference 3 with permission.82. DESIRABLE PROPERTIES OF SPE PARTICLES 33 ciently crosslinked to ensure that they do not dissolve or soften in organic solvents. The resins must not undergo large volume changes due to swelling and shrinking in contact with different liquids. 3.2.5 Good Surface Contact with Sample Solution Chemically bonded silica, usually with a C18 or C8 organic group is the most used material for SPE. However, the use of crosslinked polystyrene and other porous polymeric resins is increasing. Polymeric resins are more rugged and pH-stable, and they have a greater surface area than do most. silica-based materials. Research has consistently shown that the percentage recovery for many types of analytes is significantly higher with polymeric resins than with silica particles (5). Chemically bonded silica and porous polystyrene resins have several shortcomings for use in SPE (6). While silica itself is hydrophilic, the hydrocarbon chains make the surface hydrophobic. The consequence is poor surface contact with predominantly aqueous solutions, Porous poly- styrene resins also have a hydrophobic surface (7). Pretreatment of the SPE. ‘materials with an activating solvent (such as methanol, acetone, or acetoni- trile) must be used to obtain better surface contact with the aqueous solution being extracted. However, the activating solvent can be gradually leached out of the resin, thereby causing the extraction to become ineffective. This, is particularly true if the SPE column inadvertently becomes dry, causing air to be sucked into the column (8). A better approach is to make the surface of a solid-phase extractant permanently hydrophilic through a chemical reaction. Sun and Fritz (9,10) introduced an acetyl, hydroxymethyl, or cyanomethyl group into crosslinked polystyrene resins at a capacity of approximately 1 mmol/g. Whereas dry, untreated resins float and clump together on the surface of ‘water, the modified resins have a more hydrophilic surface and are easily wetted by water. The derivatized resins can be used in SPE without any pretreatment with an “activating” solvent such as methanol. Mild sulfonation of resins also increases their hydrophilicity and makes them wettable with water alone (11). The best performance for SPE was found to be with resins containing around 0.6 meq/g of sulfonate groups. ‘The degree of sulfonation should be sufficient to render the resin surface hydrophilic, but not enough to compete with the more hydrophobic resin matrix, which is responsible for extraction of the organic analytes from water.34 EXTRACTION OF ORGANIC COMPOUNDS FROM WATER 3.2.6 High-Percentage Recoveries Quantitative analyses involving SPE depends on recovery of a known, fixed percentage of each analyte during the extraction and elution steps. Ideally, quantification will be based on essentially complete (~100%) recovery. Using known standards, quantification based on a lower percentage recov- ery is feasible, but precision and accuracy are better when the overall percentage recovery is as high as possible. 3.3. BONDED-PHASE SILICA PARTICLES. 3.3.1. Types Available At this writing bonded-phase (BP) silica materials are the dominant part. cles used in SPE. One reason for this is their widespread availability. When the advantages of SPE as a sample pretreatment technique became apparent and the use of SPE began to increase, commercial suppliers made products available to meet the new demand. Since the technology of bonded-phase silica packings for HPLC columns was already well developed, it was a relatively easy task to offer similar materials in a form suitable for SPE. As the use of BP silica materials for SPE increased, several supply houses offered bibliographies and other literature to users. It then became possible for chemists to use this information as a guide for specific analytical problems they might encounter. Several supply houses offer a full line of silica particles for SPE. The ‘major types are listed in Table 3.4. The materials in this table are listed in their approximate order (top to bottom) of increasing polarity. The desired ‘major functional group is introduced into porous silica particles by reaction of silanol groups on the silica with chloro- or methoxyorganosilane, For example, the C8-bonded phase material has the following structure: CH | Silica —OH + C+SI—CH;CH:CH,CH;CH;CH,CH,CH, | > CH cH I Siliea-O-Si—CH,CH,CH,CH,CH,CH,CH;CH, (3.1) far33. BONDED-PHASE SILICA PARTICLES. 35 TABLE 3.4 Bonded-Phase Silica-Modified Materials Used in Solid-Phase Extraction Phase Polarity of Phase Designation ‘Octadecyl, endcapped ‘Strongly apolar CBee Octadecyl Strongly apolar cis ety! Apolar cs Ethyl Slightly polar C2 Cyclohexyl Slightly polar cH Phenyl Slightly polar PH Cyanopropyl Polar cN Diol Polar 20H Silica gel Polar SiOH Carboxymethyl ‘Weak cation exchanger CBA ‘Aminopropyl Weak anion exchanger NH, Propylbenzene sulfonic acid Strong cation exchanger SCX ‘Trimethylaminopropyl Strong anion exchanger SAX ‘The alkali chain in the octadecylsilane (ODS) material is even longer (18 carbon atoms in the alkyl chain). These particles may be visualized as porous solids with a high surface area and the alkylsilane sticking out from the various surfaces like tall trees. Actually the long alkyl chains have a kind of zigzag structure. It will be scen that longer alkyl chains make the silica, particles more hydrophobic and also provide a stronger overlap with the hydrophobic parts of extracted analyte molecules. For this reason, extrac~ uibility generally increases with chain length. The bulk of the chloro alkyl silanes used in the derivatization reaction prevents their reaction with all of the SiOH groups on the silica. These free silanol groups can attract polar analytes that might not otherwise be ex- tracted by hydrogen bonding. Alcohols and amines are exampl36 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER ‘To avoid this situation, a process known as endeapping is often used. The remaining silanol groups are endcapped by reaction with a smaller, more reactive organosilane such as chlorotrimethylsitane, Silica-OH + Cl Si(CH,), > Silica~O-Si(CH,), 2) Octadecylsilane silica, often abbreviated as ODS or C18, is by far the most widely used bonded-phase silica for reversed-phase SPE. Octylsilane (C8), ethylsilane (C2), and occasionally cyclohexylsilane (CH) also find some application, especially for SPE of more bulky analytes. Phenyl-sub- stituted silica may be used for reversed-phase extraction of somewhat more polar analytes. Cyanopropyl silica and silica with a diol bonded phase are generally too polar to be used in the reversed-phase mode. Instead, these materials are often used for SPE in the “normal-phase” extraction mode. Silica gel and several other inorganic particles are also used for normal- phase SPE (see Section 3.8). The last four BP silicas listed in Table 3.3 are used in ion-exchange SPE (Chapter 5). Bonded-phase silicas are chemically stable in the presence of most organic solvents. The average pore size of approximately 60 A makes it possible to extract organic compounds of up to ~15,000 molecular weight. Sample solutions much above pH 8 cause some hydrolysis of BP silica particles, 3.3.2 Cartridges SPE with BP silica particles may be carried out in pre-prepared cartridges, in prepacked tubes of varying dimensions, or in various devices in which the silica incorporated in a disk (Chapter 6). The basic design of cartridges has changed little since their introduction in the late 1970s. A typical cartridge consists of a polyethylene body with a female Iuer tip at the top for attachment to a positive-pressure source and a male luer tip atthe bottom. ‘The packing material is held in place by a 20-m polyethylene frit at each end. The particle size of the packing material varies, but typically averages around 40 or 50 jim in diameter. The dimensions of the sorbent bed are small ‘enough to permit easy flow of the sample through the cartridge, either by zravity or by suction-aided flow. The small dimensions also minimize the volume of organic solvent needed to condition the column and elute the sorbed analytes. Properties of two batches of a popular SPE cartridge are summarized in Table 3.5. The C18 sorbent has small pores and a high surface area. It has a heavy loading of surface-bonded C18 groups. The particle size range is,83 BONDED-PHASE SILICA PARTICLES. 37 TABLE 35 Sorbent Properties of C18 Silica Extraction Cartridges as Determined by HPLC 1. Baker Octadecylsiloxane Sorbent Parameter Column 1 ‘Column 2 Dimensions 25cm 25cm 2.1 mm ID 4.6mm ID Lot number F37502 G10508 ‘Weight of packing (g) 0.570 3.3793 Column volume (ml) 0.785 4.155 Packing density (g/ml) 0.726 0813 Total porosity 0.49-0.51 047 Interparticle porosity 0.45-0.42 oat Intraparticle porosity 0.06-0.07 0.06 Apparent average particle diameter (jum) 55.1 56.5 Source: Data from Reference 12. ‘much larger than that used in HPLC, and the distribution is skewed toward smaller particles. Miller and Poole used HPLC to study the properties of typical octadecyl- siloxane silica cartridges used for SPE (12). The characteristics of the HPLC columns used in their study are summarized in Table 3.5. The total porosity of the column includes the Tiquid between the silica particles (interparticle porosity) plus the liquid within the pores of the particles. The latter is called the intraparticle porosity. The total porosity (E) is measured by injection of a nonretained chemical marker that can freely enter the pores within the particles. Fulm G3) mre where F, is the volume flow rate, fy is the time for the marker to elute, ris the column radius, and L is the length of the resin bed. The intercolumn porosity is calculated using dilute sodium nitrate as the marker because it is excluded from the intraparticle pores by ion exclusion. The intraparticle porosity is the difference between the total porosity and the interparticle porosity. It will be seen from Table 3.5 that the intraparticle porosity of the CIB silica is quite small. This may be the result of plugging many of the small pores due to the heavy loading of bonded octadecylsiloxane groups.38 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER These results indicate that the material does not have an accessible internal pore structure, The average packing densities of several SPE cartridges were found to be 0.667 g/ml (n = 5, s = 0.014), with a range of 0.653-0.683 g/ml. These packing densities are considerably smaller than those found for C18 HPLC ‘columns in Table 3.5. This fact, together with the rather large spread in packing densities, suggests that the cartridge beds are not very stable. This, heterogeneous packing structure of the cartridges leads to channeling and reduces their effectiveness for SPE. 3.3.3 SPE Tubes Prepacked SPE tubes are very efficient and easy to use. They seem destined to replace the older cartridges. A typical SPE tube consists of a syringe barrel packed with 40-50-14m sorbent material. There is a male luer tip at the bottom and the packing is held in place by polypropylene frits. SPE tubes are fairly low in cost and are disposable. A wide variety of SPE tubes is available from chromatographic supply houses. For example, tubes packed with C18 silica are available in the following sizes: 1, 3, 6, 8, 12, 20, and 60 ml. The weight of sorbent in these tubes is 0.1, 0.5, 0.5, 1.0, 2.0, 5.0, and 10 g, respectively. Tubes and other devices for SPE are described ‘more fully in Chapter 4, The availability of this equipment means that the scale of SPE (sample size, volume of eluting solvent) can be varied over a broad range. Although they are pure, are convenient to use, and generally give accept- able results, bonded-phase silica materials do have several drawbacks. The need to pretreat the tube or cartridge with a solvent such as methanol was pointed out in Section 3.1.5. The recommended pH range for use of BP silica materials is generally specified as pH 2-8. Especially at highly alkaline pH values, the amount of organic coating is reduced by hydrolysis. This causes the particles to lose part of their extractive power and may result in unexpectedly low recoveries of sample compounds. The overall recoveries (extraction + desorption) obtained with BP silica particles varies considerably, Recoveries of some analytes are close to 100%, while others have much lower recoveries. Published procedures often involve the use of internal standards for quantification and fail to give the actual recoveries of the various analytes. The advantages of high-per- centage recoveries were pointed out in Section 3.2.6. The main point here is that BP silicas often give significantly lower overall recoveries than do other SPE materials, particularly polymeric sorbents184 ORGANIC POLYMERIC ADSORBENTS. 39 3.4 ORGANIC POLYMERIC ADSORBENTS Porous organic polymers have a number of advantages over activated carbons and bonded-phase silicas for solid-phase extraction. Unlike bonded-phase silicas, polymeric organic particles can be used at virtually any pH, and they contain no troublesome silanol groups. The surface area is generally higher than that of silica particles, and uptake of organic analytes thus tends to be more complete. In most cases adsorbed analytes are easily and completely eluted from polymeric adsorbents by a small volume of an organic solvent. The use of polymeric adsorbents for extraction has been reviewed (13). ‘The Porapak (Waters) and Chromasorb (Johns-Manville) adsorbents listed in this review have been available for some years, but they are infrequently mentioned in the more recent literature on SPE. Table 3.6 lists polymeric adsorbents that are finding extensive use in current practice. The XAD. materials, developed by Rohm and Haas and now marketed by Supelco and other supply houses, have had a major impact on SPE. Theit properties are given in Table 3.1, and some of their earlier uses are described in Chapter 2. Although these are excellent sorbents, in many cases they continue to be available only in a 20-60 mesh particle size range (approximately 400-250 im). This is much too large for efficient solid-phase extraction and means that additional grinding and sizing will be required. Even when the original particles have been purified, grinding opens up new surfaces from which ‘organic impurities can be leached over along period of time. However, some newer versions of these resins (CG sorbents) are listed in Table 3.6 that have ‘been purified and are available in a more suitable particle size range. Several companies produce porous crosslinked polystyrene—divinylben- zene (PS-DVB) resins of smaller particle size that are primarily intended for liquid chromatographic columns. These resins are also useful for SPE. ‘The PRP-1 and PRP-3 resins from Hamilton, PLRP-S from Polymer Labs, and MP-DVB from Interaction are examples of this type. However, chro- matographic columns packed with 5—10-um resins have a substantial backpressure. A SPE column packed with particles of this small size is feasible if the resin bed height is only a few millimeters. Careful packing and design of such a column is essential to avoid channeling through the very short resin bed. Several sorbents with special properties are listed in Table 3.6. MP-1 (Interaction) contains C18 alkyl functional groups, which makes this mate- rial more hydrophobic than ordinary crosslinked polystyrene polymers. At the other end of the adsorption spectrum, Oasis (Waters) is easily wettable40 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER, Polymeric Sorbents for SPE. Surface Area, Pore Size, mig AC Patt Supplier/Sorbent ‘Type Hamilton PRP-I PS-DVB 100 10um PRP-3 PS-DVB = 300 10um PRP-infinity PS-DVB — — Nonporous 104 Interaction MP-1 PS-DVB+C18 = = 20-60 pp groups MP2 Polyvinyl pyridine = — = 20-60 um MP-DVB PS-DVB = — 20-60 um Lab instruments Ostion SP. PS-DVB 350 85 = Synachrom _PS-DVB-EVB 520-620 90 = Spheron MD Methacrylate-DVB 320 = = Polymer Labs PLRP-S PS-DVB = = 10 um Supelco XAD-2 PS-DVB 300 90 20-60 mesh CG-161 PS-DVB 900 150 80-160 um Polymethacrylate 500 250 80-160 1m Polymethacrylate 160 225 40-60 mesh PS-DVB 700 300 20-50 um PS-DVB 700 300 50-100 pm PS-DVB 250 1000 20-50 um Tessek (Czeck Rep.) Seperon HEMA Hydroxyethyl- 20-60 . ‘methacrylate Waters Oasis Poly-DVB- 830 82 _ vinylpyrrolidone bby water alone and is able to take up many hydrophilic compounds effi- ciently. Oasis may be used for serum and other biological samples without any sample pretreatment. For the most part crosslinked polystyrene resins have been the most successful for solid-phase extraction. Polyacrylates such as Rohm and Haas XAD-7 and XAD-8 are more polar than polystyrene particles and have been84. ORGANIC POLYMERIC ADSORRENTS a found to be superior for uptake of polar compounds such as fulvic acid (14) or phenol (15). Spheron SE (16) and Separon SE (17) have been found suitable for preconcentration of toluene, cresols, phenoxycarboxylic acids, and S-triazi- nes, The ability of Separon HEMA (hydroxyethylmethacrylate) to retain solutes would appear to be limited by its rather low surface area (20-60 m/s). Tenax GC, poly(2,6-diphenyl-p-phenyleneoxide), is a very popular ma- terial for determining organic analytes in their vapor form in purge-and-trap procedures (18-22). In this method air or another gas is bubbled through an aqueous sample causing volatile compounds to be vaporized. The vapors are then trapped on a short column of Tenax resin. The success of this ‘method lies in the fact that Tenax retains organic vapors well but has almost no affinity for water vapor. The unusually high temperature stability of ‘Tenax (up to 380°C) and high purity permit adsorbed compounds to be cluted simply by heating the column, Often the adsorbed analytes are desorbed thermally directly into a gas chromatograph for analysis. Tenax has also found some use for accumulating organic solutes from aqueous samples (23-26). Preconcentration efficiency has been found to be best for nonpolar compounds with low solubility in water such as polychlorinated biphenyls (27), polycyclic aromatic hydrocarbons (26,28), and various pesticides (26,28). However, the low surface area of Tenax 20 m/g can lead to low recoveries of analytes from water (24). Another drawback of Tenax is its low resistance to organic solvents that might be used to elute adsorbed analytes. Some organic solvents will actually dis- solve the Tenax. 3.4.1 Modified Polymeric Sorbents Bonded-phase silica materials are readily available with a variety of hydro- phobic (C18, C8, C2, etc.) and hydrophilic substituents (cyano, diol, amino, etc.). The sorptive properties of crosslinked polystyrene resins have also been modified by the introduction of various functional groups (29). The modified resins are easily prepared by a Friedel-Crafts reaction with the benzene ring of the polymer. Examples are (29) -CH,OH Derivatives. React with formaldehyde, acetic acid, acetic anhydride, and anhydrous zinc chloride to form ~CH,OCOCH,, then hydrolyze with methanol-hydrochloric acid42 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER. -COCHCH;CO;H Derivatives. React with succinic anhydride and aluminum chloride in tetrachloroethane—nitrobenzene. ~COCH, Derivatives, React with acety] chloride and aluminum chloride inCS,. -C(CH,); Derivatives. React with tert-butyl chloride and aluminum chloride in nitrobenzene solution. ~CH,CH Derivatives, React with chloroacetonitrile and aluminum chlo- ride in tetrachloroethane. These resins have been used for HPLC as well as for SPE. Capacity factors for several organic analytes are compared in Table 3.7 with 50% aqueous acetonitrile as the mobile phase. Introduction of nonpolar tert-butyl groups er capacity factors for the more hydrophobic analytes (cumene, toluene, etc.). However, the capacity factors for these analytes are all lower on the resins substituted with polar groups (-CH,OH, ~COCH,, etc.). The capacity factors for the most polar analytes (p-cresol and phenol) are higher on these resins, particularly on resins containing an acetyl group. The effect of various functional groups on capacity factor can be seen from the data on Table 3.8, where the ratio of capacity factors of the substituted benzene to benzene itself are compared on three PS-DVB resins: underivatized, acetyl, and tert-butyl. The ratio (R) of the capacity factor on the substituted resin to that on the unsubstituted resin is also given for each analyte. Solid-phase extraction of organic analytes is usually carried out with primarily aqueous samples. The capacity factors given above for 50% acetonitrile are useful in predicting the behavior of analytes with various resins in aqueous solution. The capacity factors increase considerably as the percentage of acetonitrile in the aqueous solution is decreased, but the order of & values of the various analytes tends to remain the same. In the middle range of organic-water solvent composition (~20-80% organic), log k increases nearly linearly with decreasing percentage of organic in the liquid phase. The increase of log k with decreasing organic content often becomes steeper as the liquid composition approaches that of pure water. This effect is shown in Figure 3.1 (PS-DVB) and Figure 3.2 (acetyl PS-DVB). It will be noted that capacity factors of the more polar analytes approach a higher value with the acetyl resin. Some of the k values were too high to measure in water alone. For comparison a plot of log k versus percent ACN with C18 silica particles is given in Figure 3.3. In this case the change in log k between 0% and 10% acetonitrile is much less than with the two PS-DVB materials. The‘worssquuad yp 6 20U3!2}>y Wass -20uN0S 1H Wes LL He on pars (05-05) s1em-apgntaor2oesutunjoo uu = 1°Z x9S7 sHoNIpUOD a1ydesTOEULO “Hse y "HOOD-GAG-Sd sanan93 UE alo sant, uo saapeatiag suauog, 10} (ouszuag) 7 01 aanepy OPEY APEdeD pue (Z) siopEE AIPLME FE ATAVL"Y IDE tm payproe (os-95) Jonen-aqunuorsDe sea TURNID SL "US pazHeALapUn 2xp or pareasoo zo ro ero wo 0 x0 £60 960 060 680 60 160 901 60 vl coyssqunad yp § 20u2s2}29 Woy :29um0g 2usnuag & Jo ore Kiedeo aun PZTBALDp 249 UO JO ONE Sty oro oro. so 8r0 Ico «0 sro 090 80 sso s60 Ico seo sor svt ocr 66'S zo sso wi ue ero eso 670 Iro ero iro ero S80 360 LO 1 £60 ssz cr 6 wz 961 soz 80% oy ist urs 90 8L0 el eT 6s ot oo os ws ars 66'S TouogTe |Kzuog, low stozuag Touayg jouaydontn-+ ‘suazuagoronygonnurg-p'Z auazuaqoniy, swrozueq Kuno ‘uzzttoqor0n|}46 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER. 2500 2000 100 fa 1000 500 ° 10 20 30 40 Percent Acetonitrile = Phenol + pthylphenol + Acetophenone —#- Toluene FIGURE 3.1 Capacity factors on PS-DVB resin as a function of acetonitrile percentage in the organic~aqueous mobile phase values of log k are also significantly lower with the silica sorbent for each of the analytes. This behavior may help explain the observation that in SPE the percentage recoveries of analytes often tend to be higher with PS-DVB. resins than with C18 silica, In SPE intimate interfacial surface contact between the aqueous sample and the resin is essential for efficient extraction of organic analytes. The bulk of the solid particles must be hydrophobic to extract the analytes effectively, but the solid particle surfaces need to be somewhat hydrophilic for proper interfacial contact. Octadecylsiloxane and PS-DVB sorbents both, present some problems in that their surfaces are too hydrophobic to be effectively wetted by water alone. The usual answer to this problem is to pretreat the SPE column with a water-miscible organic solvent such as34 ORGANIC POLYMERIC ADSORBENTS ar 2500 2000 1500 x 1000 500 ° ° 10 20 30 40 Percent Acetonitrile © Phenol + peEthyiphenol “te Acetophenone -® Toluene FIGURE 3.2 Capacity factors on acetyl PS-DVB resin ‘methanol. Some of the methanol is adsorbed, thus giving a more wettable surface. But the organic solvent can be gradually lost from the sorbent, particularly if the liquid level falls below the top of the resin bed, thus allowing air to be sucked into the column. A better answer to the surface contact problem is to use a PS-DVB resin with hydrophilic groups on its surfaces. Sun and Fritz found acetyl- and hydroxymethyl-substituted PS-DVB to be the most effective in this regard (5). They compared C18 silica, underivatized Amberchrome (a PS-DVB resin from Rohm and Haas), hydroxymethol Amberchrome, and acetyl Amberchrome for SPE of a number of organic analytes. In each case the extractions were done both with and without the usual methanol pretreat- ment, The results in Table 3.9 show higher average recoveries of the test compounds with the derivatized for the “wet” runs (methanol pretreatment), All the PS-DVB resins gave higher average recoveries than did the C1848 [EXTRACTION OF ORGANIC COMPOUNDS FROM WATER. 300 250 200 3 150 100 50 ° 10 20 30 40 Percent Acetonitrile “= Phenol + p-Ethylphenol “+ Acetophenone -® Toluene FIGURE 3.3 Capacity factors on a ODS silica column. silica materials, For the “dry” runs (no methanol pretreatment) the differ~ ences were even greater. Here, recoveries were unacceptably low except for the two derivatized resins where the average recoveries were only about 10% lower than those obtained with methanol pretreatment, Other results reported by Sun and Fritz (5) and by continuing studies in the author's laboratory confirm the following conclusions: Some test compounds give just as high recoveries on C18 silica as on PS-DVB materials, but in general recoveries are significantly ’igher on the polymeric sorbents, 2. On the average, recoveries of test compounds are better on resins with hydrophilic- surface substituent than with the parent, unfunc- tionalized resin.20>) 66 ss 86 ool 06, 96, 16 oot 06 96 16 98 ss 56 66 86 REE RRS +6 1 AN pue IMA J9pUN gas Aq spunodwo; spEWOAY KxoupAystog pue ‘spunoduio;) speWIOAY ‘sJoUrHA JO 66 +6 &6 faq ut 2 9% 6 9s 16 os 88 a 56 ° 18 > 18 Lp £6 ie 96 zl 16 © 16 ag AN suo1ya1oquUIy ° ° 8s or 1 > ° uu sl eS m Toyoopeycauog 6 ounruy 8 jostuy 8 louayd)Aing-1421-p uw jousyd Sowa ° 2 > Sr 99 on 9 ra one unodwo ISss19 pep Bape aa oo GE ATANL 49‘uoysqutiod ype 6 a0Uau942q WOH [200s wy 95 0 % 0 0 0 ° Md +6 66 s 86 0 u 0 o auounboupsysinayy 8 8 0 zw 0 9% 0 0 2uoutnboupsHy 96 66 oO 6 0 8 0 0 Joursuosaryktpayy-o 6 6 0 88 0 19 oO 0 joursosay 0€ sk 6 68 0 we 0 0 yoypareD %6 86 88 +6 se o8 9 Pea aiuony 06 ool v8 96 8S 1g 1. 06 arepernydi cpa, $8 96 +8 68 el z 0 +8 aivocuogyswodosy 6 96 6 6 vs $8 9 88 suousydoror 25 ml syringe. The sample solution is added directly to the column reservoir. The70 PRACTICAL CONSIDERATIONS: EQUIPMENT AND TECHNIQUES Reservoirs Adaptor Columns Valves Vacuum Discharge box of SS Vacuum Regulator FIGURE 4.4 SPE with vacuum technique. Vacuum FIGURE 4.5 Vacuum technique with large samples (3)49. COMPLETENESS OF EXTRACTION n 0 ° FIGURE 4.6 Pressure technique with syringe. syringe is filled with air and attached to the SPE column as shown in Figure 4.6 (4). The sample is then pushed through the SPE tube at the desired rate by applying pressure manually to the syringe piston. A commercial single SPE tube processor uses a threaded piston (5). By rotating a knurled knob, pressure is applied smoothly for slow flow through the system. The plunger can be pressed for more rapid flow. ‘Another simple way to overcome the resistance of the packed tube to liquid flow is to apply air or nitrogen pressure to the liquid sample in the reservoir. A simple needle valve is used to regulate the pressure applied and hence the rate of flow. An applied pressure of 1 or 2 bars is usually sufficient to achieve the desired flow rate. 4.3. COMPLETENESS OF EXTRACTION 4.3.1 Breakthrough Volume For efficient extraction to occur, distribution of an analyte between the sample liquid and the solid SPE phase must strongly favor the latter. This ‘means that the distribution ratio, D, must be as large as possible [see Eq. (1.7) In column chromatography the distribution ratio, represented by the symbol k (or sometimes &), is called the retention factor or the capacity7 PRACTICAL CONSIDERATIONS: EQUIPMENT AND TECHNIQUES ‘factor. In elution chromatography, k is defined as the ratio of the amount (not the concentration) of analyte in the stationary phase to the amount in the mobile (liquid) phase. After introduction of the sample and pumping, mobile phase continuously through the system, the analyte will emerge as 4 peak with the shape of a normal distribution curve. The retention volume, Vey of this peak is given by the equation Vq=Vo(1+4) 4.) where Vp is the volume of liquid within the column and detector system, In SPE the situation is a little different from that in elution chromatog- raphy. Now the sample itself is the mobile phase. When a sufficient volume of sample has passed through the SPE tube, the analyte begins to emerge from the tube in a concentration profile much like that of the chroma- tographic peak. However, in SPE the concentration of analyte does not reach. ‘a maximum and then decrease as it does in elution chromatography because ‘more sample continues to flow into the system. The point at which the first analyte leaves the packed tube is called the breakthrough volume, Vz (See Fig. 4.7) (6,7). The breakthrough volume is sometimes taken as the volume when the concentration of analyte leaving the tube constitutes 1% of its initial concentration in the sample. ‘The shape of a chromatographic peak depends on the number of theo- retical plates (V) generated by the column: 4.2) hi (4.2) UV response Ao Vy V, Vm, Sample volume FIGURE4.7 Theoretical breakthrough curve obtained by percolation ofa spiked sample (UV absorbance Ao) through a precolumn. See text for definintion of Vo, Vg, and Vin. (From Ref. 6 with permission.)ioe where Vp is the retention volume of the analyte and w is the width of the peak at its base (w = 46, where o is the standard deviation), For any given Vg the peak will be sharper (smaller w) as N increases. A sharp breakthrough profile is also advantageous for SPE because Vp will occur later than when the breakthrough profile is more drawn out. Thus a larger number of theoretical plates in a SPE column is conducive to longer retention of analytes before breakthrough. From this discussion itis apparent that two major conditions are needed for good SPE: a high retention factor (k) and a column with good partition- ing efficiency. The latter can be represented by its number of theoretical plates (N). However, for a very short column such as an extraction tube, N will not be very large even for an efficient system. Even so, the efficiency of most SPE tubes in terms of Nis not very good. The value of N fora typical cartridge has been estimated to be 20 plates (7). There are several reasons for this. The particle size of typical commercial SPE device is at least 40-50 um, compared to particles of 3 or 5 um in an HPLC column. Partition, equilibrium of analytes between the liquid and solid phases is much slower with the larger particles, and NV is therefore smaller than for a comparable column packed with the smaller particles. The particle size range in SPE is much broader than in HPLC, leading to additional broadening of solute zones, Finally, SPE tubes are not as carefully packed as columns intended for chromatographic use. In short, SPE devices are just good enough to do their intended job. Equation (4.1) indicates that the retention volume of an analyte depends strongly on its retention factor, k. For a substance to be extracted strongly and have a large breakthrough volume (Vg), it must have a large k. Values of k vary tremendously with the chemical structure of the analyte as well as with the type of packing used in the SPE tube, Some typical values of k for ‘an assortment of chemical compounds on a bonded-phase silica column are listed in Table 4.2. These values for ky (retention factor in water) were selected from a paper by Braumann (8). The k, values listed are only approximate because they had to be obtained by linear extrapolation from methanol—water solutions, where the k values were low enough to be ‘measured experimentally, to pure water. Polar organic molecules have rather low ky values. As a result, their extraction by SPE tubes might be incomplete. At the other extreme, large fused-ring compounds such as anthracene and pyrene have very high k,, values. These compounds will be extracted very strongly but complete elution could pose a problem unless ‘an eluting solvent can be found where the & value would be low enough to favor fast elution.