Note: more detailed answers to most of these questions can be found in “A guide for the use of semipermeable membrane devices (SPMDs) as samplers of waterborne hydrophobic organic contaminants” by Huckins et al. (2002), a USGS report to the American Petroleum Institute (API), Washington, D.C., API no. 4690, pp 186.

Answers to Frequently Asked Questions about SPMD Technology

by

James Huckins

Jim Petty

1. Is there a standard or a reference SPMD design?

Yes, the standard “commercial” SPMD configuration consists of a thin film of triolein (³95% pure) sealed in a low-density polyethylene (LDPE) layflat tube (70-90 µm wall thickness) that is manufactured without additives. The fractional lipid content is » 0.2, or 20%. Typically, the standard SPMD contains 1 mL triolein, has dimensions of 2.54 cm wide by 91.4 cm long, and the membrane surface area is about 450 cm².

Note: Changing the LDPE membrane surface area and/or the size of the SPMD generally does not constitute a deviation from the standard design. However, changing the membrane thickness, fractional lipid content, or using LDPE with additives may result in significant changes in performance characteristics relative to the standard design. Finally, nearly all SPMD calibration data for estimating ambient environmental concentrations are based on the standard SPMD design, thus significant deviations from the standard design may invalidate the use of these calibration data.

back to List of Questions

2. Why have a standard SPMD design when SPMDs can be tailor-made for specific applications?

SPMD technology has two great advantages besides its passive mode of operation: (a) the uniformity or reproducibility of the sampling matrix and (b) the suitability of the devices for monitoring multiple media under a wide range of environmental conditions. Thus, the ability to compare data from global, regional, local, and contiguous sites is dependent on sampling methods that are uniform, reproducible and usable in multiple media. The need for global uniformity is clearly evident when the only calibration data available for estimating ambient chemical concentrations are derived from standard-SPMDs.

However, sometimes data comparability beyond a particular set of sites is not an issue, and customized SPMD designs or changes in the recovery-cleanup protocol may provide a significant advantage over the standard approach. Never the less, when investigators use devices that significantly deviate from the standard configuration, several standard SPMDs should be included as well to help demonstrate the level of enhancement over the performance of the standard-SPMD.

back to List of Questions

3. What are the current applications of SPMD technology?

Current applications of SPMDs include: a) determination of pollutant sources and relative levels at different locations, b) estimation of ambient solute or vapor phase time weighted average concentrations, c) in situ biomimetic concentration of ambient bioavailable chemicals for bioassay and immunoassay, d) estimation of organism exposure or bioconcentration potential, e) analytical enrichment of contaminant residues, and f) use in toxicity identification evaluation procedures.

back to List of Questions

4. What chemicals do SPMDs sample?

SPMDs may sample any nonionic organic compound with a K_ow value > 1, but in practice, a chemical's K_ow should be greater than 300. The following classes of compounds (not all-inclusive) have been shown to concentrate in SPMDs:

a. Polycyclic aromatic hydrocarbons (PAHs)

b. Polychlorinated biphenyls (PCBs)

c. Polychlorinated dioxins and furans

d. Organochlorine pesticides and several “new generation” pesticides

e. Pyrethroid insecticides

f. Nonyl phenols

g. Several herbicides and many industrial chemicals

h. Tributyl tin and alkylated selenides

i. Others

back to List of Questions

5. What types of environmental media can SPMDs sample?

SPMDs can be used to sample air (vapor phase), water (surface and groundwater), and sediment-soil (pore-water and soil vapor phase). They have been used in environments from the tropics to the Arctic and Antarctic, and in highly turbulent systems to stagnant backwater areas. In general, water quality parameters do not affect SPMD sampling of solutes, but in some cases (e.g., pH and DOC) may affect the amount of chemical available for uptake.

back to List of Questions

6. What advantages (if any) do SPMDs offer over grab sampling?

Grab samples provide data only on a single point in time and, because small volumes are typically sampled (< 5 L), the approach is often inadequate for detecting trace bioconcentratable residues. For compounds with relatively large K_ow values, SPMDs integratively sample residues, i.e., losses of accumulated chemicals are insignificant during an exposure period. Integrative SPMD sampling mimics the initial part of the bioconcentration process and generally permits the concentration of trace chemicals to levels sufficient for bioindicator tests. The amount of water extracted by a standard 1 mL triolein SPMD may approach 100 L for a thirty-day exposure to compounds with moderate to high K_ows. Also, episodic events can be detected without an intensive sampling program and, in some cases, estimates of total daily solute loads in flowing waters can be made from SPMD concentrations.

back to List of Questions

7. Do SPMDs sample only dissolved or vapor phase chemicals?

Nonporous polymeric films such as low-density polyethylene (membrane of choice for SPMDs) contain transient cavities with maximum diameters of about 10 Å. These cavities are far too small to accommodate colloids or macromolecular dissolved organic carbon (DOC) such as humic acids.

Also, comparisons of chemical concentrations determined by using traditional analytical methods for ultra-filtered river water (colloids and DOC > 50 Å diameter were removed) and those estimated from SPMDs exposed to river water appear to confirm that SPMDs sample only dissolved residues, which are readily bioavailable.

The importance of measuring the vapor and dissolved phases of chemicals in environmental systems be mentioned. The fact that threshold limit values (human exposure) and toxicity- databases (aquatic organisms) are based on vapor and dissolved concentrations, respectively, demonstrates the value of the SPMD approach.

back to List of Questions

8. What factors determine how much of a bioavailable chemical will be sequestered by an SPMD?

The amount depends on the chemical’s sampling rate (liters of water or air extracted per day) by an SPMD, SPMD capacity for the analyte, the water concentration (see question 9), and exposure conditions (temperature, flow velocity/turbulence, and biofouling) and duration. Factors affecting sampling rate are complex, but nonpolar lipophilic organic compounds with log K_ows between 5.0 and 6.5 and molecules with cross-sectional diameters < 10 Å are sampled at the highest rates for a specific set of environmental conditions. Of the environmental conditions affecting sampling rates, the level of flow velocity/turbulence may have the greatest impact.

Assuming biofouling is minimal, a 1-g triolein SPMD will daily extract all components of complex mixtures such as PAHs, PCBs, and organochlorine pesticides from about 0.5 to 10 L of water. As the lipophilicity or the K_ow of a series of organic compounds rises to » log K_OW of 6.5, so does the capacity of the SPMD triolein to sequester them. For compounds with relatively large K_ow values (i.e., > 100,000), extended exposure periods will typically result in greater masses of analytes sequestered, but increased biofouling may reduce the daily amount of residues sequestered during the latter part of long exposures.

back to List of Questions

9. How can SPMDs have a constant sampling rate independent of environmental concentration when the amount of chemical sequestered is proportional to the ambient chemical concentration?

Sampling rate can be defined several different ways. Often sampling rate is expressed as mass of chemical extracted (taken up) per unit time. This is a zero-order expression of uptake rate and is clearly dependent on chemical concentration.

However with SPMDs, sampling rates are normally expressed as first-order uptake rate constants, which are given in volume of medium cleared of chemical, per gram of sampler per unit time, or Lg^-1d^-1. Several SPMD studies have shown that the first order rate constant does not vary with solute concentration but is affected by flow/turbulence, temperature and biofouling. Therefore, the amount of chemical recovered will be proportional to the average environmental concentration but the magnitude of the rate constant is not. This first-order behavior has also been observed in selected studies of contaminant bioconcentration.

back to List of Questions

10. How does SPMD size affect the sampling rate of a chemical?

SPMD sampling rates are directly proportional to SPMD membrane surface area. For example, a standard 1-g triolein SPMD (surface area » 450 cm²) may extract 5 L of water per day for a PCB congener, whereas a standard triolein SPMD with half the surface area » 225 cm² (0.5-g of lipid) can be expected to extract 2.5 L of water per day of the same congener, assuming similar conditions for the exposures.

back to List of Questions

11. How do the concentrations of contaminants in exposed SPMDs relate to those found in aquatic organisms from the same site?

Direct comparisons of bivalves to SPMDs have shown that for nonionic organic contaminants, SPMDs accumulate a broader range of chemicals from water than bivalves. Note that mollusks are often the organisms of choice for biomonitoring, because their capacity to metabolize most contaminants is very limited. SPMDs are passive in situ partitioning systems without the active depuration mechanisms of living organisms. SPMD concentration factors appear to mimic the worst-case scenario of the bioconcentration of organic chemicals with log K_ows < 6.0. Even chemicals whose primary uptake route by organisms is via the diet-food chain (i.e., log K_ow ³ 6.0) are readily concentrated by SPMDs, because ultra-trace levels of their residues are always present in the water. However, the use of longer exposure periods or an increased number of SPMDs per sample may be necessary to detect some of these compounds.

Comparisons of SPMD, bivalve and fish uptake rates have shown that SPMDs concentrate PCBs and PAHs at a rate that is about 0.5 to 1.0 times and 1.0 to 2.0 times that of bivalves and fishes, respectively. However, unlike organisms, SPMDs seldom reach steady state (i.e., equilibrium) with the ambient environment, because they contain much higher levels of lipid than most organisms. Never the less, it is possible to estimate steady-state concentrations of contaminants in an organism's tissues by using the SPMD-derived water concentration along with the compound's K_ow. Finally, since organism uptake rates and bioconcentration factors differ greatly for the same chemical among different species, no single device can be expected to mimic all organisms.

back to List of Questions

12. Does water quality or temperature affect SPMD sampling?

In general, water quality (salinity, alkalinity, etc.) has very little effect on SPMD sampling. Temperature is an exception because SPMD-sampling rates rise with temperature in water and fall when temperature increases in air. For example, a 16 ^oC change (10-26^oC) in water temperature resulted in a two- to four-fold increase in chlorinated pesticide sampling rates. Also, pH affects the amount of ionizable compounds (e.g., chlorinated phenols) sampled, and the magnitude of this effect is proportional to the increase or decrease in the neutral species.

back to List of Questions

13. How much does biofouling affect the uptake rate of chemicals by SPMDs?

Biofouling impedes but does not stop the uptake of chemicals by SPMDs. Laboratory studies have shown that the uptake of some compounds by heavily fouled SPMDs is reduced by as much as 69 %. In particular, compounds with high K_ows are impeded more than those with low K_ows. Fouling impedance is generally insignificant for the first 2 weeks of an exposure but may become significant in the second two-week interval of an exposure. However, observations suggest that biofouling of the exterior membrane surface may reach a maximum after about 1 month (obviously the time to this point is dependent on the nature of the aquatic test system) and does not increase much thereafter. In most cases, biofouling causes a decrease in the slope of the SPMD uptake line (plot of time vs. SPMD concentration), and in some cases may limit the value of longer SPMD exposures. Note that permeability/performance reference compounds (PRCs) can be used to correct sampling rates for biofouling.

back to List of Questions

14. What is a permeability/performance reference compound (PRC) and how does it work?

A PRC is an analytically non-interfering compound, such as certain perdeuterated PAHs, which has moderate-to-fairly high SPMD fugacity (escaping tendency). PRCs are added to the SPMD lipid before field studies and calibration exposures. Measured values of PRC loss rates are used to account for any rate affecting differences between field and calibration study (i.e., sampling rates measured under a specific set of conditions) exposure conditions. This approach to in situ SPMD calibration is based on the principle that the rate of residue loss is proportional to the rate of residue uptake. Thus, PRC loss rate data can be used to adjust SPMD-derived estimates of ambient concentrations to reflect site-specific environmental conditions of an exposure. Using the PRC method and previously developed models, SPMD based estimates of ambient water concentrations are within 2 fold of separately measure values.

back to List of Questions

15. What effect does water or air velocities have on the uptake of chemicals by SPMDs?

Medium flow rates should have only small effects on the sampling rates of organic compounds under membrane control (i.e., log K_ow < 4.4), because transport across the membrane is the rate-limiting step. In this case, the membrane is like a barrier or valve that allows only some fraction of the molecules contacting the membrane surface to permeate into the lipid. Increased flow or turbulence does not cause a significant rise in chemical concentration at the membrane surface but merely ensures that chemical concentration in the aqueous diffusion layer reflects that of the bulk water. As long as the chemical concentration at the membrane surface is not depleted by sampling (i.e., encounter volume or water volume exchanged at the membrane surface is sufficient), then chemical uptake is independent of flow regime.

For compounds with log K_ow values ³ 4.4, the external boundary layer will likely control uptake rates for all but extremely turbulent conditions and very large molecules. Because flow/turbulence can vary greatly in environmental exposures, turbulence-induced changes in the sampling rates of chemicals with diffusion layer control can be as high as tenfold. However, the effect of turbulence can be corrected by using a PRC as described earlier and by designing deployment apparatuses with the capacity to baffle flow.

back to List of Questions

16. How are contaminant residues in exposed SPMDs recovered and how much cleanup of the recovered residue is necessary before in vitro bioassay or analytical detection?

A major portion of the sequestered residues can be recovered by opening the ends of the SPMD polyethylene tube and rinsing out the lipid with an organic solvent. However, analytes are generally recovered by dialyzing the intact SPMD (which requires removing periphytic growths, minerals, and debris from the exterior membrane surface) in an organic solvent such as hexane. Using this approach, contaminant residues present in the membrane (sometimes representing as much as 50% of total) are also recovered for analysis and the dialysis process separates nearly all of the bulk lipid from the chemicals of interest. In nearly all cases, additional lipid and polyethylene wax removal is required for analytical detection of trace compounds. Also, some additional steps may be required for tests such as EROD and ELISA. The in vitro bioassays Microtox® and Mutatox® may not require lipid removal if high purity lipid is used in SPMDs.

back to List of Questions

17. How do in vitro bioassay test results, such as Microtox® EC₅₀values, relate back to the volume of water sampled?

To answer this question, we use as an example Microtox® EC₅₀ values that are given in units of mg SPMD/mL of sample carrier solvent. If the Microtox® EC₅₀ value is 3.1 mg/mL, then it only took an equivalent of 3.1 mg of a standard 1-g triolein SPMD to elicit a toxic response. Depending on the sampling rates of the toxicant(s) detected in the bioassay, 1 mg of an SPMD would have extracted chemicals from about 0.5 to 10 mL of water per day.

Assuming that the SPMD sampled toxicants from a site at about 5 mL day^-1 mg^-1, and the exposure was 7 days, then the 3.1 mg/mL EC₅₀ value represents the response to the amount of toxicant contained in 108 mLs of sample water. Typically residue extracts are transfered into 1-mL of carrier solvent for assay, resulting in a 108 fold concentration factor. However, if only 0.1 mL of the carrier solvent is used for the assay, then this sample represents the chemical in only 10.8 mL of test water. Clearly, if only 1 mL of sample water was extracted or used directly for Microtox, the result may have been “no effect”. Note that the same type of analysis can be performed for the whole SPMD if dialysis is used and much greater concentration factors can be achieved with SPMDs if desired.

back to List of Questions

18. When comparing SPMD data to residue levels in biota, should lipid normalization be used?

Although residue concentration data of organisms are often lipid normalized, the procedure is generally inappropriate for SPMD data. A key assumption in the normalization procedure is that residues in an organism’s lipids have reached steady state with ambient waters. This assumption may be valid for feral organisms, but it is frequently not valid for biomonitoring organisms, especially when high-K_ow contaminants are the target analytes and exposure durations are short-term to moderate (less than 30-day). In general, the capacities of SPMDs for hydrophobic analytes are much greater than those exhibited by aquatic organisms, because SPMDs have higher percent lipid contents. In addition, the lipid-like membrane significantly adds to the capacity of SPMDs for hydrophobic organics, whereas non-lipoidal tissues add little to an organism’s capacity to accumulate these compounds.

Several studies have shown the similarity of the magnitude of the first-order uptake rate constants of SPMDs and organisms and the greater analyte capacity of SPMDs. Thus, biomonitoring organisms will have much shorter linear uptake phases or times to equilibrium than SPMDs. When SPMDs and biota are both in the linear or integrative phase of uptake, lipid weighting of data is certainly not justified. For example, if equal weights of fish or SPMDs were in the linear phase of chemical uptake (constant concentration) and their sampling rates were identical, both matrices would contain the same amount of chemical. However, if the fish contained 2% lipid and the SPMD contained 20% lipid, normalization of their chemical concentrations to lipid weight would result in an apparent tenfold greater concentration in the fish with 2% lipid. This procedural artifact contrasts with the similar lipid-weighted values (assuming lipid normalization theory is truly applicable) that would be obtained if both sampling matrices were at equilibrium. In the case of bivalve biomonitoring organisms, where percent lipid content varies from only about 0.5 to 2.0 % (dry weight versus wet weight), the potential for large errors in SPMD-bivalve comparisons is high. Even when the organism has reached equilibrium and the SPMD has not, significant errors can still be introduced by the normalization procedure. In these cases comparisons of whole-body and whole-SPMD concentrations and/or total mass of chemical accumulated per sample is more appropriate.

back to List of Questions

19. Can SPMDs be used for both integrative and equilibrium partition sampling? What are the pros and cons of the two approaches?

The answer to the first question depends on the magnitude of the K_ows of the target analytes, the SPMD design and environmental conditions. Investigators have shown that SPMDs (standard configuration as described in the tutorial) integratively sample compounds with log SPMD-water partition coefficients (K_SPMDs) ³ 5.0 during exposures of less than 1 month at 18 ^oC, and a flow velocity <1 cm/sec. Note that in some cases small amounts of dispersed carbon adsorbent in SPMD triolein increased SPMD capacity, and permitted integrative sampling of even low K_ow compounds (e.g., naphthalene log K_ow = 3.4 and for low K_ows, K_SPMD » K_ow) as well. Also, it has been shown for compounds with log K_ows < 4.0, 90% of equilibrium concentrations may be achieved in < 40 days using the standard SPMD design. The time to equilibrium can be decreased by increasing the flow-rate of water and air (the likely switch to membrane control limits the effects of flow for compounds with log K_ows < 4.5) and the temperature.

Let us examine the second part of the question. The American Conference of Governmental Industrial Hygienists adopted the integrative time-weighted average (TWA) approach as the most satisfactory means of monitoring chemicals that are not fast-acting toxicants. This approach provides an estimate of the cumulative dose (organic vapors, or dissolved residues as in our case) of contaminants during a specified exposure period. Unlike monitors that rapidly achieve equilibrium, episodic chemical events during the initial part of an extended exposure can usually be detected. This is because the rate of residue loss from an integrative sampler is vanishingly small (i.e., these monitors have very high chemical capacities). Note that an important feature of TWA concentration estimates is that they do not require constant analyte concentrations.

The equilibrium partitioning approach has been widely used in biomonitoring studies and to model chemical concentrations in environmental media. For an organism or SPMD to reach steady-state concentrations of lipophilic chemicals in a 30-day exposure, uptake rates (k_u) must be rapid and the total residue capacity must be relatively low. Also, since the equilibrium SPMD-water or K_SPMD partition coefficient is equal to k_u/k_e, where k_e is the depuration or dissipation rate constant, then the loss rate (k_e) must be great as well. Thus, it is possible that sequestered residues from episodic events could fall below the analytical detection limits. Basically, organisms and samplers, selected or designed to rapidly reach equilibrium, also rapidly lose or rise in chemical concentration as ambient chemical levels change, and their chemical concentrations do not necessarily represent the entire exposure period. However, if multiple samples are taken through time, peaks in analyte concentrations can be delineated.

back to List of Questions

20. Can a chemical's K_ow be used to estimate how long it will be integratively sampled by an SPMD or the exposure time required to reach equilibrium?

At equilibrium, the volume of ambient water cleared of chemical or the thermodynamic capacity of the SPMD is given by

V_wEQ = V_SPMDK_SPMD

where V_wEQ is the equilibrium clearance volume (water), V_SPMD is the volume of the SPMD, and K_SPMDis the SPMD-water partition coefficient. Because sampling rates are expressed volumetrically in L day^-1 (R_s or L g^-1day^-1[k_u]) of water extracted of chemical, the fractional amount (F_r) of water cleared relative to the equilibrium capacity of the lipid is given by

F_r = V_wcL/(K_SPMDV_L)

where V_wcL is R_st or the total volume of ambient water cleared of chemical at time t and R_s is the sampling rate in Ld^-1. The following equations provide some guidance on the estimation of the time a chemical will be in the linear phase of uptake and the time required to reach 90 % of the equilibrium concentration.

t₅₀ = t_1/2 = –ln 0.5 K_SPMD V_SPMD/ R_s = -ln 0.5 K_SPMD/ d_SPMD k_u

t₉₀ = -ln 0.1 K_SPMD V_SPMD/ R_s = -ln 0.1 K_SPMD/ d_SPMD k_u

t_1/2 » –ln 0.5 K_ow V_SPMD/ R_s » -ln 0.5 K_ow/ d_o k_u

where d is the density of the specified material. Keep in mind that SPMDs sample integratively (linear uptake) during one half-time (t_½) and it takes about four t_½s to reach > 90% of equilibrium concentrations. However, note that biofouling can increase t_½ in both instances. See Modeling section in the Tutorial for a similar analysis.

back to List of Questions

Basics.. Glossary.. Overview...Tutorial.. . FAQs...References... USGS/WRD SPMD Activities.. Other SPMD Links

If the reader has additional questions concerning SPMDs, contact Dave Alvarez: (573) 441-2970; e-mail: dalvarez@usgs.gov