Study: Organic compounds in produced waters from shale gas wells

Environmental ScienceProcesses & Impacts rsc.li/process-impacts

ISSN 2050-7887

PAPERSamuel J. Maguire-Boyle and Andrew R. BarronOrganic compounds in produced waters from shale gas wells

Volume 16 Number 10 October 2014 Pages 2217–2462

EnvironmentalScienceProcesses & Impacts

PAPER

Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article OnlineView Journal | View Issue

Organic compou

aDepartment of Chemistry, Rice University,

rice.edu; Tel: +1 713 348 5610bDepartment of Materials Science and Nano

77005, USAcEnergy Safety Research Institute, Colleg

Singleton Park, Swansea SA2 8PP, Wales, U

† Electronic Supplementary Informationcontent versus conductivity, and pH ver

ICP-OES. Tables of full GC/MS analysis10.1039/c4em00376d

Cite this: Environ. Sci.: ProcessesImpacts, 2014, 16, 2237

Received 9th July 2014Accepted 13th August 2014

DOI: 10.1039/c4em00376d

rsc.li/process-impacts

This journal is © The Royal Society of C

nds in produced waters from shalegas wells†

Samuel J. Maguire-Boylea and Andrew R. Barron*abc

A detailed analysis is reported of the organic composition of produced water samples from typical shale gas

wells in the Marcellus (PA), Eagle Ford (TX), and Barnett (NM) formations. The quality of shale gas produced

(and frac flowback) waters is a current environmental concern and disposal problem for producers. Re-use

of produced water for hydraulic fracturing is being encouraged; however, knowledge of the organic

impurities is important in determining the method of treatment. The metal content was determined by

inductively coupled plasma optical emission spectrometry (ICP-OES). Mineral elements are expected

depending on the reservoir geology and salts used in hydraulic fracturing; however, significant levels of

other transition metals and heavier main group elements are observed. The presence of scaling elements

(Ca and Ba) is related to the pH of the water rather than total dissolved solids (TDS). Using gas

chromatography mass spectrometry (GC/MS) analysis of the chloroform extracts of the produced water

samples, a plethora of organic compounds were identified. In each water sample, the majority of

organics are saturated (aliphatic), and only a small fraction comes under aromatic, resin, and asphaltene

categories. Unlike coalbed methane produced water it appears that shale oil/gas produced water does

not contain significant quantities of polyaromatic hydrocarbons reducing the potential health hazard.

Marcellus and Barnett produced waters contain predominantly C6–C16 hydrocarbons, while the Eagle

Ford produced water shows the highest concentration in the C17–C30 range. The structures of the

saturated hydrocarbons identified generally follows the trend of linear > branched > cyclic. Heterocyclic

compounds are identified with the largest fraction being fatty alcohols, esters, and ethers. However, the

presence of various fatty acid phthalate esters in the Barnett and Marcellus produced waters can be

related to their use in drilling fluids and breaker additives rather than their presence in connate fluids.

Halogen containing compounds are found in each of the water samples, and although the fluorocarbon

compounds identified are used as tracers, the presence of chlorocarbons and organobromides formed

as a consequence of using chlorine containing oxidants (to remove bacteria from source water),

suggests that industry should concentrate on non-chemical treatments of frac and produced waters.

Environmental impact

Hydraulic fracturing has become a controversial technology in the development of unconventional shale gas reserves. The large volumes of water traditionallyused during the hydraulic fracturing process and the resulting produced and owback waters are of current environmental concern and disposal problem forproducers. Knowledge of the types of organic impurities is important in determining the method of treatment and potential water re-use. We report on ndingsof recent studies and offer recommendations and limitations of specic water treatment methods.

Houston, TX 77005, USA. E-mail: arb@

engineering, Rice University, Houston, TX

e of Engineering, Swansea University,

K

(ESI) available: Plots of total cationsus [Ca]. Rankings of wavelengths forresults for water samples. See DOI:

hemistry 2014

1. Introduction

The fastest growing trend for US domestic energy generation isoverwhelmingly un-conventional natural gas,1 and the envi-ronmental benets of energy generation from gas compared tocoal is obvious.2,3 The vast amount of hydrocarbons potentiallyavailable in the continental United States is in the form ofonshore shale gas as seen from the developments currentlyunder way, including: the Marcellus, Haynesville, Fayetteville,Barnett, Eagle Ford, Bakken, Antrim, Utica, Niobara, NewAlbany, Woodford, and Bossier plays. Unconventional natural

Environ. Sci.: Processes Impacts, 2014, 16, 2237–2248 | 2237

http://crossmark.crossref.org/dialog/?doi=10.1039/c4em00376d&domain=pdf&date_stamp=2014-09-22

http://dx.doi.org/10.1039/c4em00376d

http://pubs.rsc.org/en/journals/journal/EM

http://pubs.rsc.org/en/journals/journal/EM?issueid=EM016010

Environmental Science: Processes & Impacts Paper

Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

gas production has increased 10-fold between 2001 and 2011,4,5

and shale gas is projected to increase from 23% of US naturalgas production in 2009 to 47% by 2035, offsetting decliningproduction from conventional sources.

The ability to extract shale gas in an economic and timelymanner has been achieved by the development and use ofhydraulic fracturing and horizontal drilling techniques.Hydraulic fracturing (also known as “fracing” or “fracking”)uses water, proppant, and various chemical additives, pumpedat high pressures into the well bore, to induce fracturing of theshale source rock and thus create greater permeability so thegas can migrate into the well bore and to the surface. Thefracturing uid chemistry is tailored on a case-by-case basis foreach geographical area and sometimes even on a well-by-wellbasis. Although typical fracturing uid contains predominantlywater (ca. 90%) as well as sand or a ceramic proppant (8–9%),the uid also contains all or some of the following classes ofchemicals: salts, friction reducers, scale inhibitors, biocides,gelling agents, gel breakers, and organic and inorganic acids.6,7

There have been wide-spread concerns over the use of theseadditives; however, the large volumes of water used duringhydraulic fracturing is also of a concern. In Texas, estimates ofwater use for hydraulic fracturing vary with median values of10 600 m3 (2.8 million gallons) for the Eagle Ford play to 21 500m3 (5.6 million gallons) for the Texas portion of the Haynesvilleplay.8 It is worth noting that while, on a state wide basis, this isapproximately 1.5% of the amount used in farming andmunicipalities, in local usage it can be a considerable burden.Aer the hydraulic fracture the pressure is released and the fracuid is allowed to return to the surface. The chemically engi-neered water that returns to the surface is called “owbackwater”, and is mixed with water that is already in the sourcerock, “connate uids”. Generally the ow rate of the returningwater is very high with very little hydrocarbon. The amount ofload water (the amount actually injected during the frac)recovered aer a fracture ranges from 5–15% in the Eagle Fordand Haynesville to 50–60% in the Barnett and some parts of theMarcellus. However once this water has returned to the surfaceand the well begins to produce hydrocarbons the water thatcomes to the surface is now called “produced water”. Generallythis water has far lower or no amounts of the engineeredcomponents that were added to the frac water initially.However, the chemical make-up of the production water maystill include quantities of chemicals and materials, such asinorganic salts, bacteria and organic molecules, with theorganic molecules being either naturally occurring or a residuefrom the added components. Unlike conventional wells inwhich volumes of produced water generated over the lifetime ofa well are oen enormous and far exceed amounts of petroleumrecovered,9 shale gas wells produce most of their water withinthe rst few weeks of production. Subsequently, a few barrels aday are commonly produced.

Given that projections of cumulative net water used in allshale plays during the next 50 years totals 4350 Mm3 (over 1trillion gallons), there is an incentive to reuse frac owback andproduced water for hydraulic fracturing. Unfortunately,produced water is oen unsuitable for reuse in frac uids due to

2238 | Environ. Sci.: Processes Impacts, 2014, 16, 2237–2248

high dissolved salts and high organic content, including:hydrocarbons, greases, and biological matter. Furthermore,produced water discharges offshore have been shown to be toxicto marine organisms,10,11 and hence it is illegal for recoveredwaters from shales to be released into rivers or groundwaters.However, concerns still linger with regard to accidental spillageand exposure. In an attempt to eliminate exposure of wasteproduced water to the environment companies use encloseduid capture systems. One common disposal practice in theBarnett Shale production area of Texas involves re-injecting thewastewater uids back into the ground. Unfortunately, this isnot a long-term solution, and it will be necessary to cleanproduced water for reuse in hydraulic fracturing.

The chemical and physical characteristics of produced waterfrom conventional and unconventional oil and gas reservoirsworldwide and the potential treatment options for the watershave been reported.12–14 However, in most cases the organiccontent has been described by the total organic content (TOC)rather than the individual species present. A detailed study byOrem et al. has investigated the identity of organic compoundsin produced water for coalbed gas wells;15 but this sourcepresents different issues to those of shale reservoirs. As part ofour studies into the treatment of oil–water mixtures,16 andunderstanding the challenges in treating frac and producedwater to make it ideal for re-use, we are interested in thecomposition and compositional variation between variousproduced waters. Such information will assist in understandingof whether a particular treatment process can be used generallyor which treatment processes should be applied to differentproduced waters. Witter and Jones have previously reported thatdichloromethane extraction of acidied conventional oil eldwastewater facilitates analysis by gas chromatography/massspectrometry (GC/MS).17 We have adapted this methodology forproduced waters using both neutral and acid extraction todifferentiate high from low polarity organic compounds. Ourstudy is presented herein along with suggestions for futuretreatment protocols based upon the results.

2. Experimental

Chloroform (anhydrous $99%, 0.5–1.0% ethanol as stabilizerspectroscopic grade), nitric acid (ACS reagent grade, 70%purity), hydrochloric acid (ACS reagent, 37% purity), potassiumhydroxide ($85% purity), anhydrous sodium sulphate (ACSreagent, $99%) and ethanol (anhydrous $99.5%) werepurchased from Sigma-Aldrich. Isopropyl alcohol (99.5%) waspurchased from EMD. All water used was puried via Milliporeltration to 18 MU deionized (DI). ICP standards (IV-ICPMS-71A) were obtained from Inorganic Ventures. Whatmann ltersno. 40 were obtained from Fischer Scientic. Thermo Scientic50 mL and 15 mL sterilized graduated conical centrifuge tubeswere obtained from Fischer Scientic.

2.1. Water sampling

Produced water samples have been collected at three well sites.In our initial study the goal was to determine the variability

This journal is © The Royal Society of Chemistry 2014


Paper Environmental Science: Processes & Impacts

Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

across a wide geographic range, rather than a well-to-well vari-ation within a single eld. In this regard the three samples weredrawn from wells in the Marcellus (PA), Eagle Ford (TX), andBarnett (NM). In each case the water was produced rather thanfrac owback, i.e., the wells were all producing gas.

Water samples were collected in 1 L mason jars. All jars usedfor sample collection were cleaned using detergent and rinsedthoroughly with 18 MU water. The jars were then soaked in basebath containing KOH in isopropyl alcohol for one week thenrinsed thoroughly with ethanol then DI. The jars were thensoaked in concentrated nitric acid from three days and againrinsed thoroughly with DI. All glassware used followed the samecleaning procedure. Samples were collected directly from fractanks into which the produced water was directly piped. Nomixing with other water sources was observed. All samples werecollected without a headspace to reduce oxidation of thesample. All samples were stored and transported in the dark insealed containers under refrigeration. Blank samples weretested using 18 MU deionized water. These showed that nochemicals were leached from the O-rings in the jar or contam-ination from other sources during handling and storage.

2.2. Conductivity and pH analysis

The as collected water sample jars were decanted and pre-ltered three times withWhatmann Filter no. 40 before analysisto remove any samples containing visible particulate matter ornon-dissolved matter. The ltered water was added to a glass 20mL scintillation vial, which had been cleaned as above.Conductivity wasmeasured on a calibrated pH/CON 510 Oaktonanalyser (Model # WD-35610-10). The pH of the ltered watersamples was measured as is. Conductivity measurements of thesamples were taken at full concentration of the water and atregular dilution intervals until a ve-point linear calibrationcurve was achieved. The actual conductivity of the sample wasthus extrapolated from this curve.

2.3. ICP-OES

The as collected water sample was shaken vigorously to inducehomogeneity and pre-ltered three times with a Whatmannlter no. 40 before sample preparation to remove any particu-late matter or non-dissolved matter. The ltered water 0.5 mLwas added to a centrifuge tube, followed by 9.5 mL of 70%concentrated nitric acid was then added to the tube. The tubewas closed, shaken and allowed to digest for 72 h 0.5 mL ofdigested sample was then diluted again with 9.5 mL of 3.0%nitric acid in DI. This sample was then analysed using ICP-OES(Inductively Coupled Plasma Optical Emission Spectrometer)carried out using an Optima 4300 DV spectrophotometer ana-lyser with an AS-93 + auto sampler. Inorganic Ventures ICPmetal cation standard (IV-ICPMS-71A, 10 ppm) was dilutedusing 3.0% HNO3 in DI to 1/10th and 1/100th diluted fraction ofthe original. The ICP standard, the two diluted fractions, andthe 3.0% HNO3 in DI as a blank were used to obtain a four-pointcalibration curve. During analysis all possible wavelengths weremeasured for each atom (Table S1 in ESI†). The wavelengths foreach atom analysed are ranked. If similar trends were observed


for two or more samples that have very similar wavelengths thenthe measurement with the highest wavelength rank was takenand the other one was disregarded.

2.4. Carbon analysis

The as collected water sample jars were shaken vigorously toinduce homogeneity then decanted and pre-ltered three timeswith a Whatmann lter no. 40 before carbon analysis to removeany samples containing visible particulate matter of non-dis-solved matter. The ltered water was then diluted to a 1/100th

fraction using DI. The ltered water was added to glass vials,which had been cleaned as outlined in the above procedure.Samples were analysed using a Shimadzu TOC analyser (Model# TOC-VCSH) using an auto-sampler (Model # ASi-V) in totalcarbon (TC) and non-purgeable organic carbon (NPOC) mode.Replicate samples were run in all cases.

2.5. GC/MS analysis

Two aliquots (500 mL) were withdrawn from the collection jarsfor GC/MS analysis. Before decanting of the two aliquotsamples, the jar was shaken vigorously to achieve homogeneity.One of the samples was adjusted to pH 2 with 37% concentratedHCl to facilitate extraction of organic acids and analysed againstthe non-acidied aliquot for comparison. The aliquots wereeach extracted four times with 25 mL volumes of chloroform.The organic phase was ltered through anhydrous sodiumsulphate. The chloroform extract was transferred to an evapo-ration vessel, which had been cleaned as above. The vessel wascooled to 0 �C and the solvent concentrated under a stream ofnitrogen over 24 h. The concentrated extract was re-diluted withchloroform and transferred to a GC/MS vials for analysis.

Gas chromatography/mass spectroscopy analyses were per-formed on a Agilent Technologies 5973 network mass selectivedetector with a quadrupole mass spectrometer with an AgilentTechnologies 5973 network GC system, using a 30 m DB-1capillary column (0.25 mm I.D., 0.25 mm lm) and helium ascarrier. All samples were taken using a 7693A Automatic LiquidSampler with split/splitless injections of 1.0 mL using an auto-mated injector temperature set at 285 �C. The GC oventemperature was programmed 60 �C to 300 �C (20 �C min�1),and held at 300 �C for 10 min. Peaks were identied automat-ically by using Agilent Technologies Chemstation Sowarelinked with National Institute of Standards and Technologymass spectrum library (NIST 08MS Library). Data was processedusing Automated Mass Spectroscopy Deconvolution and Iden-tication System (AMDIS). The NIST library is searched auto-matically for Chemstation integrated and identied peaks.From this a quality of t for each integrated peak area isestablished. Abundance was initially determined by the use ofexternal standard solutions of acetone/water, and checked by“spiking” the samples with a known aliquot of acetone. Repli-cate samples were run in all cases. Acetone was also used as aninternal standard for all samples to corroborate retention timesand MS tting. In addition, components such as hexane,heptane, decane, and toluene were run as for conrmation.



Table 2 Chemical analysis (mg L�1) of the produced water samplesa

Element Marcellus (PA) Eagle Ford (TX) Barnett (NM)

Na 523.6 45.9 5548.9K 2605.8 17043.3 4566.5Li 0.0 1200.6 84407.4Rb 47.0 0.0 0.0Mg 289.7 28.2 5747.2Ca 1387.5 111.2 33971.8Sr 92.9 34.5 2461.8Ba 0.0 4.7 17.2Ti 0.0 16.2 15.1V 4.2 16.2 14.6Cr 11.0 13.6 11.5Mn 0.0 11.5 9.4Fe 8.4 1246.5 75.7Ni 0.0 36.5 12.0Cu 2.6 0.0 0.0Zn 65.8 0.0 684.9Hg 14.6 0.0 0.0B 0.0 40.2 70.5Si 727.1 4416.6 0.0Sn 206.2 3.1 124.2P 0.0 29.2 1177.1As 26.1 0.0 2.1Sb 0.5 2.1 9.9Bi 36.5 50.1 161.3S 189.0 413.9 290.8

a Cs, Sc, Y, La, and Cd probed but not detected.

Fig. 1 The concentration of the 10 most abundant elements found inthe produced water samples.


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

3. Results and discussion3.1. Conductivity, pH and salt content

Before considering the organic content, we wanted to determinethe inorganic content of the produced waters under study. Aspreviously discussed, the salt content of produced waters havealready been extensively reported.13 Table 1 shows the conduc-tivity and pH of the as-collected water. There is no direct rela-tionship between the conductivity and pH for the samplesindicating that the conductivity is a function of salt content andidentity rather than acidity (see below).

The chemistry of a shale reservoir is unlike that of aconventional oil or gas reservoir that is ushed with hundredsof pore volumes of transient water resulting in leaching of therock and other components to an equilibrium level. Shale has avery low permeability (concrete is 102 to 104 more permeable)and there has been little or no movement of fresh water (orwaters of a different mineral content) since the rock wasformed. Furthermore, shales are under-saturated to water andthe levels of salt in the connate waters within the shales areoen at salinity equal to the seawater the shale was depositedfrom. In other words, shale is a reactor waiting for an inux offresh ingredients, and thus when under saturated fresh water oreven moderate salinity water is introduced during a frac, salts,some organics, and other minerals that were in equilibriumwith the connate waters are solubilized.

The ion content for each of the produced water samples wasdetermined by ICP-OES. The results are summarized in Table 2.The presence of common mineral elements (such as Na, K, Li,Ca, Si, and Fe) is expected depending on the reservoir geology(see below) and salts added to the frac uid; however, some ofthe levels of the transition metals and heavier main groupelements are more surprising. For example the nickel content ofthe Eagle Ford and Barnett waters (36.5 and 12.0 mg L�1,respectively) are comparable to that observed in typical acidmining discharge (AMD) waters from coal production (ca. 1–10mg L�1), while the zinc content of the Barnett water (684.9 mgL�1) is signicantly higher than the typical <100 mg L�1

observed in AMD water.18 It should also be noted that Cr, Hg,and As are above the levels prescribed by the US EPA themaximum contamination levels (MCLs) for drinking water.

For ease of discussion and comparison of the differentsamples, we will consider the 10 most abundant elements in thefollowing discussions. Fig. 1 shows a plot of the concentrationof these elements for each of the produced waters. From Fig. 1and Table 2 it may be seen that the Barnett sample has thehighest inorganic content, followed by Eagle Ford and thenMarcellus. This follows the trend observed in the water

Table 1 Conductivity and pH of as collected produced water samples

Water Conductivity (mS) pH

Marcellus (PA) 28.5 6.85Eagle Ford (TX) 31.1 5.95Barnett (NM) 52.8 7.43


conductivity. In fact, a plot of the total cation content versusconductivity (Table 1) shows a linear correlation (Fig. S1 inESI†).

Consideration of the 10 most abundant elements measuredfor each water sample provides useful information (Fig. 2). Withregard to alkali (Group 1) metals the Marcellus and Eagle Fordsamples are potassium rich (Fig. 2a and b); while the Barnettproduced water is lithium rich (Fig. 2c). Potassium appears tobe associated with silicon, since the relative K : Si ratio issimilar in the two water samples: 3.5 : 1 (Marcellus) and 3.8 : 1(Eagle Ford). Generally, these high alkali metal levels are not anissue with regard to the re-use of the produced water insubsequent hydraulic fracturing. In contrast to the alkalimetals, alkaline earth (Group 2) metals are associated with scale



Fig. 2 Concentration of the ten most abundant elementals found in(a) Marcellus (PA), (b) Eagle Ford (TX), and (c) Barnett (NM) producedwaters.

Table 3 Total carbon (TC), non-purgeable organic carbon (NPOC),and total inorganic carbon (TIC) for produced water samplesa

Produced water TC (mg L�1) NPOC (mg L�1) TIC (mg L�1)

Marcellus (PA) 3808 (82) 2348 (22) 1460 (82)Eagle Ford (TX) 9285 (100) 6095 (300) 3190 (200)Barnett (NM) 58 550 (995) 43 550 (730) 15 000 (862)

a Estimated standard deviations are in parenthesis.

Fig. 3 Carbon content measurements for each produced watersample.

Fig. 4 Representative GC trace for Marcellus produced water afterCHCl3 extraction.


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

formation.19,20 In particular, when calcium and barium levelsare above ca. 20 000 mg L�1 scale inhibitors must be employedand or the salt content lowered before the water can be re-useddown hole.4

It is interesting to note that the samples with the highestcalcium content are closest to neutral pH. Given that CaCO3 isused as a neutralizing agent for AMD and other acidic waters theacidity of produced water in the absence of calcium (i.e., theEagle Ford sample) is expected. Although total dissolved solids(TDS) measured from conductivity21 is ordinarily consideredwhen thinking of scaling, the calcium content is more a causa-tive factor. Based upon these waters, the pH of the water appearsto be a better predictor of scaling potential than conductivity (seeFig. S2 in ESI†). Finally, the presence of strontium (and due totheir lower radioactive isotope percentage, to a lesser extentbarium and calcium) may also be associated with naturallyoccurring radioactive materials (NORM).

3.2. Carbon content

The total carbon (TC), non-purgeable organic carbon (NPOC)also known as total organic content (TOC), and total inorganic


carbon (TIC) for each produced water sample was measured(Table 3) and the results are shown in Fig. 3. For all of theproduced water samples the NPOC is signicantly higher thanthe TIC. However, the ratio of TIC : NPOC range from 0.62 forthe Marcellus produced water to 0.34 for the Barnett producedwater. Thus, not only does the Barnett sample contain signi-cantly more carbon than the other samples but also thepercentage of the carbon due to organic compounds is higher.

3.3. Identication of organic compounds

Fig. 4 shows a representative GC for Marcellus producedwater: additional plots are given in the ESI.† The peakassignment is provided based upon the tting of the inte-grated mass spectrum for each peak. While all the peakscould be assigned a suitable compound, there is a qualityparameter (Q) that provides a goodness of t of the data, i.e.,



Fig. 5 Percentage of peaks in the GC/MS of the produced waters witha quality range based upon the peak number and peak area.


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

a condence level in the assignment. Fig. 5 shows thepercentage of peaks in the GC of the produced waters withina particular quality range of the assignment by mass spec-trometry. The data is shown as a function of the totalnumber of peaks (i.e., fraction of the different compounds)and also the peak area (i.e., the molar fraction of thesample). It should be noted that a low Q is not an indicationthat the peak is incorrectly assigned per se, but a measure ofthe integrated MS's t to the known library (NIST 08 MSLibrary). The full analyses for each water sample, using bothas-received and acid extraction, are given in Tables S2–S7(ESI†). We note that in the work of Orem et al. less than20% of potential organic compounds were actuallyidentied.15

For simplicity in giving a representative example of thetypes of organic compound found in each water sample, wehave limited the contents of Tables 4–6 to those compoundsthat are assigned with condence in more than one wellsample. Tables 4–6 list the saturate, aromatic, and resin andasphaltene (SARA) groupings, respectively. SARA is a tradi-tional analysis method that divides crude oil componentsaccording to their polarizability and polarity.22 The saturatefraction consists of nonpolar material including linear,branched, and cyclic saturated hydrocarbons, in our presentdiscussion we have included unsaturated aliphaticcompounds in the general class of “saturated” compounds.Aromatics are dened as molecules containing one or morearomatic rings, and are slightly more polarizable. Resins andasphaltenes have polar (heteroatom) substituents. Thedistinction between the two is that asphaltenes are insolublein heptane whereas resins are miscible with heptane. As suchthe resins should be observed by as-received (neutral) extrac-tion, while any asphaltenes are more likely to be extractedfrom acid solution.

It is interesting to note that the greater overlap is betweenthe Marcellus and the other two produced water samplesindependently, rather than all three or the Eagle Ford andBarnett. This is despite the Marcellus produced water showingthe lowest NPOC (Fig. 3). We note that the Eagle Ford is


potassium rich (Fig. 2b), while the Barnett is lithium andcalcium rich (Fig. 2a). In contrast, the Marcellus has bothpotassium and calcium. It would be interesting to determinewith further analysis as to whether there is a relationshipbetween the salt content and the particular organic compoundsfound in the produced water.

Hetero compound: natural versus additives. The presence ineach of the produced water samples of a wide range of satu-rated and aromatic compounds is expected given the natureand composition of oil. In addition, it is expected thatsubstituted compounds such as carboxylic acids and otheroxidation products be observed given their relationship tomaterials such as humic and fulvic acids that is produced bybiodegradation of dead organic matter.23 In particular, theobservation of 10,4-dihydroxy-70-methoxy-2,30-dimethyl-,(�)-[1,20-binaphthalene]-5,50,8,80-tetrone (I) in both the Mar-cellus and Barnett produced water is related to the types ofstructures proposed to humic acids.23 However, some of thecompounds observed are man-made in origin and are relatedto drilling uids, frac uids, or tracers. Conversely, some of thehetero substituted compounds observed must be the result ofsecondary reaction chemistry due to water treatment ratherthan their presence in either the connate waters or the fracuid.

The presence of various fatty acid phthalate esters (II) inthe Barnett and Marcellus produced waters can be related totheir use in drilling uids and breaker additives,24 while thedioctadecyl ester of phosphate phosphoric acid[(C18H37O)2P(O)2] found in the Marcellus produced water is acommon lubricant. Thus, these chemicals are most likelyresidues from the well drilling and hydraulic fracturing oper-ations, and would be expected to be present in frac owbackor in the earlier stages of production. The lack of suchchemicals in the Eagle Ford sample is either due to the rela-tive age of the well or the use of other chemicals by the driller.We note that the presence of the uorinated compounds2,2,3,3,4,4,4-heptauorobutyl undecylate [2,2,3,3,4,4,4-hepta-uoro-butanoic acid undecylate ester, C3F7C(O)OC11H24] in theMarcellus and Eagle Ford, cis-4-ethyl-5-octyl-2,2-bis(tri-uoromethyl)-1,3-dioxolane (III) in the Barnett, and tri-uoromethy tetradecylate [triuoroacetic acid tetradecyl ester,CF3C(O)OC14H29] in the Eagle Ford, are commonly employedas ow tracers.



Table 4 Selected compounds (present in at least two well samples) identified in the CHCl3 extraction of produced water

Formula Compound name Produced water source

C5H10 2-Methyl-1-butene Marcellus, Eagle Ford2-Methyl-2-butene Marcellus, Eagle Ford1,1-Dimethylcyclopropane Marcellus, Barnett

C6H14 Hexane Marcellus, BarnettC7H14 Methylcyclohexane Marcellus, Barnett

1,3-Dimethylcyclopentane Marcellus, Barnett(E)-3-Heptene Marcellus, Barnett

C8H16 2-Octene Marcellus, Barnett1,3-Dimethylcyclohexane Marcellus, BarnettEthylcyclohexane Marcellus, BarnettPropylcyclopentane Marcellus, Barnett1,2,4-Trimethylcyclopentane Marcellus, Barnett1,1,3-Trimethylcyclopentane Marcellus, Barnett3-Methyl-1-heptene Marcellus, Barnett

C8H18 2-Methyl-heptane Marcellus, BarnettC9H18 Propyl-cyclohexane Marcellus, Barnett

1,1,3-Trimethylcyclohexane Marcellus, Barnett1-Ethyl-4-methylcyclohexane Marcellus, Barnett1,2,4-Trimethylcyclohexane Marcellus, Barnett

C9H20 2-Methyloctane Marcellus, BarnettNonane Marcellus, Barnett2,6-Dimethylheptane Marcellus, Eagle Ford

C10H20 4-Propyl-3-heptene Marcellus, BarnettC10H22 Decane Marcellus, Barnett

2,6-Dimethyloctane Marcellus, Barnett2,2,3,3-Tetramethylhexane Eagle Ford, Barnett4-Methylnonane Marcellus, Barnett

C11H22 2-Ethyl-1,1,3-trimethylcyclohexane Marcellus, BarnettPentylcyclohexane Marcellus, Barnett

C11H24 Undecane Marcellus, Barnett4-Methyldecane Marcellus, Barnett2-Methyldecane Marcellus, Barnett3-Methyldecane Marcellus, Barnett

C12H26 Dodecane Marcellus, Eagle Ford, Barnett4-Methylundecane Marcellu, Barnett

C13H26 1-Pentyl-2-propylcyclopentane Marcellus, BarnettC13H28 Tridecane Marcellus, Eagle Ford, Barnett

2,6-Dimethylundecane Marcellus, Barnett2,4-Dimethyl-undecane Marcellus, Eagle Ford

C14H28 Cyclotetradecane Marcellus, Eagle FordC14H30 Tetradecane Marcellus, Eagle Ford, Barnett

2-Methyl-Tridecane Marcellus, Eagle Ford7-Methyltridecane Marcellus, Barnett4-Methyltridecane Marcellus, Eagle Ford

C15H32 Pentadecane Marcellus, Eagle Ford, Barnett2,6,10-Trimethyldodecane Marcellus, Eagle Ford, Barnett2,6,11-Trimethyl-dodecane Marcellus, Barnett5-Methyltetradecane Eagle Ford, Barnett

C16H32 1-Hexadecene Marcellus, BarnettC16H34 Hexadecane Marcellus, Eagle Ford, BarnettC17H36 Heptadecane Marcellus, Eagle Ford, BarnettC18H36 1-Octadecene Marcellus, BarnettC18H38 Octadecane Marcellus, Eagle Ford, Barnett

2,6,10-Trimethylpentadecane Marcellus, BarnettC19H40 Nonadecane Marcellus, Eagle Ford, Barnett

2,6,10,14-Tetramethylpentadecane Marcellus, BarnettC20H40 1,7,11-Trimethyl-4-(1-methylethyl)-cyclotetradecane Marcellus, BarnettC20H42 Eicosane Marcellus, Eagle Ford, Barnett

2,6,10,14-Tetramethylhexadecane (phytane) Marcellus, Barnett10-methylnonadecane Marcellus, Eagle Ford

C21H44 Heneicosane Marcellus, Eagle Ford, BarnettC22H46 Docosane Marcellus, Eagle Ford, BarnettC23H46 9-Tricosene Marcellus, Barnett

This journal is © The Royal Society of Chemistry 2014 Environ. Sci.: Processes Impacts, 2014, 16, 2237–2248 | 2243


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online


Table 4 (Contd. )


C23H48 Tricosane Marcellus, Eagle Ford, BarnettC24H50 Tetracosane Marcellus, BarnettC25H52 Pentacosane Marcellus, Eagle Ford, BarnettC26H54 Hexacosane Marcellus, Barnett

11-(1-Ethylpropyl)-heneicosane Eagle Ford, BarnettC27H56 Heptacosane Marcellus, BarnettC28H58 Octacosane Marcellus, Eagle Ford, BarnettC29H60 Nonacosane Eagle Ford, BarnettC30H62 Triacontane Eagle Ford, BarnettC32H66 Dotriacontane Eagle Ford, BarnettC35H70 17-Pentatriacontene Marcellus, Eagle Ford, BarnettC35H72 Pentatriacontane Marcellus, Eagle Ford, BarnettC36H74 Hexatriacontane Marcellus, Eagle Ford, BarnettC43H88 Tritetracontane Marcellus, Eagle Ford, BarnettC44H90 Tetratetracontane Eagle Ford, Barnett

Table 5 Selected aromatic compounds (present in at least two wellsamples) identified in the CHCl3 extraction of produced water


C7H8 Toluene Marcellus, BarnettC8H10 Ethylbenzene Marcellus, BarnettC8H18 2-Methyl-heptane Marcellus, BarnettC9H12 1,2,3-Trimethylbenzene Marcellus, Barnett

1,2,4-Trimethylbenzene Marcellus, Barnett1-Ethyl-3-methyl-benzene Marcellus, Barnett

C10H14 1,2,4,5-Tetramethylbenzene Marcellus, Barnett1-Methyl-3-propylbenzene Marcellus, Barnett

C11H20 Decahydro-2-methyl-naphthalene Marcellus, BarnettC12H12 1,5-Dimethyl-naphthalene Marcellus, Barnett

2,7-Dimethyl-naphthalene Marcellus, BarnettC13H14 1,6,7-Trimethylnaphthalene Marcellus, Barnett

2,3,6-Trimethylnaphthalene Marcellus, BarnettC14H18 1,2,3,4-Tetramethyl-naphthalene Marcellus, Barnett

Table 6 Selected resin and asphaltine compounds (present in at least tw

Formula Compound name

C10H18O 3,7-Dimethyl-7-octenalC15H14O 10,11-Dihydro-5H-dibenzo[a,d]cyclohepten-5-olC16H30O2 13-Tetradecen-1-ol acetateC16H32O 2-TetradecyloxiraneC16H33Cl 1-Chloro-hexadecaneC18H37Cl 1-ChlorooctadecaneC18H38S 1-OctadecanethiolC22H34O4 1,2-Benzenedicarboxylic acid, 1-butyl 2-(8-methylnoC24H18O8 10,4-Dihydroxy-70-methoxy-2,30-dimethyl-,(�)-[1,20-biC24H38O4 1,2-Benzenedicarboxylic acid, 1,2-bis(2-ethylhexyl) eC27H55Cl 1-Chloro-HeptacosaneC28H46O4 1,2-Benzenedicarboxylic acid, 1,2-didecyl ester

1,2-Benzenedicarboxylic acid, 1,2-bis(8-methylnonyC32H66O 1-DotriacontanolC54H108Br2 1,54-Dibromo-tetrapentacontane



Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

Relative distribution of organic compounds. Neutral versusacid extraction should provide an indication of the relativepolarity of the hydrocarbons. That is, extraction from neutralwater should advantage non-polar hydrocarbons as they havethe lowest solubility in water. Acid extraction should enhance

o well samples) identified in the CHCl3 extraction of produced water

Produced water source

Marcellus, BarnettMarcellus, BarnettMarcellus, BarnettMarcellus, BarnettMarcellus, Eagle FordMarcellus, BarnettMarcellus, Barnett

nyl) ester Marcellus, Barnettnaphthalene]-5,50,8,80-tetrone Marcellus, Barnettster Marcellus, Barnett

Eagle Ford, BarnettMarcellus, Barnett

l) ester Marcellus, BarnettEagle Ford, BarnettMarcellus, Eagle Ford, Barnett



Fig. 7 The relative composition of organic molecules identified inproduced water as a function of carbon content: (a) neutral and (b)acid extraction.


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

the removal (and therefore analysis) of polar hydrocarbons withhigher solubility in non-acidic water. Generally this latter cate-gory will include carboxylic acid, amine, and nitratefunctionality.

Fig. 6 shows the normalized saturate, aromatic, resin andasphaltene (SARA) composition for each of the produced waters.In each case the majority of organics are saturated, and only asmall fraction come under the other three categories. This is incontrast to the prior analysis of coalbed produced water,15

which showed signicant concentrations of polyaromatichydrocarbons (PAH). Given that the source of PAHs in coalbedmethane produced water is thought to be due to their leachingfrom the coal, it is unsurprising that low levels should be seenin shale derived produced water. Further analysis shows thatthe aromatic compounds (Tables S2–S7†) are exclusivelysubstituted benzene derivatives, rather than PAHs. The lack ofhazardous (carcinogenic) PAHs, and generally low aromaticcontent, in shale produced water is a positive result andsignicantly lowers the toxic effects of the water compared tocoalbed methane produced water15 and off-shore producedwaters from conventional oil and gas production.10,11 The lowlevels of resins and asphaltenes are also consistent with the“mature” nature of a gas reservoir as compared to coal (and to alesser extent oil) formations since condensates are virtuallydevoid of asphaltenes.

An alternative differentiation of the organic compoundsidentied is by carbon content (i.e., Cn). Fig. 7 shows thenormalized composition for each of the produced watersgrouped into four categories of carbon content: C1–C5, C6–C10,C11–C16, and C17–C30. An observation of higher Cn is consistentwith lower volatility of higher molecular weight hydrocarbons.

Fig. 6 The relative saturate, aromatic, resin and asphaltene (SARA)composition for produced water: (a) neutral and (b) acid extraction.


As may be seen from Fig. 7a, the as-received (“neutral”)extraction fromMarcellus and Barnett produced waters containpredominantly C6–C16 hydrocarbons, while the Eagle Fordproduced water shows the highest concentration in the C17–C30

range. This is consistent with the greater resin/asphaltenecontent for the Eagle Ford produced water (Fig. 6).

Acid extraction shows increased fraction with increased Cn

for all the water samples, suggesting that there is a higherpercentage of high molecular weight polar molecules. Thissuggests that the non-acidied results for the Eagle Fordmay bepartially as a consequence of the higher acidity (lower pH).Furthermore the lack of difference between as-received andacidied extraction suggest that organic species are weaklyacidic since they are almost all protonated at pH ¼ 5.95. This isconsistent with the identication of the heteronuclear organiccompounds in the Eagle Ford produced water (see above),which shows them to consist of mainly fatty alcohols and ethers(and halocarbons) with no carboxylic acids identied (seeTables S4 and S5 in ESI†).

As is discussed above, the major category of organiccompounds identied is the saturate group. In order tounderstand the composition of this group they can be furthersubdivided into linear, branched, and cyclic structures (Fig. 8).The cyclic compounds in this group do not include aromaticcompounds; however, alkenes and alkynes are included in thisgrouping. It is interesting to note that the relative order followslinear > branched > cyclic, except in the case of the Eagle Fordand Marcellus water under acidic extraction conditions(Fig. 8b). It is also interesting that the presence of cyclicaliphatic hydrocarbons in the Eagle Ford produced water is



Fig. 8 The relative composition of aliphatic compounds identified inproducedwater as a function of structure: linear, branched, and cyclic:(a) neutral and (b) acid extraction.

Fig. 9 The relative composition of heteroatom compounds identifiedin produced water as a function of the heteroatom: (a) neutral and (b)acid extraction.


Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

coincidental with the lack of aromatic hydrocarbons (unsatu-rated cyclic compounds) (see Fig. 6b), suggestive of a higherthermal maturity of the Eagle Ford hydrocarbon.

As noted above the resin and asphaltene componentscontain polar (heteroatom) substituents; however, NSOs (N-, S-,and O-containing heterocyclic compounds) can account forother derivatives identied. Fig. 9 shows the normalizedcomposition for each of the produced waters grouped intocategories of heteroatom containing compounds. The as-received extraction shows that most of the heteroatom-substituted hydrocarbons contain oxygen (Fig. 9a). Chemicalcharacterization (Table 6) shows that these are predominantlyfatty alcohols, esters, and ethers. As discussed above, these areall associated with the biodegradation of dead organic matter,and, as such, it is reasonable to expect that they will be themajor heteronuclear organic components.23 The presence ofuorinated organic molecules for the Eagle Ford water isconsistent in that the produced water contains peruorocarbontracers (PFTs) that are commonly used for oil reservoirsmapping (see above).25 The difference between the oxygencontaining compound content of the Eagle Ford samplecompared to either the Barnett or Marcellus samples is high-lighted with the acid extraction (Fig. 9b). This is consistent withthe high percentage of phthalate esters in the Barnett andMarcellus samples (see above).

Acid extraction (Fig. 9b) enhances the detection of the otherhalogenated (chloro and bromo) organics. It is important tonote that unlike uorocarbons, chlorocarbons or organo-bromides are not used in drilling or frac uid or as tracers. Thisindicates that these chemicals are sourced from the reservoir


material rather as a result of man-made pollution. So where dothese compounds come from?

A consideration of the chlorocarbons and organobromidesobserved shows they are related to the alkene, alcohol orcarboxylic acid also present. Anti-Markovnikov addition to 1-octadecene (present in each water sample) would allow for 1-bromooctadecane. However, 1-bromooctadecane or 1-chlor-ooctadecane can be prepared from alcohols or the carboxylicacids (via halodecarboxylation) directly by the reaction ofbromide salts. Both sodium and calcium bromide are both usedas a drilling uid additive.26,27 In addition, 1-bromooctadecanecould be formed from 1-chlorooctadecane using LiBr or NaBr(especially in the case of a phase transfer reaction that wouldoccur at a oil–water interface).28 It has been previously reportedthat chlorocarbons may be formed during oxidative chlorina-tion of waste waters, and that subsequent interaction withbromide salts results in organobromide formation.29 The pres-ence of chlorocarbons could cause a potential issue sincechlorocarbons are considered an environmental pollutant. Aswith chlorocarbons, the presence of quantities of organo-bromides has a potential health effect since they cause mal-functioning of the nervous system and cause damage to organssuch as liver, kidneys, lungs, and cause stomach and gastroin-testinal issues.30

4. Conclusions

With the understanding that the present report is part of awider analysis, the goal of our study was to provide a detailedbaseline in the future mapping of produced water associatedwith shale plays. Although the samples used in the present




Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

study are from shale wells in the US, the processes and meth-odology, as well as the lessons learned, are directly applicable toother regions. In particular the initiation of hydraulic fracturingin the UK warrants a consideration of these results, especiallywith respect to shale gas produced water treatment. If hydraulicfracturing is going to occur in the UK then possibly the mostimportant issue should be how the risk of environmentalcontamination should be reduced.31 Thus, an important ques-tion can be posed: What are the implications for shale gasproduced water treatment?

In presenting the variation in composition and particularchemicals analysed, we canmake a number of statements aboutproduced water from shale gas wells with regard to re-use anddisposal, as well as guide researchers in future treatmentefforts. In particular, given the salt content correlates with theconcentration of hydrocarbon (c.f., Tables 2 and 3), it will beimportant to determine if the salt content and ion ratios withinactual shale connate waters provide a link to explaining thepresence of higher concentrations of some of the organics thatare observed.

In shale, the ow mechanism is different from conventionalreservoirs. Shale is extremely low permeability and uid owthrough these reservoirs has not occurred. Thus, the equilib-rium within shale is radically different from conventionalreservoirs that have been exposed to water leaching from tran-sient water and bacterial activity over geologic time. The lowpermeability nature of shale means that when they are intro-duced to under-saturated water solubilisation will occur of salts,organics and other minerals that were at equilibrium in theconnate waters. Aer fracture treating shale, the recoveredwater volume peaks in a few days (5 to 15 days is common) andthen sharply declines. This short-lived burst of water recoverydescribes a rapidly changing ow environment. When thesalinity prole is compared, the produced water quickly movesfrom: a base frac uid to a mixture of base fracturing uid andconnate water to a rich leachate and then sharply decreases involume, salt and mineral content to a purer connate composi-tion with very low recovered volumes.

Our data highlights a few issues that should be taken intoaccount when developing treatment strategies for frac andproduced waters to allow for their reuse. The rst issue is thatunlike coal bed methane produced waters, no polyaromatichydrocarbons (PAH) are observed meaning that the generaltoxicity of the produced (frac owback) water is potentiallylower for shale plays. This is important since many PAHs haveserious health effects in the >0.2 mg L�1 level. The second issueis that while many of the chemicals used in the frac process areidentied as being present in the produced water, they are lowerin concentration than those that were naturally present in theconnate waters within the shale. Thus, subsequent treatment ofshale produced (and owback) water should concentrate on theremoval of saturates rather than the remaining aromatic, resinand asphaltene groupings. This analysis should also be used asa guide for creating a “standard” or “idealized” shale producedwater so that the efficacy of any treatment may be measured.

The presence of chlorocarbons and organobromides is, webelieve, most probably due to the interactions of produced


water with the salts used in hydraulic fracturing and drilling orthe chemical treatments (and pre-treatment) of produced water.At present, due to the high bacteriological content of naturalwaters (and also produced water upon standing in ponds whichis blended with “fresh” water), chemical treatment is appliedbefore it can be used (re-used) for hydraulic fracturing. In manycases this involves chlorine dioxide or hypochlorite treatments.Such treatments will enhance the active species present toconvert the naturally occurring hydrocarbons to chlorocarbonsand organobromides. Further studies will be required todetermine if the reactions are occurring downhole or duringtreatment of produced water. In either case this would suggestthat chemical treatments should be limited since they cause theformation of unwanted non-naturally occurring compounds;treatments involving separation are preferred on long-termsafety grounds. We believe that analysis of the detailedcomposition of the produced waters also highlights the prob-lems in future treatment protocols, in particular the develop-ment of a process that allows for the removal of a wide range oforganic compounds.16,32–34

Acknowledgements

This work was supported by the Robert A. Welch Foundation (C-0002) and the Welsh Government Ser Cymru Programme. Theauthors acknowledge the encouragement and advice of JakeDavis and Waymon Votaw (Lance Energy Services).

References

1 P. Jaramillo, W. M. Griffin and H. S. Matthews, Environ. Sci.Technol., 2007, 41, 6290–6296.

2 A. Burnham, J. Han, C. E. Clark, M. Wang, J. B. Dunn andI. Palou-Rivera, Environ. Sci. Technol., 2012, 46, 619–627.

3 R. W. Howarth, R. Santoro and A. Ingraffea, Clim. Change,2011, 106, 679–690.

4 K. B. Gregory, R. D. Vidic and D. A. Dzombak, Elements, 2011,7, 181–186.

5 J. A. Dammel, J. M. Bielicki, M. F. Pollak and E. J. Wilson,Environ. Sci. Technol., 2011, 45, 5075–5086.

6 G. E. King, SPE Annual Technical Conference, Florence, Italy,September 19–22, 2010, Abstract SPE 133456.

7 G. T. Tellez and N. Nirmalakhandan, Produced Water:Technological/Environmental Issues and Solutions, ed. J. P.Ray and F. R. Englehardt, Plenum, New York, pp 522–523.

8 B. G. Rahm and S. J. Riha, Environ. Sci.: Processes Impacts,2014, 16, 1400–1412; B. G. Rahm and S. J. Riha, Environ.Sci. Policy, 2012, 17, 12–23.

9 C. A. Menzie, Environ. Sci. Technol., 1982, 16, 454A–472A.10 H. W. Armstrong, K. Fucik, J. W. Anderson and J. M. Neff,

Mar. Environ. Res., 1979, 2, 55–69.11 J. K. Andreasen and R. W. Spears, Bull. Environ. Contam.

Toxicol., 1983, 30, 277–283.12 T. Sirivedhin and L. Dallbauman, Chemosphere, 2004, 57,

463–469.13 A. S. Andrew, D. J. Whitford, M. D. Berry, S. A. Barclay and

A. M. Giblin, Appl. Geochem., 2005, 20, 727–747.




Publ

ishe

d on

13

Aug

ust 2

014.

Dow

nloa

ded

on 0

7/11

/201

4 14

:16:

55.

View Article Online

14 A. Fakhru'l-Razi, A. Pendashteh, L. Chuah Abdullah,D. Radiah Awang Biak, S. Siavash Madaeni and Z. ZainalAbidin, J. Hazard. Mater., 2009, 170, 530–551.

15 W. H. Orem, C. A. Tatu, H. E. Lerch, C. A. Rice, T. T. Bartos,A. L. Bates, S. Tewalt andM. D. Corum, Appl. Geochem., 2007,22, 2240–2256.

16 S. J. Maguire-Boyle and A. R. Barron, J. Membr. Sci., 2011,382, 107–115.

17 E. A. Witter and A. D. Jones, Environ. Toxicol. Chem., 1999,18, 1920–1926.

18 N. F. Gray, Water Res., 1998, 32, 2122–2134.19 H. A. Stiff Jr and L. E. Davis, J. Pet. Technol., 1952, 4, 213–216.20 O. J. G. Vetter and R. C. Phillips, J. Pet. Technol., 1970, 22,

1299–1308.21 E. A. Atekwana, E. Atekwana, R. S. Rowe, D. D. Werkema and

F. D. Legall, Journal of Applied Geophysics, 2004, 56, 281–294.22 O. C. Mullins, E. Y. Sheu, A. Hammami and A. G. Marshall,

Asphaltenes, Heavy Oils, and Petroleomics, Springer, New YorkCity, 2007.

23 F. J. Stevenson, Humus Chemistry: Genesis, Composition,Reactions, John Wiley & Sons, New York, 1994.

24 R. Hofer and J. Bigorra, Green Chem., 2007, 9, 203–212.25 S. N. Davis, G. M. Thompson, H. W. Bentley and G. Stiles,

Groundwater, 1980, 18, 14–23.


26 G. Dakka and Y. Sasson, Tetrahedron Lett., 1987, 28, 1223–1224.

27 J. K. Kochi, J. Am. Chem. Soc., 1965, 87, 2500–2502.28 Y. Sasson, M. Wiess, A. Loupy, G. Bram and C. Pardo, J.

Chem. Soc., Chem. Commun., 1986, 1250–1252.29 C. D. Volz, K. Ferrar, D. Michanowicz, C. Christen,

S. Kearney, M. Kelso, and S. Malone, Contaminantcharacterization of effluent from Pennsylvania brine treatmentInc., Josephine facility being released into Blacklick Creek,Indiana County, Pennsylvania, 2011, http://go.nature.com/5otd59.

30 A. Luch andW.M. Baird, The Carcinogenic Effects of PolycyclicAromatic Hydrocarbons, ed. A. Luch, Imperial College Press,London, 2004, pp. 9–96.

31 P. Ziemkiewicz, J. D. Quaranta and M. McCawley, Environ.Sci.: Processes Impacts, 2014, 16, 1692–1699.

32 P. Xu and J. E. Drewes, Sep. Purif. Technol., 2006, 52, 67–76.33 S. Mondal and S. R. Wickramasinghe, J. Membr. Sci., 2008,

322, 162–170.34 F.-R. Ahmadun, A. Pendashtech, L. C. Abdullah,

D. R. A. Biak, S. S. Madaeni and Z. Z. Abidin, J. Hazard.Mater., 2009, 170, 530–551.



News & Politics

Study: Organic compounds in produced waters from shale gas wells