3 8 8 A A N A LY T I C A L C H E M I S T R Y / N O V E M B E R 1 , 2 0 0 4 © 2 0 0 4 A M E R I C A N C H E M I C A L S O C I E T Y

a c d e t e c t i v e

Asurvey of drinking-water wells inwestern Texas, which began inthe summer of 2002, turned up

some surprising results. Perchlorate, amain ingredient in rocket fuel, was foundin >80% of the wells tested over an area of~60,000 square miles. The levels of per-chlorate varied—most were quite low, but~25% were ≥4 ppb, the level many consid-er to be of concern. The Texas Commis-sion on Environmental Quality (TCEQ)couldn’t figure out where the perchloratewas coming from.

TCEQ turned to researchers at TexasTech University (TTU) for help. Envi-ronmental engineer Andrew Jackson, an-alytical chemist Purnendu “Sandy” Das-gupta, and their colleagues have beentrying to solve the mystery ever since.

It has been, and still is, a huge effort.“We looked at vertical distribution inthe aquifer and at unsaturated soils. Welooked at land use in the area to see ifthere was any correlation with a certainindustry or practice. We looked at a num-ber of other constituents to see if we couldfind correlations with the occurrence ofperchlorate. We did some modeling anddetermined in which geologic features itoccurred,” says Jackson. “What it comesdown to in our best guess is that whatwe are dealing with is natural perchlo-rate,” he says.

Most of the media attention sur-rounding perchlorate has focused onrocket-fuel contamination. The Texaspanhandle wells, however, are nowherenear any source of munitions, says Das-gupta. Other sources, such as roadside

flares, fireworks, and explosives, have alsobeen ruled out. Analytical methods fordetecting perchlorate have become sosensitive that the chemical is turning upjust about everywhere people look, saysDasgupta. The question now is: Howmuch of it is manmade?

Stable-isotope analysisAlthough scientists are fairly certain thatthe perchlorate in the Texas wells is ofnatural origin, they still need more evi-dence to prove it. Researchers at Loui-siana State University (LSU) and OakRidge National Laboratory (ORNL)have developed a new stable-isotoperatio technique that could provide themissing piece of information. “We cantell easily whether it is a natural sourceor not by looking at both the 17O/16Oand 18O/16O ratios,” says geochemistHuiming Bao of LSU.

In the new approach, all three stableoxygen isotopes are measured after athermal decomposition method gener-ates O2 from perchlorate crystals. Mea-suring just 18O/16O won’t give you theanswer, because 18O/16O values for an-thropogenic and atmospheric perchlo-rate overlap, says Bao. But if you mea-sure 17O/16O at the same time andcompare the two ratios, you will be ableto tell, he says. Most oxygen-containingcompounds on earth have a correlationbetween 18O/16O and 17O/16O. Anydeviation from this relationship is re-ferred to as the 17O anomaly, or �17O.

In photochemical reactions, 18O/16Oand 17O/16O often do not follow this

mass-dependent isotope fractionationrelationship; there is a large deviationand a positive �17O. Atmospheric ozone(O3) is known to have a highly positive�17O. If atmospheric O3 is involved inthe creation of perchlorate, then naturalperchlorate would also have a highly posi-tive �17O. Although the researchers onlyhave a limited data set, “So far all naturalperchlorate has shown this 17O anomaly,”says Bao.

With the help of environmental chem-ist Baohua Gu of ORNL, Bao and col-leagues measured the three stable oxygenisotopes in perchlorate extracted from ahandful of Chilean salt deposits, whichare known to contain trace levels of per-chlorate. In all the samples tested, theyfound that the perchlorate had a uniqueoxygen isotope signature. Gu is now col-laborating with Neil Sturchio of the Uni-versity of Illinois at Chicago to analyze37Cl/35Cl in the samples. So far, naturalperchlorate also appears to have a uniquechlorine isotope signature, says Gu.

To perform the isotopic analyses, theresearchers need a minimum of 0.1 mgof pure crystallized perchlorate. One ofthe biggest challenges was extracting andanalyzing trace levels of perchlorate fromthe deposits, which also contain largeamounts of nitrate and other impurities,says Bao. And that is where Gu’s exper-tise came in handy.

Several years ago, Gu and colleaguesat ORNL developed an ion-exchangeresin to remove parts-per-trillion levels ofradioactive pertechnetate from ground-water. They later demonstrated that the

Tracing the origin of perchlorateA main ingredient in rocket fuel is showing up almost everywhere researchers look, but where is it all coming from?

N O V E M B E R 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y 3 8 9 A

a c d e t e c t i v e

resin also worked well for removing per-chlorate. “The resin is so selective that itpicks up perchlorate even in the presenceof high concentrations of nitrate, sulfate,or chloride in water,” says Gu. The re-searchers eventually developed a proce-dure for removing the perchlorate fromthe resin so that they could reuse it. “Per-chlorate is so strongly [ad]sorbedyou cannot remove it with a sodiumchloride brine solution,” says Gu.Instead, the researchers used a ferricchloride/hydrochloric acid mixture,which resulted in ~99% recovery ofthe perchlorate. That same proce-dure turned out to be useful forconcentrating perchlorate for iso-topic analysis.

Jackson and colleagues at TTUare currently pumping perchlorate-contaminated groundwater fromwestern Texas through columnscontaining the special resin so thatthey can obtain stable-oxygen-isotope signatures. “It’ll say whetherthe perchlorate is anthropogenic,but it won’t be the nail in the cof-fin,” says Jackson. We cannot positive-ly rule out Chilean fertilizer, which isknown to contain ~0.1% perchlorate, asthe source, he says. However, it is high-ly unlikely because perchlorate has beendetected in several areas where there hasnever been any agricultural activity, headds. All the evidence points to perchlo-rate being generated in the atmosphereor by surface oxidation, he says.

The Texas researchers are also tryingto produce perchlorate in the laboratoryby processes that mimic those in the loweratmosphere or on surfaces such as sand.“We are looking at ozone reactions andUV reactions with chloride in aerosolsand in sands,” says Jackson. Dasgupta’sgroup is investigating mechanisms thatinvolve electrical discharge or lightning.Although the researchers have been ableto produce significant quantities of per-chlorate using various techniques, exact-ly how perchlorate forms in nature is stillelusive. “It is all speculation,” says Bao.“What we can provide is a stable isotopeconstraint. If you propose a model, ourdata provide two constraints. One isthe 18O/16O ratio. The other is �17O,”he says.

The next step is to obtain more sta-ble-oxygen-isotope data on samples thatare likely to contain natural perchlorate,says Bao. He and his colleagues intendto analyze more soils, including soilfrom southern California, the Nevadadesert, and possibly other deserts. “Aslong as we have enough of those data,

we can go ahead and put a very goodconstraint in terms of the formationpathway,” says Bao.

The regulatory sideIf some of the perchlorate found in drink-ing water really is of natural origin, itcould affect how the U.S. EnvironmentalProtection Agency (EPA) decides to reg-ulate the compound in drinking water.Although EPA recommended a limit of1 ppb in 2002, it is waiting for the Na-tional Academies’ Institute of Medicineto finish reviewing the issue before it ac-tually sets a perchlorate limit for drink-ing water. Experts predict that the levelwill be >1 ppb.

Meanwhile, the U.S. Food and DrugAdministration (FDA) has become con-cerned about perchlorate in the food sup-ply. It all started when a nonprofit envi-ronmental group released a report in thespring of 2003 showing that lettuce irri-gated with contaminated water in Califor-nia and Arizona contained detectable lev-els of perchlorate. “We found that 4 outof 18 lettuce samples contained perchlo-rate,” recalls Dasgupta, who performedthe analyses for the environmental group.

“It was really not a scientifically largeenough sample,” he says. Nonetheless,the report received a lot of attentionfrom the press.

Perchlorate was once thought to besafe because it is not very reactive. Now,however, many consider it to be a cumu-lative toxin because it interferes with the

transport of iodide in the body.Without iodide, the thyroid glandwill not function properly. Thyroidhormones are particularly impor-tant for infants and young chil-dren, who rely on them for prop-er development.

Because kids don’t typically eatlettuce, Dasgupta and his graduatestudent Andrea Kirk decided toinvestigate the occurrence of per-chlorate in milk. In the summerof 2003, they analyzed seven milksamples purchased randomly fromLubbock, Texas, supermarkets. Allof them contained perchlorate, atlevels ranging from 1.7 to 6.4 ppb.

The results prompted FDA totake a closer look at perchlorate in

the U.S. food supply. In December 2003,the agency announced its plans to ana-lyze lettuce, bottled drinking water, andmany other foods, as appropriate samplepreparation methods become available.

Although the final report from FDA isnot expected until next year, some of theearly results were published in the Sep-tember 15 issue of Analytical Chemistry(2004, 76, 5518–5522). In that paper,Alexander Krynitsky and colleagues de-scribe a new ion chromatography (IC)/MS/MS method for analyzing perchlo-rate in lettuce, cantaloupe, bottled water,and milk. The method incorporates an18O4-labeled perchlorate internal stan-dard, which corrects for matrix effectsthat can enhance or suppress the MS sig-nal. According to the researchers, the la-beled internal standard greatly simplifiedthe extraction procedure and cleanupprior to IC/MS/MS analysis.

The analysis of perchlorate in foodsis still in its infancy, and some technicalhurdles need to be overcome before theresults are meaningful. Only time willtell how widespread perchlorate has be-come in our food chain. a

—Britt E. Erickson

Mystery in the desert. Nitrate salt deposits from the Ataca-ma Desert in Chile are known to contain trace levels of nat-urally occurring perchlorate. Researchers are still trying todetermine how the perchlorate got there.

CO

UR

TES

Y O

F H

UIM

ING

BA

O, L

OU

ISIA

NA

STA

TE U

NIV

ER

SIT

Y