A new laser-based, field-deployable analyzer for laboratory-class stable isotope measurements in water

A new laser-based, field-deployable analyzer for laboratory-class stable isotope measurements in water
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Demonstration of high-precision continuousmeasurements of water vapor isotopologues in laboratoryand remote field deployments using wavelength-scannedcavity ring-down spectroscopy (WS-CRDS) technology y Priya Gupta 1 *, David Noone 2,3 , Joseph Galewsky 4 , Colm Sweeney 3,5 andBruce H. Vaughn 6 1 Picarro, Inc., 480 Oakmead Pkwy, Sunnyvale, CA 94085, USA 2 Department of Atmospheric and Oceanic Sciences, University of Colorado, Boulder, CO 80309, USA 3 Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO 80309, USA 4 Department of Earth and Planetary Sciences, University of New Mexico, Albuquerque, NM 87131, USA 5 National Oceanographic and Atmospheric Administration-Earth System Research Laboratory, 325 Broadway, Boulder, CO 80304, USA 6 Institute of Alpine and Arctic Research, University of Colorado, Boulder, CO 80309, USA Received 16 February 2009; Revised 24 April 2009; Accepted 27 April 2009 This study demonstrates the application of Wavelength-Scanned Cavity Ring-Down Spectroscopy(WS-CRDS) technology which is used to measure the stable isotopic composition of water. Thisisotopic water analyzer incorporates an evaporator system that allows liquid water as well as water vaportobemeasuredwithhighprecision.TheanalyzercanmeasureH 182  O,H 162  OandHD 16 Ocontentof the water sample simultaneously. The results of a laboratory test and two field trials with thisanalyzer are described. The results of these trials show that the isotopic water analyzer gives precise,accurate measurements with little or no instrument drift for the two most common isotopologues ofwater. In the laboratory the analyzer has a precision of 0.5 per mil for   d D and 0.1 per mil for  d 18 Owhichissimilartotheprecisionobtainedbylaboratory-basedisotoperatiomassspectrometers.In the field, when measuring vapor samples, the analyzer has a precision of 1.0 per mil for   d D and0.2 permil for   d 18 O. Theseresults demonstrate that the isotopic water analyzer isa powerful tool thatis appropriate for use in a wide range of applications and environments. Copyright # 2009 JohnWiley & Sons, Ltd. Measurements of the stable isotopic composition of waterhave historically facilitated a greater understanding of manyphysical processes in hydrology, meteorology, ecology andmore recently, global climate change to name just a few.Specifically, the stable isotopes of hydrogen and oxygen(hereafter denoted by the common  d D and  d 18 O) have givenrise to deeper understanding of transport and exchangeprocesses in hydrological systems at many scales. Acting asindicators, integrators, tracers and recorders they have, forexample, allowed us the ability to reconstruct paleoclimaterecords from ice cores 1,2 that form one of the pillars of globalclimatechangesciencetoday.Inotherapplicationshydrogenand oxygen isotopes act as natural tracers driving models of soil water content and pore water that produce informationabout near-surface water dynamics.Although a vast amount of literature exists on  d D and d 18 O measurements of liquid meteoric water, 3–9 fewerisotopic measurements of meteoric water vapor have beenperformed. 10–15 This is because isotopic analyses aretypically done with isotope ratio mass spectrometry (IRMS)systems in a laboratory setting, which precludes convenientreal-time field use and limits the samples to those which can be reliably prepared manually in the field and transported tothe lab. Recently, isotopic water vapor analyzers based onlaser spectroscopy have advanced to the point where fielddeployment is now possible. The relatively small size,weight, and ruggedness of these laser-based analyzers couldtransform the way in which scientists think about acquiring d D and  d 18 O measurements of meteoric water vapor. Thistechnology will enable systematic, continuous studies of theisotopic composition of meteoric water vapor to beperformed at locations never before possible. In addition,the wide deployment of such analyzers will help greatlyimprove the coverage of isotope data associated withhydrological cycling and recycling. 16 Theobjectiveofthispaperistodemonstratetheabilityofarecently developed isotopic water analyzer to make precisemeasurementsoftheisotopiccontentofwatersamplesinthe RAPID COMMUNICATIONS IN MASS SPECTROMETRY Rapid Commun. Mass Spectrom.  2009;  23 : 2534–2542Published online in Wiley InterScience ( DOI: 10.1002/rcm.4100 * Correspondence to : P. Gupta, Picarro, Inc., 480 Oakmead Pkwy,Sunnyvale, CA 94085, USA.E-mail: y Presentedatthe2 nd  JointEuropeanStableIsotopeUserMeeting(JESIUM), Presqu’ıˆle de Giens, France, 31 August–5 September,2008.Contract/grantsponsor:NationalCenterforResearchResources(NCRR),acomponent oftheNationalInstitutesof Health(NIH);contract/grant number: 2 R44 RR021297-02. Copyright # 2009 John Wiley & Sons, Ltd.  liquid phase as well as in the vapor phase. Here, we reportthe results of three trials undertaken with the Wavelength-Scanned Cavity Ring-Down Spectroscopy (WS-CRDS) iso-topic water analyzer that test and confirm its performanceandalsoidentifyany shortcomings. Laboratorytestingofthestability, memory, and drift of this analyzer for liquidsamples was undertaken at the University of ColoradoInstitute for Alpine and Arctic Research (INSTAAR)Stable Isotope Laboratory. Two separate field campaignswereundertakentotestthewatervaporanalysismodeoftheinstrument. One was a 10-day field trial undertaken at theWoods Hole Oceanographic Institute where the ambient airwasveryhumidandtemperatureswerewarm.Thesecond,amore strenuous test, was a 26-day field trial undertaken atthe National Oceanographic and Atmospheric Adminis-tration’s Mauna Loa Observatory in Hawaii. The range of variation in  d D and  d 18 O, and generally low H 2 O concen-trations at Mauna Loa, make for a particularly challengingenvironment for measurements and so it is an ideal site atwhich to assess the performance of the analyzer. Together,these three trials demonstrate the practical performance of the isotopic water analyzer in making continuous andcalibrated measurements of liquid water and ambient vaporsamples. EXPERIMENTAL The isotopic water analyzer has three components: a gas-phase instrument that measures the concentration andisotopic content of water in vapor form, a liquid evaporatorthat converts liquid water samples into water vapor and anautosampler that injects liquid water samples into theevaporator. Although the measurement of water vaporsamples only requires the gas-phase instrument, theinstrument should be calibrated periodically to compensatefor any instrumental drift. In order to calibrate theinstrument, known water standards were measured everyfew hours so that the data could be scaled, based on the trueversus measured values of these standards. While isotopicstandardsexist invaporform for otherisotopicanalyses (e.g.CO 2  gas), the primary standards (e.g. VSMOW, SLAP, GISP)for water exist only in liquid form. Hence, the calibrationprocess requires that these liquid standards be vaporized sothat they can be measured by the gas-phase instrument.The introduction of liquid water samples for vaporanalysis provides two major challenges. First, the watermust be vaporized with little or no isotopic fractionationduring sample preparation otherwise there will be systema-tic errors in the measurement. Second, both the vaporizedliquid sample and the water vapor samples must bemeasured identically.In order to overcome these challenges, a liquid waterevaporator (Picarro Inc., Sunnyvale, CA, USA) has beendeveloped for use with the gas-phase instrument. Theevaporator allows for flash evaporation of the liquid watersamples and subsequent equilibration of the resulting vaporwith dry carrier gas to minimize isotopic fractionation. It isintended to enable the measurement of liquid water sampleswith great precision and also facilitate switching betweenambient atmospheric gas sampling and liquid samples. Thiscapability permits the use of liquid standards for real-timerecalibration even in the field. With the help of thisevaporator, variations in the isotopic content of samplestaken from a variety of environmental sources can bemeasured with high precision and confidence. Gas-phase instrument – WS-CRDS technology WS-CRDS has emerged as a proven cavity-enhanced opticalmethod for detecting trace quantities of gases 17,18 and is now beginning to be applied to stable isotope measurement. Thetechnology employs absorption spectroscopy using aninfrared laser to detect the various isotopologues of water.A full description of the method can be found elsewhere. 18 The temperature and pressure of the ring-down cavity aremonitored and actively stabilized at 80 8 C  20mK and35  0.1 Torr, respectively. This stability is required fortwo reasons. First, the strengths of the absorption peaks of the various water isotopologues have different temperaturesensitivities, so an unstable temperature could result inunstable and erroneous isotopologue measurements. Sec-ond, this stability enables the absolute concentration (molefraction) of HH 18 O (conc[H 182  O]), HDO (conc[HD 16 O]) andHH 16 O (conc[H 162  O]) in air samples to be simultaneouslymeasured. In addition, a wavelength monitor, built into thegas-phase instrument, is continuously used to target specificwavelengths of the spectrum with about 1MHz precision,helping to counteract laser drift.WS-CRDS aims to measure the rate loss of light emergingfrom an optical cell that contains the gas sample. BecauseWS-CRDS is a time-based measurement, several orders of magnitude of dynamic range of loss can be measuredwithoutchangestotheanalyzer.Thedecaytimeconstantcan be expressed as: t   ¼  1 c ð L þ  A Þ  (1)where t  isthelaserlightdecaytimeconstant, c isthespeedof light,  L  is the loss in the empty cavity and  A  is the absorptionloss due to the gas species to be measured. Assuming anempty cavity loss of 1ppm/cm, an absorption loss of 0.001ppm/cm gives a decay time constant,  t  ¼ 33.3 m s, whilean absorption loss of 1ppm/cm gives a decay time constantof 16.67 m s. This dynamic range allows for accuratemeasurements of both high and low abundance isotopolo-gues.The ratios of the concentrations,  R 18 O  and  RD , of the twoisotopologues are defined as (currently the H 172  O data is notrecorded): R 18 O ¼ conc ½ H 182  O  conc ½ H 162  O  ; RD ¼ conc ½ HD 16 O  conc ½ H 162  O  (2)and can be used to express the ‘delta values’ ( d ) defined as: d ¼  RR standard  1    1000 (3)where d isgiveninpartsperthousandorthemorecommonlyused ‘per mil’ ( % ). The international reference standard forisotopic water which defines  d 18 O ¼ d D ¼ 0 is ViennaStandard Mean Ocean Water (VSMOW). Copyright # 2009 John Wiley & Sons, Ltd.  Rapid Commun. Mass Spectrom.  2009;  23 : 2534–2542DOI: 10.1002/rcm Measurements of water vapor isotopologues in laboratory and field 2535  Liquid evaporator  The WS-CRDS gas-phase instrument measures watersamples only when they are in the vapor phase. Analysisofliquidsamplesthereforerequirescompletevaporizationof the samples and subsequent transport to the WS-CRDSinstrument. Therefore, a liquid evaporator was developedwhich is intended to facilitate vaporizing of the entire liquidwater sample in a controlled fashion to prevent or minimizeisotopic fractionation. This was achieved by making theevaporation chamber a small (  150cc) volume, withhydrophobic coatings, and maintaining a constant hot(140 8 C) temperature of the wetted parts.The evaporator (see Fig. 1) consists of a hollow cylindricalchamber, a solenoid valve for the carrier gas inlet line, asolenoidvalveforthevacuumport,aninjectionport,athree-way valve and fittings to connect them. A dry carrier gas,such as nitrogen or zero-grade dry air, enters the cylinderthroughtheinletvalve.Avacuumvalve attheexitendofthecylinder is connected to a small diaphragm vacuum pumpwhich enables full evacuation of the evaporator whenneeded. The injection port, through which liquid samplesare injected into the evaporator, is sealed and isolated fromroom air through a 9.5mm septum (Restek, Bellefonte, PA,USA). The three-way valve in the evaporator has twooperating positions: when it is powered it connects the gas-phase instrument to the evaporator and when it is notpowered it connects the instrument to a dry gas supply.While measurement of water vapor samples only requiresthe gas-phase instrument, calibration of this instrumentrequires use of the evaporator. Hence, when measuringvapor samples the inlet of the gas-phase instrument isconnected to the vapor source through an external solenoidvalve. This valve can switch the input of the WS-CRDSinstrumentfromthevaporsampletodrygas.Theinstrumentis connected to dry gas prior to being connected to theevaporator for measuring liquid water standards so that anytraces of the vapor sample are removed from the measure-ment cell. The standards can then be manually orautomatically injected into the evaporator and measuredusing the protocol described in the following section. Operation with an autosampler  The WS-CRDS instrument and the liquid evaporator can becoupled with a CTC HTC-Pal liquid autosampler (LEAPTechnologies, Carrboro, NC, USA) to make automatedmeasurements of liquid water samples. The autosampleris fitted with an 5- m L glass syringe (SGE Inc., Austin, TX,USA) and programmed to provide injections of 2- m L watersamples from 2-mL glass vials fitted with Teflon/siliconsepta (MicroLiter Analytical Supplies, Suwanee, GA, USA).If cleaned on a daily basis with organic solutions, thesyringes can make up to 3000 consistent injections beforetheir plungers start to malfunction due to clogging.In operation, a needle-rinsing protocol is employed thatinvolves aspirating several aliquots of the sample water anddispensing it into a waste port. While the needle is rinsing,theevaporator and theanalyzerare purged withdrygas andevacuated multiple times to clean the surfaces and removeany memory of the previous sample. When the variouscomponents have completed this cleaning cycle, the needlecollects the final 2 m L of the water sample and injects it intothe evaporator through the injection port.Within a few seconds after injection of the liquid watersample, the sample is completely vaporized because theevaporator is maintained at 140 8 C and is under vacuum. Thewater vapor sample is then mixed with dry carrier gas toform a uniform, homogenous mixture of all the isotopolo-guesofwatervapor.Mixingthewatervaporsamplewithdrygasfacilitatestheflowofthesamplefromtheevaporatorintothe WS-CRDS analyzer with limited adsorption on thevarious surfaces and it also dilutes the vapor sample goinginto the gas-phase instrument. It is advantageous to dilutethe sample as there is an optimum water concentration thatgives the highest possible precision from the WS-CRDSinstrument, although the instrument is designed to meetspecifications over a wide concentration range.After equilibration with itself and the carrier gas for  2min, thehomogeneousmixtureof thesampleanddrygasflows to the WS-CRDS instrument at a constant flow rate viathe three-way valve, where its isotopic content is repeatedlymeasured over the next few minutes. The 2- m L injection of  Figure 1.  Schematic of the evaporator. The evaporator facilitates vaporization of liquid watersamples in a controlled fashion to prevent isotopic fractionation. Dry carrier gas is introduced intothe cylinder through the inlet valve and the vapor sample equilibrates with it to form a homo-geneous gas sample. The three-way valve allows this equilibrated sample to flow from theevaporator to the analyzer. Copyright # 2009 John Wiley & Sons, Ltd.  Rapid Commun. Mass Spectrom.  2009;  23 : 2534–2542DOI: 10.1002/rcm 2536 P. Gupta  et al.  water, after dilution in the evaporator, yields a pulse of approximately 20 000  220ppm (5.9  10 17 water moleculesper cm 3 ) water vapor in the measurement cell held at   35Torr during analyses. Since the sample has a uniformconcentrationandtheflowrateintothegasphaseinstrumentis constant, the pulse of water measured by the analyzer isflat, as shown in Fig. 2. Measurements made during the flatpart of this concentration profile are used to determine theisotopic ratios in the sample. After a few minutes of measurement time, the evaporator is isolated from thegas-phase instrument again and the two components gothrough their automatic cleaning cycle in parallel to removeany memory of the old sample. Once the components areclean again, the new sample is injected into the evaporatorwhich goes through the measurement cycle and so on. Whenusing an autosampler, the entire measurement of more than100 samples is fully automated and requires no userintervention.The isotopic water analyzer is designed to have minimaldrift and as such the measurement of the delta values can becalibrated to the internationally accepted VSMOW scale byintermittently measuring standards. When liquid samplesare being measured, standards can be interspersed withinthe unknown liquid samples. The data can then be scaledusing a linear fit of the true versus measured delta values of the standards. Laboratory test The isotopic water analyzer was tested in a laboratoryenvironment at the University of Colorado Institute of Alpine and Arctic Research (INSTAAR) Stable IsotopeLaboratory.Forthetests,asuiteoffiveinternallabstandardswas used that had previously been calibrated by IRMS toprimary reference materials VSMOW, SLAP and GISPsupplied by IAEA, Vienna, Austria. The IRMS method usedfor  d 18 O was CO 2 /H 2 O head-space equilibration, and themethod for  d D determination was an automated uraniumreduction system. 19 The isotopic range of the lab standardswas approximately 430 per mil for  d D (over 55 per mil  d 18 O),as shown in Table 1.Determinations of sample precision and drift were made by making 90 injections (2- m L sample volume each) of purewaters (deionized local waters) over a period of 48h. Thestandard deviations were 0.29 %  for  d D and 0.05 %  for d 18 O (n ¼ 90), and the drift over the entire 48-h period wastowards lighter values for both isotopes, but was very small(0.26 % d D and 0.05 % for  d 18 O over 48h). No drift correctionshave been applied and the results of this experiment areplotted in Fig. 3. There also appears to be random noise overthe period and, while the srcin of this noise is unknown, thestability and reproducibility over this time period appear toexceed those obtained with the other IRMS methods in thispaper.ThehydrogenresultsaresimilartothosefoundbyLis et al . 20 for an off-axis integrated cavity output spectroscopy(OA-ICOS) instrument, but the oxygen results in this studyappear to be more reproducible, by a factor of 3 (  0.16 %  byLis  et al . to  0.05 % in this study).In any system that converts liquid water into vapor therewill be residual water molecules retained in the system andthus any one analysis will contain some small fraction of theprevious water injected into the system. 19 To quantify thissystem memory, multiple analyses of isotopically differingwaters were made. Figure 4 shows the memory effect for both hydrogen and oxygen obtained from 25 experiments,moving from 15 injections of one water to 15 injections of anisotopically different water. In all cases, the true value of thewater samples was taken to be the average of the last 4 of the15 measurements (injections), and it was assumed that this isthe asymptotic value. In theory, the measured value M is afunctionofX  T þ (1–X)P,whereMisthemeasuredvalueof current injection, X is the unknown memory coefficient of this n th injection, T is the true value of the current watersample and P is the true value of the immediately previouswatersample.Therefore,thememorycoefficientXforthisn th injection is (M – P)/(T – P). For comparison, it is useful toexpress a memory coefficient as a percentage of the previoussample, or 1 – X.We found that the memory for hydrogen over the firstthree injections is significantly greater than for oxygen overthesamespan.Thereasonforthisisnotwellunderstood,butmay possibly relate to interactions at the molecular level between the water and the hydrophobic coatings on the Figure 2.  Sample data showing a pulse of water vapor. Theanalyzer is purged with dry gas while a new sample is pre-paredinthe evaporator.Thethree-wayvalvethenopenstoletthis homogeneous sample into the analyzer. After a fewminutes of sampling, the three-way valve closes and theevaporator and analyzer are both purged to remove anymemory of the sample. Table 1.  The isotopic water standards and correspondingdelta values used in the laboratory tests Isotopic standardH 182  O/H 162  O ( d 18 O) HD 16 O/H 162  O ( d D)( % ) ( % )Florida   0.56  0.15   3.75  1Boulder   16.73  0.15   124.39  1WAIS   26.07  0.15   205.08  1Greenland   38.43  0.15   300.41  1Vostok   56.13  0.15   437.52  1  Internal laboratory standards. Copyright # 2009 John Wiley & Sons, Ltd.  Rapid Commun. Mass Spectrom.  2009;  23 : 2534–2542DOI: 10.1002/rcm Measurements of water vapor isotopologues in laboratory and field 2537  wetted surfaces within the system. The memory coefficientsfor both isotopes were quite reproducible for both the firstandsecondinjectionsat(1–X) ¼ 0.5  0.06and0.25  0.01forhydrogen and 0.15  0.02 and 0.08  0.02 for oxygen,respectively. Because of the consistency of the memorycoefficients,onecouldmakefewermeasurementsofanywatersampleandcalculatethetruevalueusingT ¼ (M–(1–X)P)/Xto facilitate a higher throughput with the analyzer.Earlier efforts on similar systems have shown that smallmass variations can adversely and dramatically affect themeasuredisotoperatios. 20 Totestthisdependence,aseriesof different size injections was made of the same water into theanalyzer. The results showed minimal dependence of thewater mass on the delta values. There is a slightly greaterdependence on water mass for oxygen isotopes thanhydrogen. In both cases, however, the correction is smalland was applied in the software, based on these character-izations.Theresultsofallthebenchtestswereveryreproducibleondifferentdays,includingtestsofstability,memory,drift, anddependence on water vapor concentration. Because of thisapparent consistency, it is then possible to apply post-analysis calculations that can correct for memory and scalemeasurements to isotopic standards to achieve results thatareaccuratelytiedtotheVSMOWscale.Thedriftcoefficientswere found to be very small (less than 0.002 %  d D and0.0005 %  d 18 O over 48h), and were too insignificant andwithin the noise of the instrument to correct on timescalesless than 48h. Memory effects, especially for hydrogen, weremuch smaller (  5 percent 1 st injection) than those observedin other IRMS water injection systems (up to   9 percenton 1 st injection 19 ) and can be characterized and removedfrom the final data. The apparatus to convert liquid waterinto vapor is crucial for calibrating any water vapor isotopicmeasurements, and because calibrations are an essentialelement of producing credible isotopic analyses of water orwater vapor, this represents an important step forward innon-IRMS water isotope analyses. Field test at Woods Hole OceanographicInstitute In an attempt to gain insight into the performancecharacteristics of the isotopic water analyzer outside thelaboratory, a 10-day field trial was undertaken at WoodsHole Oceanographic Institute. The measurement site was onthe roof of an ocean-side building roughly 30m above thewater. While the analyzer was installed inside a protectiveshelter, a 4m long and 9.5mm diameter Dekoron line wasmountedoutside theshelter totake ambientair samples.Thestable isotopes  d D and  d 18 O as well as the water vaporconcentrations were continuously monitored; liquid watercalibration standards were measured every 8h. In order toswitch between liquid and vapor samples, the three-wayvalve in the evaporator was configured such that it eitherconnected the vaporized liquid sample in the evaporator tothe analyzer or the ambient vapor sample to the analyzer(Note: this configuration is different from the one describedearlier). Water vapor measurements were produced every15s. Dry carrier gas was provided to the analyzer by flowingambient air through a dryer (Drierite, Xenia, OH, USA).TheH 2 Oconcentrationattheexittothedryerwas < 250ppm.The H 2 O concentration of ambient air over the 10-dayperiod varied from 14000ppm to 24000ppm while the d 18 O varied from   10 %  to   22 %  and  d D spanned from  93 % to   170 % , as shown in Fig. 5. Although the ambient Figure 3.  Plot of 90 analyses of deionized water, performedover a 48-h period. The standard deviations of the determi-nations are 0.05 % for oxygen and 0.3 % for hydrogen. Drift istowardslighter valuesover the48-hperiod,butitisverysmall,and is easily normalized with periodic standards during theanalysis run. Figure 4.  The system memory of previous water, expressedas a percentage of the previous sample, for both hydrogenand oxygen. For oxygen, after the 3rd injection, the memoryhas been reduced to zero, whereas hydrogen has a slightlymorepersistentmemory.Errorbarsarestandarddeviationsofmultiple determinations (n ¼ 25) for each of the first threeinjections. Copyright # 2009 John Wiley & Sons, Ltd.  Rapid Commun. Mass Spectrom.  2009;  23 : 2534–2542DOI: 10.1002/rcm 2538 P. Gupta  et al.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks