Anal Bioanal Chem (2006) 385: 248–255DOI 10.1007/s00216-006-0362-0

SPECIAL ISSUE PAPER

Friederike Weritz . Dieter Schaurich .Alexander Taffe . Gerd Wilsch

Effect of heterogeneity on the quantitative determination of traceelements in concrete

Received: 31 October 2005 / Revised: 7 February 2006 / Accepted: 13 February 2006 / Published online: 8 March 2006# Springer-Verlag 2006

Abstract Laser-induced breakdown spectroscopy hasbeen used for quantitative measurement of trace elements,e.g. sulfur and chlorine, in concrete. Chloride and sulfateions have a large effect on the durability of concretestructures, and quantitative measurement is important forcondition assessment and quality assurance. Concrete is ahighly heterogeneous material in composition and grain-size distribution, i.e. the spatial distribution of elements.Calibration plots were determined by use of laboratory-made reference samples consisting of pressings of cementpowder, hydrated cement, cement mortar, and concrete, inwhich the heterogeneity of the material is increasingbecause of the aggregates. Coarse aggregate and cementpaste are distinguishable by the intensity of the Ca spectrallines. More advanced evaluation is necessary to account forthe effect of the fine aggregate. The three series ofreference samples enable systematic study of the effects ofheterogeneity on spectral intensity, signal fluctuation,uncertainty, and limits of detection. Spatially resolvedmeasurements and many spectra enable statistical evalua-tion of the data. The heterogeneity has an effect onmeasurement of the sulfur and chlorine content, becauseboth occur mainly in the cement matrix. Critical chlorideconcentrations are approximately 0.04% (m/m). The chlo-rine spectral line at 837.6 nm is evaluated. The naturalsulfur content of concrete is approximately 0.1% (m/m).The spectral line at 921.3 nm is evaluated. One futureapplication may be simultaneous determination of theamount of damaging trace elements and the cement contentof the concrete.

Keywords LIBS . Heterogeneity . Concrete . Chlorine .Sulfur

Introduction

Concrete is a highly heterogeneous material consisting ofcement, water, fine and coarse aggregate, and additives.Reinforcement is also included during construction. Theproperties of concrete, for example strength and resistanceagainst external diffusion of aqueous solutions, depend onthe hydration products, the water-to-cement ratio, and thepore structure. The hydration products are cement minerals,which act as binders of the aggregates. Typical cementminerals are complex silicates and aluminates with majorelements Ca, Si, Al, Fe, and O, and minor elements such asNa, K, and Mg. Because most damaging species penetratethe concrete via the pores when dissolved in water, thedevelopment of concrete with small pores and a smallamount of cross-linking is an important strategy for avoidingdamage. Chloride ingress is one of the most important typesof damage occurring to reinforced concrete structures.Dissolved chlorides from external sources penetrate theconcrete via the pore structure of the cement paste. Whenchlorides reach the steel of the reinforcement the passivelayer is locally destroyed (in the presence of water and O2)and the chlorides act as catalyst for corrosion of the steel(pitting corrosion) [1, 2]. Critical concentrations are 0.04%(m/m) and higher, depending on the material, construction,environmental conditions and use [2]. High chlorideconcentrations occur in marine environments and whereroad salts are applied, e.g. parking garages. Only a smallpercentage of chloride can be bound in cement minerals, e.g.as Friedel’s salt (3CaO.Al2O3.CaCl2.10H2O) [1].

A second important species involved in damage mechan-isms are sulfates, which can react with the cement minerals.Sulfates are abundant in concrete, because gypsum is used asa component of cement to regulate the time of hydration.Sulfates are part of cement minerals, e.g. gypsum (CaSO4.2H2O), ettringite (Ca6[Al(OH)6](SO4)3.26H2O), thaumasite(Ca3Si(OH)6(CO3)(SO4)3.12 H2O). Formation of theseminerals by solid-state reaction of the hydrated cementleads to concrete damage as consequence of lower bindingability or greater volume. Damage symptoms are loss ofstrength and mass or cracks and spalling [3]. The source of

F. Weritz (*) . D. Schaurich . A. Taffe . G. WilschBAM - Bundesanstalt für Materialforschung und -prüfung,Berlin (D),Unter den Eichen 87,12205 Berlin, Germanye-mail: Friederike.Weritz@bam.deTel.: +49-30-81044228Fax: +49-30-81041447

sulfates can be internal or external, whereas the reactionsdepend on the ambient conditions. External sources can besewage waters, industrial waste, or soils with a high sulfurcontent. The natural sulfur content of concrete is approxi-mately 0.1% (m/m).

Thus, chloride and sulfate ions have a large effect on thedurability of concrete structures, and quantitative measure-ment is important for condition assessment and qualityassurance.

Until now chloride and sulfate content have beendetermined by chemical analysis. Concrete cores aresampled at the structure and taken to the laboratory. Thecores are cut into depth slices of minimum thickness10 mm, because of the coarse aggregate. The concrete ofeach slice is ground and dissolved in acid. The methods arespecified in technical standards, usually potentiometrictitration for chloride e.g. [4, 5] and gravimetric precipita-tion for sulfates [6]. Because the analysis is time-consuming and quite expensive only a few concrete coresare investigated and the results are taken as representativeof the complete structure.

Laser-induced breakdown spectroscopy is a new ap-proach for quantitative measurement of the chlorine andsulfur content of concrete. (The chlorine content is equal tothe chloride content, the sulfate content can be calculatedstoichiometrically from the sulfur content, with someuncertainty because not all sulfur may be bound in sulfate.)General advantages of LIBS compared with chemicalanalysis are [e.g. 7]:

– information about more than one element is obtainedwith the same measurement;

– depth profiles with millimeter resolution are obtained;– imaging of element distribution is possible;– more measurements both of single cores and complete

structures are possible; and– the method is, in principle, portable and can be used in

harsh environments.

These properties of LIBS recommend it for applicationsin civil engineering.

Heterogeneity has an effect on measurement of sulfateand chloride content by both chemical analysis and LIBS,because the ions occur mainly in the cement matrix, whichis less than 35% (m/m) of concrete. In each ground depthslice (chemical analysis) and in each ablated volume(LIBS) the percentage of aggregate and hydrated cementvaries. This can be understood by thinking of a 10 mmdepth slice of a 30 mm diameter concrete core, which caneither consist of a large aggregate only, with 0% (m/m)chloride or sulfate, or of a great percentage of hydratedcement, containing a certain amount of chloride or sulfate.

Calibration plots are constructed by use of laboratory-prepared reference samples with element contents inconcentration ranges occurring at real sites. Otherwisethe normalized Ca intensity could be used. Both normal-ization methods provide almost the same results asindications of spectra mainly originating from cementitiousmaterial; the uncertainty is slightly better for the Ca/Oratio.

Experimental set-up and evaluation

The experimental set-up is shown in Fig. 1. A NdYAG-Laser (1064 nm, 10 Hz, 400 mJ/pulse; Big Sky) is focusedon the sample surface to generate the plasma plume.Because concrete samples may have a high surfaceroughness, a large focal length is chosen, so that thepoint of focus is not critical. The plasma emission is guidedthrough an optical fibre to the detection unit. In front of theoptical fiber a spectral filter (transmits light >550 nm) isinstalled to avoid interference with higher spectral orders.The wavelength range is adjusted by use of a mono-chromator (Acton Research SpectraPro 150) with a1200 lines mm−1 grating. An OMA 4 detector (EG&G,CCD-Array 1024×256 pixel, 18 bit intensity resolution,

Fig. 1 Experimental set-up

249

detector temperature −20°C) is used. The detector is a non-intensified system with a comparative high quantum-efficiency in the NIR (approximately 15% at 930 nm).Time-resolved measurements are not possible with thissetup. The exposure time chosen is 860 ms. During thistime the signal of eight laser shots are accumulated on theCCD, thus improving the signal-to-noise ratio. The sampleis positioned on an xy-translation stage and the measure-ment area is flushed with helium. The gas flow is adjustedto exclude ambient air completely from the measurementarea and to remove the evaporated material. Use of heliumalso results in a better signal-to-noise ratio for both chlorineand sulfur [8–10]. The samples are moved on the xy-translation stage. Meander scans are performed for hydrat-ed cement and cement mortar samples and line scans forconcrete samples. The measurements are performed in thecenter of a freshly prepared surface, to avoid effects of theformwork margins. Each scan comprises 50 single spectra.On concrete samples six linear scans with 50 single spectra,a total of 300 single spectra per sample, are measured.

The concrete cores from real sites are moved, so thatmeasurements along a line parallel to the front side(x-direction) are accomplished with a spatial resolution of1 mm. The distance between linear scans is 2 mm (depthresolution). The exposure time is 860 ms.

Quantitative analysis of the spectral data uses thecorrelation between the intensity of the spectral line andthe element content. For quantitative determination ofchlorine and sulfur the spectral lines at 837.6 nm and at921.3 nm, respectively, are evaluated. For sulfur thespectral line is fitted with a Gauss function. For chlorinethe peak intensity is evaluated, because this method ispreferable for weak spectral lines close to the detectionlimit. Both methods have been compared for both chlorineand sulfur and the results are almost the same. The integralis determined and normalized by division by the calculatedbaseline value at the position of the peak maximum. Thevalues are averaged for each sample (reference samples) oreach depth (concrete cores from real sites) and the standarddeviation is determined. For spatial resolved measurementseach spectrum is evaluated and the result is assigned to aposition, giving images. Because concrete consists of at

least two different classes of material—aggregate andcementitious materials—which are spatially heteroge-neously distributed, there are no elements with constantconcentrations (except for oxygen, the content of which issimilar in both). It is, therefore, not possible to use spectralline ratios, and the spectral intensities are normalized to thebackground signal.

Composition of calibration samples

The three sets of reference samples used—hydratedcement, cement mortar, and concrete—were of increasinghetereogeneity because of the aggregates (Fig. 2). Hydratedcement consists of cement powder and water. Cementmortar also contains fine aggregates, i.e. sand with amaximum grain size of 2 mm, and concrete has,additionally, coarse aggregates (SiO2) with a maximumgrain size of 16 mm. All samples are prepared withPortland cement. More details of preparation of the sulfursamples are given elsewhere [9]. The chlorine sampleswere prepared analogously.

The trace elements are bound only in the cementitiousmaterial. Thus, addition of aggregates dilutes the traceelement signal and, because the amount of aggregate ismore than 65% (m/m) this is an important effect. In Table 1the composition is given in more detail, especially therange of trace element concentrations. The elementcontents were chosen as realistic concentrations. Becausesulfur is always abundant in cement, there are no sampleswith 0% (m/m) sulfur.

Effect of heterogeneity on spectra from concrete

With each laser pulse a small amount of material is ablatedand the measured optical emission gives information aboutthe composition of this ablated material. If this material isrepresentative of the bulk, information about the bulkcomposition can be obtained. Concrete is a heterogeneousmaterial with regard to element composition and the spatialdistribution of element composition (and surface rough-

Hydrated cement- NaCl + cement + water- CaSO4 + cement + water

Cement mortar- NaCl + cement + water + fine aggregate (0-2mm)- CaSO4 + cement + water + fine aggregate (0-2mm)

Concrete- NaCl + cement + water + fine + coarse aggregate (2-16mm)- CaSO4 + cement + water + fine + coarse aggregate (2-16mm)

HETEROGENEITY

Fig. 2 Composition of refer-ence samples (dimensions of thesurface of interest are5 mm×5 mm for hydrated ce-ment and cement mortar, and15 mm×15 mm for concrete).The traces of the measurementscan be seen on the photographs

250

ness). In previous work the effect of concrete heterogeneityon the main elements is described [11].

Typical spectra obtained from concrete samples areshown in Fig. 3. On the left a spectrum with the Cl spectralline (top), measured in the range from 810 nm to 870 nm,and a spectrum with the S spectral line (bottom), measuredin the range from 890 nm to 940 nm, are shown, andassignment of the spectral lines to elements is given.

On the right, two spectra measured on the same concretesample are compared for both chlorine (top) and sulfur(bottom) and assigned to either aggregate or (mainly)cementitious material. Comparison of these spectra iscrucial for understanding the effect of the heterogeneity ofthe fine and coarse aggregates.

The spectrum assigned to aggregate is measured on acoarse aggregate, where coarse can be defined in thisexperiment as a surface exposed to the laser more than fivetimes the laserfocus (1 mm2), so that any effect of materialother than the aggregate itself can be excluded. Theaggregates in the reference samples consist of SiO2.Because no silicon spectral lines of sufficient intensityoccur in the wavelength regions under investigations, onlyoxygen spectral lines can be seen.

The spectrum assigned to (mainly) cementitiousmaterial is an example of a spectrum with the highestintensity of the calcium spectral lines, where calcium isan indicator of cementitious material. The intensity ofthe spectral lines of trace elements such as S and Cl,which are only abundant in the cementitious material,also have the greatest intensities, which corresponds tothe respective element concentration in the cementitiousmaterial. There is, therefore, a direct correlation betweenthe intensities of the Ca and the S or Cl spectral lines.

The intensity fluctuations are because of the varyingfractions of cementitious material and aggregate ablated ineach measurement, i.e. a question of sampling. Coarseaggregates can be identified, but in most measurements the

Table 1 Composition of reference samples

Chlorine, fourteensamples each [%(m/m)]

Sulfur, ninesamples each [%(m/m)]

Hydratedcement

Cement 76 Cement 76Water 23 Water 23NaCl 0.08–4.0 Gypsum 0–3.7i.e. chlorine 0.05–2.34 i.e. sulfur 0.9–1.7

Cementmortar

Cement 22 Cement 22Water 11 Water 11NaCl 0.02–1.15 Gypsum 0–1.3i.e. chlorine 0.01–0.70 i.e. sulfur 0.3–0.6Sand 66 Sand 66

Concrete Cement 13 Cement 13Water 8 Water 8NaCl 0.01–0.67 Gypsum 0–1.6i.e. chlorine 0.008–0.4 i.e. sulfur 0.2–0.5Aggregate 79 Aggregate 79

Fig. 3 Top left: Concretespectrum in the chlorinewavelength range, withidentification of the spectrallines. Bottom left: Concretespectrum in the sulfurwavelength range, withidentification of spectral lines.Right: Comparison of spectrameasured on coarse aggregateswith a high fraction of cementi-tious material for the Cl (top)and S (bottom) wavelengthranges

251

ablated material is a mixture of both aggregate andcementitious material in unknown proportions, whichcannot yet be quantified. But the spectra provide informa-tion about these fractions—at least qualitatively, becausethe intensity of a calcium spectral line corresponds to thecontent of cementitious material. We have defined the ratioof the intensity of a calcium spectral line to that of anoxygen spectral line (Ca/O) as a qualitative measure of thefractions of cementitious material and aggregate (Table 2).

The relationship between Ca/O ratio and cement/aggre-gate ratio of the measured material can be seen in Fig. 4. Onthe left a color-coded image of the Ca/O ratio is shown.Black corresponds to low values of the Ca/O ratio, i.e.aggregate; blue corresponds to higher values of the Ca/Oratio. Comparison with the photograph of the concretesurface proves that coarse aggregate can be identified, butfine aggregate and cementitious material cannot bedistinguished in the image.

The maximum value of the Ca/O ratio is different forchlorine and sulfur measurements, which is to be expectedas different spectral lines are considered, but can alsodepend on the sample under investigation.

In Fig. 5, left, the distribution of the Ca/O ratio is shownfor measurements on samples of concrete, cement mortar,and hydrated cement with comparable sulfur contentrelative to the mass of cement. The composition of thehydrated cement corresponds to the cementitious materialin the cement mortar and concrete samples.

As expected, the dispersion of the Ca/O ratio increaseswith increasing heterogeneity of the samples. The distri-bution is rather symmetric for hydrated cement and cementmortar, but there are two cumulation areas for concrete, oneat low values (<1) with small dispersion—corresponding topure aggregate spectra—and one at high values (>2.5) withlarge dispersion similar to cement mortar.

The average Ca/O ratio is higher for hydrated cementthan the maximum value for cement mortar and concrete.This result means that the highest fraction of cementitiousmaterial, which can be measured in cement mortar andconcrete is always mixed with aggregate, no pure spectrafrom cementitious material are available.

The maximum and average Ca/O ratios for cementmortar are higher than for concrete, but there is overlap ofthe measurement data. Because of the higher masspercentage of aggregates in concrete (Table 1) even inthe spectra from samples with the highest fraction ofcementitious material the fraction of aggregate is higher forconcrete than for cement mortar. Pure spectra are availablefrom measurements on coarse aggregate in concrete, butfor cement mortar no pure aggregate spectra are acquired,because no low (<1) Ca/O ratios occur.

The distribution of the intensity of the sulfur spectral line(IS), shown in Fig. 5, right, is in accordance with thedistribution of the Ca/O ratio, although there is not an exactcorrelation.

The dispersion of IS increases with increasing heteroge-neity. The dispersion of IS for measurements on hydratedcement is larger than the dispersion of the Ca/O ratio. Theaverage of IS is higher for hydrated cement than for eithercement mortar or concrete. For concrete there are fewerlow values than for the Ca/O ratio. The reason for thegreater variation is probably the weak signal of the Sspectral line.

It can, nevertheless, be seen that the intensity of thesulfur spectral IS line is correlated with the fraction ofcementitious material and aggregate in the respectiveinvestigated material. The effect of grain size distributioncan be seen from the different effects of coarse and fineaggregates on spectral intensity. Spectra measured oncoarse aggregates can be identified and thus taken intoaccount. The effect of fine aggregates is more difficult toaccount for.

Measurement of chlorine in the range 810 to 870 nmgives comparable results for the Ca/O ratio and forevaluation of sodium (chlorine was added as NaCl). Forchlorine the intensity of the spectral line is weaker andfor some samples below the detection limit; thus thedispersion of the signal is greater and the normalizedintensity is not clearly offset for the different materials.

Effect of heterogeneity on calibration plots

The calibration plot obtained for concrete samples withdefined sulfur content is shown in Fig. 6. The sulfur contentis determined by chemical analysis with a coulometricmethod (Coulomat CS 30 HT, Behr Labortechnik) [12].For each sample 300 single spectra were evaluated. Alinear calibration plot was obtained with an uncertainty ofapproximately 30% (average for all samples). The highuncertainty is a consequence of the above described

Table 2 Ca/O criteria for chlorine and sulfur evaluation

Chlorine Sulfur

Ca/O= Ca (849.9 nm)/O(844.6 nm)

Ca (891.2 nm)/O(926.6 nm)

Fig. 4 Spatially resolved representation of the Ca/O ratio(Ca (849.9 nm)/O (844.6 nm)) and comparison with photograph(taken after measurement—the measurement lines can be seen) ofthe measured concrete core. (Dimensions of the concrete core:30 mm×50 mm)

252

fluctuations of the intensity of the spectral lines of elementsfrom the cementitious material. Improved evaluation,which takes into account the effect of heterogeneitybecause of the aggregates, is necessary to reduce theuncertainty.

For improved evaluation a limit for the Ca/O ratio isdefined, so that only spectra with the highest fraction ofcementitious material are evaluated, but also so that at least

a statistically sufficient number of spectra is evaluated(approximately 20% of the original number of spectra).Thus, the limit value for the Ca/O ratio is not constant fordifferent types of material and so far empirically deducedfrom the spectra. Comparison of different limits for theCa/O ratio has shown that the results converge the higherthe limit value chosen and the results are constant withinthe uncertainty of the measurement. Even if a limit for theCa/O ratio is determined for each measurement, not for ameasurement series, it is valid to argue that by doing sospectra from ablated material with highest amount ofcementitious material available from the material areselected and evaluated.

Calibration plots based on all the data and those based ona reduced data set (Ca/O>limit) are compared in Fig. 7 forsamples prepared from hydrated cement (Ca/O>5.25) andconcrete (Ca/O>3.7).

For hydrated cement the effect of improved evaluation isuncertainty reduced from 8% to 3% whereas the averagevalue for each sample remains almost constant and the dataare fitted with the same linear function.

For concrete the average value for each sample is higherfor the improved evaluation than for all the data. This isbecause with the improved evaluation only the spectra withthe highest fraction of cementitious material are evaluated,

Fig. 5 Left: Ca/O ratio distribution for measurements on hydratedcement, a cement mortar, and a concrete sample with equal sulfurcontent relative to the mass of cement. (Evaluation of 300 spectra

for concrete and 100 spectra for each hydrated cement and cementmortar.) Right: Distribution of the normalized intensity of the sulfurspectral line at 921.3 nm for the same measurements

Fig. 6 Calibration plot for the sulfur content of concrete samples

Fig. 7 Left: Calibration plotsfor the sulfur content of hydrat-ed cement samples. Comparisonof evaluation of all data andreduced data (Ca/O>5.25).Right: Calibration plots for sul-fur content of concrete samples.Comparison of evaluation ofall data and reduced data(Ca/O>3.7)

253

and in these spectra the intensity of the sulfur spectral lineis also highest, as shown in Fig. 5. Thus the calibrationfunction is almost parallel offset. The uncertainty isreduced substantially—from 30% to 9%.

The results for cement mortar are similar. The limit forthe Ca/O ratio is 4.5, the average value for each sample ishigher for the improved evaluation and the uncertainty isreduced from 13% to 7%.

Almost the same results—parallel offset and reductionof the uncertainty—are obtained for sodium measured onthe chlorine calibration samples. Evaluation for chlorinealso results in a substantial decrease in uncertainty but noparallel offset. The uncertainty in the calibration plot forchlorine is 15%. The differences are probably because thesignal is weak—close to the detection limit.

Effect of heterogeneity on the limit of detection

Because sulfur is a natural component of cement, it isalways abundant in building materials. The intensity of thesulfur spectral line is always measurable for the concentra-tions present. Thus, the detection limit is always below theabundant S content and has, therefore, not been determinedfor sulfur.

In unaffected concrete the chlorine content is zero, so thedetection limit is an important issue for chlorine. Therelevant concentration range for chlorine in concrete andcritical chlorine contents leading to corrosion of thereinforcement are given in Table 3. So far we haveachieved a detection limit of 0.15% (m/m) for chlorine [13].Comparison shows that this detection limit is not sufficientfor the relevant chlorine concentration range relative tototal mass but it is sufficient for chlorine content relative to

the cement mass. For chlorine, distinction betweencementitious material and aggregate is not only necessaryto improve the uncertainty but it is important for detectionof chlorine itself.

That the improved evaluation for chlorine leads to usefulresults in practical applications is apparent from Fig. 8, inwhich results obtained from a concrete core originatingfrom a marine structure are presented. Depth-profiles arecompared on the left. LIBS results have 2 mm resolution,the results from chemical analysis have 10 mm resolution.The LIBS results resolve the expected ingress behaviorwith maximum chlorine levels slightly below the surface. Itis apparent that the chlorine content is significantlyenhanced up to a depth of 27 mm. The chlorine contentat greater depths is below the detection limit for LIBS. Theresults from chemical analysis are in accordance with theLIBS results. The right of Fig. 8 shows the spatiallyresolved distribution of chlorine on the surface of the cutconcrete core. The measurement points where chlorine isdetected are shown in green; underneath is the color-codedCa/O ratio.

Conclusions and perspectives

Summarizing, heterogeneity in the composition and spatialdistribution of concrete, mainly because of the aggregates,is the factor dominating the uncertainty of the measuredsignal—other factors, for example fluctuations of the laserenergy, are negligible. The overall abundance of theaggregates leads to broad dispersion. The distribution ofthe spectral intensities depends on the grain-size distribu-tions of the aggregates. The intensity of the spectral lines ofthe main and trace elements depends on the cement/aggregate ratio of the ablated material. Fortunately, thespectra provide information—mainly the normalized in-tensity of a Ca spectral line as a marker for cement—aboutthe cement/aggregate ratio. In a first approach the normal-ized intensity of a Ca spectral line is used to classify thespectra. For quantitative evaluation the only spectra takeninto account are those classified as arising mainly fromcementitious material.

Only spectra with Ca/O ratio greater than a limitdetermined from the data are evaluated. Because S and

Table 3 Concentration range for chlorine in concrete [2]

Cl content (% m/m)relative to cement mass

Cl content (% m/m)relative to total mass

Concentrationrange

0–2 0–0.5

Criticalcontent

0.2–0.4 0.04–0.08

Fig. 8 Results of measurementson a concrete core originatingfrom a marine structure. Left:Comparison of chlorine depthprofile measured with LIBS(mm-resolution) and by chemi-cal analysis. Right: Spatiallyresolved chlorine distribution(green: chlorine detection,underneath: Ca/O ratio). (Thedimensions of the concrete corewere 30 mm×50 mm)

254

Cl are usually bound only in the cementitious material, thiscriterion selects the highest intensities of the S and Clspectral lines and leads to a reduced uncertainty and resultswhich are approximately related to the mass of cement.This is of relevance in civil engineering, because amountsrelated to cement mass are of primary interest, as damageoccurs mainly in the cementitious material.

In future, the possibility of evaluating the Ca/O ratioquantitatively to obtain the cement/aggregate ratio for eachspectrum will be investigated; this will enable simulta-neous determination of the trace element and cementcontent of concrete. Thus, heterogeneity is not only aproperty interfering with the measurement, but providesvaluable information about the material.

LIBS enables quantitative determination of the Cl and Scontent of concrete, provides information about more thanone element from a single measurement and depth profileswith mm-resolution, enables imaging of element distribu-tions, and furnishes more measurements from single coresand complete structures, thus providing more representa-tive results for large structures.

Future work will be concerned with quantitative deter-mination of cement content and the effect of grain-sizedistribution in more detail, further improvement of thedetection limit for chlorine, and determination of calibra-tion functions for other elements, for example Na, K, Fe,and Mg.

References

1. Taylor HFW (1997) Cement Chemistry, 2nd edition. ThomasTelford Publishing, London, ISBN 0 7277 2592 0

2. Jungwirth D, Beyer E, Grübl P (1986) Dauerhafte Betonbau-werke: Substanzerhaltung und Schadensvermeidung in For-schung und Praxis. Beton, Düsseldorf, ISBN 3-7640-0210-7

3. Skalny J, Marchand J, Odler I (2002) Sulfate Attack onConcrete, SPON Press, London and New York, ISBN 0 41924550 2

4. Springenschmidt R (1989) Arbeitskreis “Prüfverfahren Chlor-ideindringtiefe” des Deutschen Ausschusses für Stahlbeton:Anleitung zur Bestimmung des Chloridgehaltes von Beton.Beuth Verlag GmbH, Berlin

5. ASTM C 1202-97: Standard Test Method for ElectricalIndication of Concrete’s Ability to Resist Chloride IonPenetration, American Society for Testing and 6 Materials,West Conshohocken, Philadelphia, USA 1997, AmericanSociety for Testing and Materials, ASTM Norm

6. DIN EN 197-1, edition 2004-08, Beuth Verlag GmbH, Berlin7. Wiens RC, Sharma SK, Thompson J, Misra A, Lucey PG

(2005) Joint analyses by laser-induced breakdown spectroscopy(LIBS) and Raman spectroscopy at stand-off distances.Spectrochim Acta A 61:2324–2334

8. Durdragne L, Adam P, Amouroux J (1998) Time-ResolvedLaser-Induced Breakdown Spectroscopy: Application for Qual-itative and Quantitative Detection of Fluorine, Chlorine, Sulfurand Carbon in Air. Appl Spectrosc 52(10):1321–1327

9. Weritz F, Ryahi S, Schaurich D, Taffe A, Wilsch G (2005)Quantitative determination of sulfur content in concrete withlaser-induced breakdown spectroscopy. Spectrochim Acta B60:1121–1131

10. Asimellis G, Hamilton S, Giannoudakos A, Kompitsas M(2005) Controlled inert gas environment for enhanced chlorineand fluorine detection in the visible and near-infrared by laser-induced breakdown spectroscopy. Spectrochim Acta B60:1132–1139

11. Wiggenhauser H, Wilsch G, Schaurich D, Wöstmann J (2000)Analysis of Concrete Using LIBS, INSIGHT 42(7):436–438

12. Coulomat CS 30 HT, Behr Labortechnik, http://www.behr-labor.com

13. Wilsch G, Weritz F, Schaurich D, Wiggenhauser H (2005)Determination of chloride content in concrete structures withlaser-induced breakdown spectroscopy. Constr Build Mater19:724–730

255