Chemosphere 95 (2014) 379–386

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Characterization and quantification of N-(3-aminopropyl)-N-dodecyl-1,3-propanediamine biocide by NMR, HPLC/MS and titrationtechniques

0045-6535/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.chemosphere.2013.09.049

⇑ Corresponding author. Tel.: +39 049 8275182; fax: +39 049 8275271.E-mail address: paolo.pastore@unipd.it (P. Pastore).

Andrea Mondin a, Sara Bogialli a, Alfonso Venzo b, Gabriella Favaro a, Denis Badocco a, Paolo Pastore a,⇑a Department of Chemical Sciences, University of Padua, Via F. Marzolo, 1, 35131 Padova, Italyb CNR-ISTM, Istituto di Scienze e Tecnologie Molecolari, via Marzolo 1, 35131 Padova, Italy

h i g h l i g h t s

� Standard-less analysis of the biocide N-(3-aminopropyl)-N-dodecyl-1,3-propanediamine.� Full characterization of the raw material by solution NMR spectroscopy.� HPLC/QTOF analysis as first chromatographic characterization of the raw material.� Active principle determination in raw material and in commercial formulations.� Simple acid–base titration in mixed solvent to easily test commercial products.

a r t i c l e i n f o

Article history:Received 28 May 2013Received in revised form 5 September 2013Accepted 6 September 2013Available online 11 October 2013

Keywords:N-(3-aminopropyl)-N-dodecyl-1,3-propanediamineNMR analysisLC/MS analysisLONZABAC

a b s t r a c t

The present paper reports the determination of the tri-amine N-(3-aminopropyl)-N-dodecyl-1,3-pro-panediamine (TA) present in a raw material called LONZABAC used to formulate various, widely usedcommercial biocides. The active principle, TA, is present in LONZABAC together with other moleculesat lower concentration levels. Three independent analytical approaches, namely solution NMR spectros-copy, liquid chromatography coupled to high resolution mass spectrometry (LC/HRMS) and acid–basetitration in mixed solvent, were used to overcome the problem of the non-availability of the active prin-ciple as high purity standard. NMR analysis of raw material, using a suitable internal standard, evidencedin all analyzed lots the presence of the active principle, the N-dodecyl-1,3-propanediamine (DA) and then-dodecylamine (MA) and the absence of non-organic, NMR-inactive species. NMR peak integration led toa rough composition of the MA:DA:TA as 1:9:90. The LC/HRMS analysis allowed the accurate determina-tion of DA and MA and confirmed in all samples the presence of the TA, which was estimated by differ-ence: MA = 1.4 ± 0.3%, DA = 11.1 ± 0.7%, TA = 87.5 ± 1.3%. The obtained results were used to setup an easy,rapid and cheap acid–base titration method able to furnish a sufficiently accurate evaluation of the activeprinciple both in the raw material and in diluted commercial products. For the raw material the resultswere: TA + MA = 91.1 ± 0.8% and DA �MA = 8.9 ± 0.8%, statistically coherent with LC/MS ones. The LC/MSapproach demonstrated also its great potentialities to recognize trace of the biocide components both inenvironmental samples and in the formulated commercial products.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

According to the modern definition, a ‘‘biocide’’ is a formulationcontaining one or more active substances that, even in very smalldoses, repels, controls or destroys harmful organisms. Biocidesare essential in our everyday life to prevent diseases and protecthuman health. They are subject to a dedicated set of regulationsand should not be confused with medicinal drugs, used to treat

the human body, or with plant protection products, directly usedon crops. More precisely, according to (‘‘Directive 98/8/EC of theEuropean Parliament and of the Council,’’ 1998), biocidal productsare ‘‘Active substances and preparations containing one or moreactive substances, put up in the form in which they are suppliedto the user, intended to destroy, deter, render harmless, preventthe action of, or otherwise exert a controlling effect on any harmfulorganism by chemical or biological means.’’ When biocides areused to remove or reduce bacteria on inanimate substances, theymay be called disinfectants or sanitizer, respectively. The problemof their accurate, quantitative determination is particularly

380 A. Mondin et al. / Chemosphere 95 (2014) 379–386

important in two main contexts: (1) to correctly formulate thecommercial product and (2) to assess their presence in the envi-ronment after use, together with their by-products as they canbe potentially harmful compounds or they can contribute to theselection of resistant bacteria (Lundén et al., 2003). There are sev-eral well-known classes of those products (see Table 1) and theirquantification is usually made via chromatographic methods (Sing-er et al., 2002; Hua et al., 2005; Coelhan et al., 2006; Hennekenet al., 2006; Martínez Bueno et al., 2007; Martínez-Carballo et al.,2007; Vincent et al., 2007; González-Mariño et al., 2009; Speksnij-der et al., 2010), consequently, the instrumental quantitative anal-ysis usually requires the availability of the active principles ascalibrant in the form of high purity standard. A well establishedand effective biocide is the N-(3-aminopropyl)-N-dodecyl-1,3-pro-panediamine(TA) a tri-amine present in the formulation of severalcommercialized disinfectants/cleaners (Domka et al., 1999;Chladkova et al., 2004). Although TA is registered as pesticides inthe US Environmental Protection Agency database (CAS number2372-82-9), no information about its toxicity or regulatory limitsin water are available in the same Web site (‘‘http://www.pesticid-einfo.org,’’ 2013). It has a broad activity spectrum against bothgram positive and gram negative bacteria, maintains high efficacyalso in presence of heavy organic soiling, such as blood and protein,is active against enveloped viruses (e.g. Hepatitis-B), good surfac-tant properties, compatible with selected anionic surfactants andhas a low viscosity. The substance is predicted to have a negligiblebioaccumulation potential because of its high solubility in waterand high biodegradation potential. Ecotoxicity studies indicatethe TA as moderately toxic to aquatic invertebrates and freshwaterfishes. No algal toxicity was evidenced. It has only rarely been rec-ognized as a contact allergen (Dejobert et al., 1997; Schliemannet al., 2010). The main problem associated to the quantitativedetermination of this active principle is that the pure standard isnot available, although its synthesis was described (Denton et al.,2007). The TA is present in the raw material together with two ma-jor impurities, the corresponding di-amine (N-dodecyl-1,3-pro-panediamine, DA) and monoamine (dodecylamine, MA). The rawmaterial is available for factories as LONZABAC 12-100 and it isused to formulate various widely used commercial products. Theabsence of the pure standard prevents the application of all analyt-ical methods usually employed for the quantification of alkyl-poly-amines (Ducros et al., 2009; Mayer et al., 2010; Triki et al., 2012) asboth direct detection and analyte derivatization protocols requirestandard solutions of the active principle. Moreover, the cited TAcannot be electrochemically quantified owing to a fast electrodepassivation as usually happens for analogous polyamines (Witekand Swain, 2001). Thus, none analytical method is nowadays pro-posed for the simultaneous determination of TA and its twoimpurities, DA and MA.

Table 1Main classes of biocides and usual analytical methods for their determination and quanti

Analyte Analysi

Esters of 4-hydroxybenzoic acid (parabens) LC–TMSQuaternary ammonium compounds (QACs) IPCb andOrtho-phenylphenol GC–MSd

2-(2,4-Dichlorophenoxy)-5-chlorophenol (triclosan) LC–TMSBiphenylol and chlorophene LC–TMSPeroxyacetic acid LC–DADIsothiazolinones LC–TMS

a Liquid chromatography–tandem mass spectrometry.b Indirect photometric chromatography.c Ionic chromatography.d Gas chromatography–mass spectrometry.e Gas chromatography–tandem mass spectrometry.f Liquid chromatography–diode array detection.

In the present study we will fulfill three tasks: (1) characteriza-tion of the LONZABAC by means of advanced NMR techniques withan internal standard; (2) optimization of a liquid chromatography/high resolution mass spectrometric (LC/HRMS) approach for thequantitative determination of the main biocide components witha low detection limit, able to detect the concentrations expectedin environmental samples; and (3) study of the equilibrium systemin order to develop a cheap and rapid titration procedure for ana-lyzing raw material and formulations based on TA for industrialinternal control purpose.

2. Experimental section

2.1. Reagents and chemicals

Deuterated dimethylsulfoxide (DMSO-d6, 0.75 mL vials, SigmaAldrich, Corp., St. Louis, MO, D-content P99.9%, H2O 620 ppm)and 1,1,2,2-tetrachloroethane (reagent grade, Sigma Aldrich, Corp.,St. Louis, MO), employed as internal standard, was used for theNMR measurements. Methanol and acetonitrile were of HPLCgrade (Fluka, AG, Buchs, Switzerland) and water was purified bya Millipore MilliQ equipment (18.2 MX cm�1, Millipore, Bedford,MA). Pentafluoropropionic acid (PFPA, 97%) and analytical stan-dards of DA and MA were from Sigma Aldrich. Individual stocksolutions of the two analytes were made by dissolving each com-pound in deionized water to obtain 1 mg mL�1 concentration.These solutions were stored at 4 �C in the dark. For LC/MS analysiscomposite working standard solutions of the two compounds wereobtained by mixing the above solutions and diluting with suitablevolumes of water 0.1% formic acid (v/v) to obtain the desired finalconcentrations. Perchloric acid (Carlo Erba, Milan, Italy) suitablystandardized was used for the potentiometric acid–base titrations.LONZABAC 12-100 and CAPTOSIL were supplied by MONDIAL s.n.c.(Limena, Italy).

2.2. Instrumentations and procedures

2.2.1. NMR analysis1H and 13C NMR spectra were obtained as DMSO-d6 solutions on

a Bruker DRX-600 Advance spectrometer operating at 600.01 and150.07 MHz, respectively. The chemical shift values are given in dunits with reference to Me4Si for both 1H and 13C. The assignmentsare according to the labeling reported in Fig. 1 relative to the TAhaving the formula C18H41N3. Suitable integral values for the pro-ton spectra were obtained with a pre-scan delay of 60 s to ensurethe complete relaxation for all the resonances. The signals wereintegrated manually between the 13C satellites after the baselinecorrection. The proton assignments were given by standard

fication.

s method

a González-Mariño et al. (2009)ICc Vincent et al. (2007), LC–TMSa Martínez-Carballo et al. (2007)(Coelhan et al. (2006)

a Hua et al. (2005), González-Mariño et al. (2009), GC–TMSe Singer et al. (2002)a Martínez Bueno et al. (2007)f Henneken et al. (2006)a Speksnijder et al. (2010)

H3C N

H2N

NH2

1

2

3 1'5

6

7

8

9

10

11

12 4 2'

3'

Fig. 1. N-(3-aminopropyl)-N-dodecyl-1,3-propanediamine (TA).

A. Mondin et al. / Chemosphere 95 (2014) 379–386 381

chemical shift correlations as well as by 2D correlation spectros-copy (COSY), total correlation spectroscopy (TOCSY), and nuclearOverhauser enhancement spectroscopy (NOESY) experiments.The 13C chemical shift values were obtained through 2D-heteronu-clear correlation experiments (heteronuclear multiple quantumcoherence, HMQC), with bilinear rotation-decoupling, BIRD, se-quence (Drobny et al., 1978; Bax and Subramanian, 1986) andquadrature along F1 achieved using the time-proportional receiverphase increment, TPPI, method (Otting and Wüthrich, 1988) for theH-bonded carbon atoms, and heteronuclear multiple bond correla-tion. HMBC (Bax and Summers, 1986) measurements were user toconfirm the HMQC correlations thorough the additional multiplebond coherences. Use of field gradients in both dimensions signif-icantly enhanced the quality of the 2D-experiments. In the quanti-tative determination using the 13C resonances, apart from a 120 srelaxation delay between the 90� pulses, the zgig pulse programof the Bruker library was chosen because it avoids the presenceof the heteronuclear Overhauser signal enhancement.

2.2.2. LC/MS analysisLC/quadrupole-Time of Flight (LC/QTOF)-MS analysis were per-

formed with an UHPLC system (Agilent Series 1200; Agilent Tech-nologies, Palo Alto, CA, USA), consisting of vacuum degasser, auto-sampler, a binary pump and a column oven coupled to both DADand QTOF-MS mass analyzer (Agilent Series 6520; Agilent Technol-ogies, Palo Alto, CA, USA).

The analytical column was an Alltima HyPurity 3.0-lm C-18(150 mm � 2.1 mm i.d., Alltech) coupled to a 5-lm C-18,7.5 � 2.1 mm i.d. guard cartridge. The separator system wasthermostated at 25 �C. The sample injected volume was 5 lL. Themobile phase components A and B were water and acetonitrile,respectively, both acidified with 10 mM PFPA. The eluant flow ratewas 0.2 mL min�1. The mobile phase gradient profile was as follow(t in min): t0, A = 10%; t9, A = 100%; t20, A = 100%; t21, A = 10%; t30,A = 10%. The QTOF system was equipped with an electrospray ion-ization interface (ESI), operating in dual ESI mode and positive ESIacquisition, with the following operation parameters: capillaryvoltage, 3500 V; nebulizer pressure, 35 psi; drying gas, 8 L min�1;gas temperature, 350 �C; fragmentor voltage, 170 V; skimmer65 V. Full scan mass spectra were recorded as centroid over therange 50–1000 m/z with a scan rate of 2 spectra/s. QTOF calibrationwas daily performed with the manufacturer’s solution. For all chro-matographic runs the m/z 391.28429 relative to the diisooctylphthalate molecular ion, always present as impurity residue, wasset as lock mass for accurate mass analysis. The instrument pro-vided a typical resolving power (FWHM) of ca 18000 at m/z311.0805. Mass spectra acquisition and data analysis was pro-cessed with Masshunter Workstation B 04.00 software (AgilentTechnologies, Palo Alto, CA, USA). Analytes quantification wasmade on the absolute area of the chromatographic signal corre-sponding to the extracted ion of the protonated molecular formulawith a windows span of 0.1 Da.

2.2.3. TitrationsAcid–base titrations of suitable amounts of raw product were

conducted in methanol/water 80/20 solvent. Titration end pointswere determined by first derivative of the potentiometric titrationperformed with a 5081 Crison (Allella, Spain) potentiometerequipped with a GLP 21 Crison glass electrode. Solutions werethermostated at 20 �C.

2.3. Samples

For titration experiment, 1 g of LONZABAC was diluted with50 mL of methanol/water80/20 and titrated with standardized0.1 M perchloric acid. The LONZABAC amount present in the com-mercially available product called CAPTOSIL, containing nominally0.2% LONZABAC, was also evaluated according to the followingprocedure: accurately weighed amount close to 75 g of 0.2% w/wCAPTOSIL was placed into a 500 mL beaker. 300 mL methanol/water80/20 were added and the mixture was pH-metrically ti-trated with perchloric acid.

For LC/MS analysis, samples of LONZABAC were diluted1:1,000,000 with an aqueous solution acidified with formic acidat 0.1% (v/v). Samples were freshly prepared prior to the analysisto prevent the degradation of the TA.

3. Results and discussion

The approach used to reliably determine the main componentsof the biocide raw material was:

(a) Characterization of the raw material composition (organicand inorganic species) and identification of the main com-pounds by NMR analysis.

(b) Determination of the organic compounds and quantitationof DA and MA, available as analytical standards, by HRMS.

(c) Setup of an easy to use acid–base titration method for deter-mination of the amines present in the raw material.

3.1. NMR analysis of the raw material

The quantitative determination of chemical species in mixtureby NMR spectroscopy requires the objective assignment of thespectral signals to all the species present in the mixture. Then,the molar ratio among the various species may be determined onthe basis of the peak integrations. Fig. 2 shows the 1H NMR spec-trum (recorded on a 600 MHz instrument) of a DMSO-d6 solutionof the commercial mixture containing nominally TA, DA, MA andother compounds at lower concentration levels. A series of quiteintense signals, at high magnetic field, is observed. In particular,a distorted triplet at d 0.830, a very intense and wide singlet at d1.22, a somewhat broaden quintet at d 1.329, a sharp quintet at d1.411, two triplets with widened central peak at d 2.270 and d2.326 and a sharp triplet at d 2.516 (these last triplets are better

Fig. 2. 1H NMR LONZABAC spectrum in DMSO-d6, mo = 600.01 MHz. Fig. 4. 13C NMR LONZABAC spectrum in DMSO-d6, mo = 150.07 MHz.

Table 21H and13C chemical shift assignment of LONZABAC (solvent, DMSO-d6; T 298 K;mo(1H) 600.01 MHz, mo(13C) 150.07 MHz. To mark the positions, the labels of Fig. 1 areused.

Position d(1H) d(13C)

1 2.270 54.542 1.332 27.623/10 1.22 27.13 29.45, 29.73 (2C), 30.06, 30.13, 30.69, 31.38 31.1211 1.247 23.0412 0.830 14.2910 2.326 52.1220 1.411 31.8830 2.516 40.27

382 A. Mondin et al. / Chemosphere 95 (2014) 379–386

visible in Fig. 3). These signals integrate in the 3:18:2:4:2:4:4 ratio,respectively, according to the TA structure. By comparison with the1H spectrum of an authentic sample of MA in the same solvent, thesignals at d 0.830, d 1.247, at d 1.329 and d 2.270 are assigned to thedodecyl chain of the TA (Denton et al., 2007). In particular, the trip-let at d 2.270 is due to the methylene protons in the a-position tothe nitrogen, as confirmed by the moderate peak broadening. Theremaining signals of the spectrum (d 2.516, d 2.326 and d 1.411),on the basis of their multiplicity and of their scalar correlationfound in the COSY and TOCSY spectra, are assigned to the twotri-methylene moieties of the two N-3-aminopropyl groups whichappear equivalent. Concerning the integrals of protons reported inFig. 3, signals at d 2.326 and d 2.516 have integrals twice largerthan that at d 2.270. By considering the molecular structure underinvestigation, the ratio suggests assignment of the signal at d2.270 ppm to the N-bonded methylene in the dodecyl chain whilstthe other two signals to the methylene moieties bonded to thenitrogen in the propyl chains. This is confirmed by the crossed sca-lar correlations in the two dimensional NMR spectroscopy. The 13CNMR spectrum of the same solution (recorded at mo 150.07 MHz) isreported in Fig. 4. The signals of the most abundant species are lo-cated at d 14.29, d 23.04, d 27.13, d 27.62, d 29.45, d 29.73, d 30.06, d30.13, d 30.69, d 31.12, d 31.38, d 40.27, d 52.12 and d 54.54.

The bi-dimensional 1H, 13C-HMQC spectrum allows correlatingevery proton with the directly-bonded carbon, and a complete listis given in Table 2. The 13C chemical shift values so obtained areagain coherent with the molecular structure of the TA. It is to notethat the signals of the N-bonded 13C nuclei are not pure singlets

Fig. 3. 1H NMR LONZABAC spectrum in DMSO-d6, mo = 600.01 MHz, d 2.6 – d 2.2zone and integrals.

and show two low-intensity satellite-like lines. In fact, this spectralshape is due to a second-order interaction with the spin I = 1 of the14N isotope. We consider this peculiarity useful to recognize easilythe N-bonded carbons, in addition to their chemical shift valueswhich usually fall in the d range of 60–40. As a final remark, thecorrect assignments of the methylene protons and carbons in the1 and 3 position of the 3-N-aminopropyl moiety have been madeby taking into consideration the long range 1H–13C correlations.In particular, we assigned the signal at d 52.12 to the 13C bondedto the ‘‘central’’ N-atom of the TA since only the corresponding pro-tons (d 2.326) show scalar correlation with the last methylene pro-tons of the dodecyl chain (d 2.270), whereas the methyleneprotons, resonating at d 2.516, do not show the same correlationand then are recognized as those near to the terminal NH2 moiety.In Fig. 3, two low intensity (about 9%) triplets are evidenced at d2.427 and d 2.481. They show scalar couplings with a quintet atd 1.432, and these three protons are respectively bonded to 13C nu-clei resonating at d 50.58 and d 48.41 (both of them showing theabove mentioned interaction with 14N) and at d 34.64. These reso-nances are probably due to the 3-aminopropyl substituent of theDA. The d values of the dodecyl chain signals are expected to notdiffer substantially from the corresponding signals of the TA, andin fact they are generally masked by those of the more abundantspecies. The low intensity triplet (about 1%) detected at d 2.481,connected with a 13C signal detected at d 42.67 is attributed tothe MA by comparison with a reference standard. Finally, we ob-served some very low intensity signals (about 0.2%, near to the lim-it value for a quantitative determination by NMR, a triplet at d 3.34,a singlet at d 3.194 and a quintet at d 1.54, in the integral ratio of2:6:6, respectively. The corresponding 13C resonances are observedat d 71.02, d 58.73 and d 34.14, respectively. Therefore, consideringthe d values, the integrals and the proton signal multiplicities, thisimpurity is very likely the 1,3-dimethoxypropane. Although the

A. Mondin et al. / Chemosphere 95 (2014) 379–386 383

NMR technique is mainly used for molecular structure identifica-tion, a rough quantification is still possible. On the basis of theabove discussion the estimated composition of the LONZABAC asMA:DA:TA was 1:9:90 in terms of percent in weight. Very similarresults were obtained in other four different lots of the LONZABAC.Finally, to determine with accuracy the total content of organicderivatives in the commercial product, a weighted amount of1,1,2,2-tetrachloroethane, was added as internal standard to asolution of LONZABAC. The 1H NMR spectrum of the solution wasrecorded in the fully-relaxed conditions and integrals were mea-sured between the 13C satellites. The LONZABAC signal at d 0.830was chosen for integration since the CH3 protons of the presentamines resonate isochronously at that field value. Taking also intoaccount the different molecular weight of the three TA, DA and MAcomponents, a total amine molecular amount >99% was measured,so the NMR-inactive specie concentration (inorganic salts, water,etc.) can be considered as not significant.

3.2. LC/MS analysis

The LC/QTOF-MS system was employed both to confirm theidentification of the TA performed with NMR analysis and to quan-titate DA and MA impurities present in the raw material. Becauseof their highly hydrophilic nature, the analysis of aliphatic poly-amines is always critic in a reversed-phase LC. The ion-pair chro-matography is normally used to retain such polar compounds ona C-18 column, using the relatively volatile trifluoroacetic acid(TFA), PFPA or heptafluorobutyric acid (Halsey and Burkin, 1962;Sánchez-López et al., 2009). The mass spectrometric parameterswere optimized using standard solutions of DA and MA at1 lg mL�1. With the aim to propose a protocol able to simulta-neously analyze TA, DA and MA, the chromatographic conditionsfor the retention and separation of the three principal componentshave been evaluated taking into consideration the content of theTA in the raw material. As TA has two primary and a tertiary aminegroups, it is reasonably the most difficult analyte to be retained ona C-18 column among the three compounds present in the mixture.So, a solution of LONZABAC diluted 1:1,000,000 was used to opti-mize the concentration of the ion-pair agent. This dilution factorwas chosen to simulate the expected concentrations in environ-mental samples and minimize the risk of the overloading and car-ry-over of the TA. All solutions and samples were diluted withformic acid at 0.1% (v/v) concentration to avoid the adsorption ofthe targeted compounds at low concentration on the silanolicgroups of the glassware (Fernández et al., 1996). In order to retainTA and the other two components on the analytical column, TFAand PFPA have been chosen as modifiers added to the mobilephases, taking care to minimize the signal suppression that suchmoderately volatile acids could provoke on ESI source. The resultsindicated that TFA added to the mobile phases even at concentra-tion of 0.1% (v/v) was not sufficient to efficiently retain TA and, to aminor extent, DA, as unacceptable broaden and tailed peaks wereobserved. Moreover, with this modifier the chromatographic con-ditions were not able to completely retain TA on the column asdemonstrated by a heavy carry-over of this compound. However,a higher concentration of TFA was not allowed due to the pH limitof the C-18 column. Both mobile phases were therefore added ofPFPA at 1 mM and 10 mM. The latter concentration was able tosolve this chromatographic issue producing a satisfactory peakshape for all components of the mixture and a reduced carry-overphenomenon. Fig. 5 reports a representative chromatogram of asample of LONZABAC diluted 1:1,000,000. When the molecularion signal of the MA was extracted from the total ion current chro-matograms with a mass window of 0.1 Da, two distinct peaks wereobserved at retention time of about 10.2 and 11.8 min (see Fig. 5).A further addition of the MA standard at the sample made it pos-

sible to recognize the signal at 11.8 min as the target analyte.The other interfering signal was attributed to the isobaric molecu-lar ion (m/z 186.2216) of the tributylamine, often used as modifierand present as residue of contamination. Anyway, the retentiontimes of the two compounds were quite different and DA andMA were identified through the comparison of both MS signalsand retention times of the sample with respect to the availableanalytical standards. Under these instrumental conditions, theidentification of TA in the sample was assured by the presence ofthe protonated molecular ion with the HRMS. The accuracy ob-tained for the experimental values related to the molecular for-mula of the TA (C18H41N3, theoretical mass 300.33783 m/z) wasin the range of 0.2–2.97 ppm for all the analysis performed. Thesignal related to the TA was very intense, with an estimated S/N va-lue of more than 1400 (see Fig. 5). This fact suggests that the ana-lytical method developed is potentially able to detect TA at verylow concentration both in formulations of LONZABAC and in envi-ronmental samples, taking into consideration the dilution factorused in this protocol. As regard as quantification of the DA andMA, two series of experiments were conducted to verify the possi-ble presence of a matrix effect, often occurring in the LC/ESI-MSanalysis. Two calibration curves were constructed using (a) stan-dard solutions of the two analytes and (b) different aliquots ofthe diluted LONZABAC spiked with DA and MA according to thestandard addition method. Both calibration curves were made at4 concentration levels, 5, 50, 100, 500 ng mL�1, corresponding toa percent range of the DA and MA of 0.5–50% of the raw material.The concentrations of the DA and MA in the sample of LONZABACobtained from the two different calibration approaches were com-parable, suggesting that no significant matrix effect was presentanalyzing samples of LONZABAC diluted 1:1,000,000. However,when polar analytes with a different number of amino-groupsare ionized in ESI source in the presence of an ion pair agent, theirresponse factor is not foreseeable, so that the concentration of theTA has to be assessed by difference from the other two amines. Theresults obtained for samples coming from five lots were compara-ble and indicated an average composition (n = 3) ofMA = 1.4 ± 0.3%, DA = 11.1 ± 0.8% and, by difference,TA = 87.5 ± 1.3%. The intervals represent the extended uncertainty.The results for both molecules are coherent with the NMR onesindicating the robustness of the proposed approaches. It has tobe pointed out that the use of the PFPA in place of TFA as ion-pair-ing agent decreases significantly the MS response. Anyway, theinstrumental limit of detections obtained for DA and MA, assessedas S/N of 3 from the lowest calibration level, were 2.4 and3.7 ng mL�1, respectively, corresponding to 0.24% and 0.37% inthe raw material (diluted 1/1,000,000), fit-for purpose for thedetermination of the two amines present as impurities in thebiocide.

3.3. Titrations

On the basis of combined results obtained by NMR and LC/MSanalysis, an easy and cheap method for the determination of theactive principle in the raw material, was carried out. The approachis based on a simple acid–base titration in mixed solvent. The in-volved reactions are resumed in Table 3. The seven reported reac-tions indicate that six end points are expected, namely, three dueto TA, two to DA and one to MA. Actually, we are aware that, forthe nature of the involved molecules, several constants are similarso that the real number of the end points will be lower than six. Apreliminary attempt to titrate the LONZABAC in water evidenced asingle end point indicating quite small differences among the pKa

values of all the involved reactions. This result indicates only thetotal amount of bases present in solution. That end point quantifiesthe 99 ± 2% of the total base equivalents corresponding to the sum

Fig. 5. LC/QTOF-MS representative chromatogram related to the analysis of a sample of LONZABAC diluted 1:1,000,000 reporting in (A) the extracted ion chromatogram ofthe TA, DA and MA (extraction window 0.1 Da); (B) extracted MS spectrum of the TA-related chromatographic peak. N-(3-aminopropyl)-N-dodecyl-1,3-propanediamine (TA);N-dodecyl-1,3-propanediamine (DA); dodecylamine (MA).

Table 3Chemical Equilibriums accounted for the titration of the LONZABAC system.

Chemical reaction Equilibrium constant

1 HClO4 + H2O = H3O+ + ClO�4 K12 C12H28N+ + H2O = C12H27N + H3O+ (Ka1)MA

3 C15H37N2þ2 þ H2O ¼ C15H36Nþ2 þ H3Oþ (Ka1)DA

4 C15H36Nþ2 + H2O = C15H35N2 + H3O+ (Ka2)DA

5 C18H44N3þ3 þ H2O ¼ C18H43N2þ

3 þ H3Oþ (Ka1)TA

6 C18H43N2þ3 þ H2O ¼ C18H42Nþ3 þ H3Oþ (Ka2)TA

7 C18H42Nþ3 þ H2O ¼ C18H41N3 þ H3Oþ (Ka3)TA

Table 4pKa values of N-(3-aminopropyl)-N-dodecyl-1,3-propanediamine, N-dodecyl-1,3-propanediamine, n-dodecylamine in water and in methanol/water 80/20. MA andDA pKa values were determined from standard solutions. TA pKa values were obtainedfrom the model. All pKa values are averages of 4 measurements.

% MeOH (pKa1)TA (pKa2)TA (pKa3)TA (pKa1)DA (pKa2)DA (pKa1)MA

0 6.7(0.3) 8.4(0.3) 10.0(0.3) 5.9(0.5) 8.2(0.5) 9.5(0.5)80 6.4(0.3) 9.2(0.3) 10.2(0.3) 5.7(0.5) 9.1(0.5) 9.7(0.5)

Fig. 6. Titration curves of the raw solution in terms of pH vs. the normalized totalamount of bases present, U. (h) Titration curve of 1.0250 g raw solution withperchloric acid. (s) Titration curve of 1.0230 g raw solution with perchloric acid inMethanol/water 80/20 (v/v). Continuous lines represent the regression curvesobtained from the model of the seven equations reported in Table 4. Experimentalconditions: perchloric acid CH = 0.1990 M. The mathematical model is the titrationof a mixture of the three studied weak bases with strong acid.

384 A. Mondin et al. / Chemosphere 95 (2014) 379–386

of MA, DA and TA estimated by LC/MS and also according to theNMR analysis. In order to increase the difference among the pKa

values, titrations with perchloric acid were performed in metha-nol/water solvent. The methanol/water ratio was varied from 0%to 90%. In Table 4 the pKa values obtained with methanol/water

0/100 and 80/20 are reported. A significant difference is evidencedand the relevant experimental data are reported in Fig. 6, showingthe two titration curves of the LONZABAC in water (squares) and inmethanol/water 80/20 (circles). The U axis represents the titratedfraction, that is, the ratio between the amount of titrated bases and

A. Mondin et al. / Chemosphere 95 (2014) 379–386 385

the perchloric acid concentration, CH. A single end point is presentin water whilst, two of them are clearly visible in the mixed sol-vent. The overlapped regression lines were obtained with theparameters reported in Table 4 relative to the 80/20 methanol/water and come from the model of the seven equations abovementioned, that is the titration of weak bases with strong acid inmixed solvent. The titration end points were not sufficient todetermine all the pKa values but succeed in determining the TAfrom some known parameter values. In particular, the pKa valueswere obtained from the available standards of DA and MA. ThepKa value of MA was confirmed also from literature data (Matulisand Bloomfield, 2001). The pKa values of TA were obtained fromthe model. The TA:DA:MA ratios determined by NMR and LC/MSwere used as starting parameters in order to accurately determinethe pKa values of TA and to interpret the titration curve shape. Itmust be noted that, although NMR, LC/MS and titration techniquesare independent one another, the titration is the weakest. For thisreason it needed some preliminary information to work correctly.The nature of the titrated species may be determined from theend points and from the proposed model. In particular, if Ui isthe titrated fraction of the single amine, then UTA, UDA and UMA

are directly related to the amount of TA, DA and MA, respectively,through CH. From this definition the second inflection point repre-sents the total amount of titrated bases:

3UTA þ 2UDA þUMA ¼ U ¼ 1 ð1Þ

The first inflection point, according to the pKa values reported inTable 4, represents

2UTA þUDA þUMA ð2Þ

so that the difference between the (1) and (2) is

3UTA þ 2UDA þUMA � ð2UTA þUDA þUMAÞ ¼ UTA þUDA ð3Þ

From the difference between the (2) and (3) we obtain

2UTA þUDA þUMA � ðUTA þUDAÞ ¼ UTA þUMA

From this result, the TA value obtained by titration is overesti-mated of the MA amount. As the MA amount is usually close to 1%(as assessed by both NMR and LC/MS analysis and determined infive different lots of LONZABAC), this result is acceptable for thequality control of the raw material with the purpose of a correctformulation of the commercial products, in which TA is usuallypresent at 1% or less (and consequently MA � 0.01%). However, ifbetter accuracy is required, LC/MS analysis is needed. End pointswere determined by the first derivative of the titration curve withrespect to the titrant added volume. From the above discussion it isevident that DA may be also quantified, although always underes-timated of the MA amount. From the titration curves we obtain thevalue of TA + MA = 91.1 ± 0.8% and DA �MA = 8.9 ± 0.8%. Thesevalues are coherent to the LC/MS ones, as proved by statisticalcomparison. Finally, the titration of a commercially available prod-uct named CAPTOSIL and containing 0.2% LONZABAC as nominalamount was also made, in the same conditions as above. The LON-ZABAC percentage was directly obtained from Eq. (4) by knowingthe composition of the LONZABAC raw material used to preparethe commercial product (%MA, %DA, %TA), the molecular weightof the three amines MWMA ¼ 185;MWDA ¼ 242;MWTA ¼ 299),from the CAPTOSIL weight in grams (wCAPTOSIL) and from the sec-ond end point volume in liter, VEP.

%LONZABAC ¼ 100 � CHClO4 � VEP

a �wCAPTOSIL¼ 100 � 0:09 � 0:015

0:0098 � 72:5¼ 0:19% ð4Þ

In this equation a is the number of base equivalents per 100 g ofLONZABAC and accounts for the stoichiometric factor of the con-sidered amines, m,

a ¼X3

i¼1

%Ai � v i

MWi¼ %MA � 1

MWMAþ%DA � 2

MWDAþ%TA � 3

MWTA

� �,100

¼ 0:0098 eq=g LONZABAC ð5Þ

As the extended uncertainty is 0.015% with a coverage factor,k = 2, the obtained result contains the declared nominal valueand demonstrates its effectiveness to test the commercial product.It must be noted that the CAPTOSIL formulation includes also thepresence of various surfactants that do not interfere with the titra-tion procedure. Degradation phenomena may be hypothesized ifanomalous ratios between the two inflection points are present.Even in this case an LC/MS analysis is suggested.

4. Conclusions

In the present paper we reported the quantitative analysis ofthe tri-amine N-(3-aminopropyl)-N-dodecyl-1,3-propanediaminepresent in a raw material named LONZABAC and used to formulatevarious, widely used commercial biocides, without the availabilityof the pure standard. At the same time the other two main compo-nents of the mixture, the N-dodecyl-1,3-propanediamine and n-docecylamine were also quantified. Three approaches were used:the NMR technique used as absolute method to determine themain components of the raw material; the LC/MS analysis for highresolution identification of TA and for the quantitation of DA andMA impurities in the raw material; the acid–base titrations as easyand rapid method for testing the biocide raw material as well asthe commercial formulation. The NMR gave a comparable compo-sition of various lots of LONZABAC equal to 1:9:90 as MA:DA:TA.The presence of non-NMR active species was excluded by using asuitable internal standard. Results obtained by LC/HRMS analysisand acid–base titration method in methanol/water 80/20 were sta-tistically comparable. Thus, titration could be useful as a cheapmethod for quality assurance purpose in the formulations of TA-based biocides, while LC/MS technique demonstrated great poten-tialities to recognize all the three compounds at the very low con-centration levels expected for environmental samples andformulated commercial products.

Acknowledgement

We gratefully acknowledge the MONDIAL s.r.l. (Limena, PD,Italy) for funding this research.

References

Bax, A., Subramanian, S., 1986. Sensitivity-enhanced two-dimensionalheteronuclear shift correlation NMR spectroscopy. J. Mag. Reson. (1969) 67,565–569.

Bax, A., Summers, M.F., 1986. Proton and carbon-13 assignments from sensitivity-enhanced detection of heteronuclear multiple-bond connectivity by 2Dmultiple quantum NMR. J. Am. Chem. Soc. 108, 2093–2094.

Chladkova, K., Hendrych, T., Gaskova, D., Goroncy-Bermes, P., Sigler, K., 2004. Effectof Biocides on S. cerevisiae: relationship between short-term membraneaffliction and long-term cell killing. Folia Microbiol. 49, 718–724.

Coelhan, M., Bromig, K.-H., Glas, K., Roberts, A.L., 2006. Determination and levels ofthe biocide ortho-phenylphenol in canned beers from different countries. J.Agric. Food Chem. 54, 5731–5735.

Dejobert, Y., Martin, P., Piette, E., Thomas, P., Bergoend, H., 1997. Contact dermatitisfrom didecyldimethylammonium chloride and bis-(aminopropyl)-laurylaminein a detergent–disinfectant used in hospital. Contact Dermatitis 37. pp. 95–95.

Denton, T.T., Joyce, A.S., Kiely, D.E., 2007. Preparation of N-Alkylbis(3-aminopropyl)amines by the Catalytic Hydrogenation of N-Alkylbis(cyanoethyl)amines. J. Org. Chem. 72, 4997–5000.

Directive 98/8/EC of the European Parliament and of the Council. 1998. Directive98/8/EC of the European Parliament and of the Council Official Journal of theEuropean Communities.

Domka, F., Brycki, B., Seifert, K., Szymanska, K., 1999. Microbiocide efficiency ofAPDA in denitrification and desulfurication processes. Polym. J. Environ. Stud. 8,299–303.

386 A. Mondin et al. / Chemosphere 95 (2014) 379–386

Drobny, G., Pines, A., Sinton, S., Weitekamp, D.P., Wemmer, D., 1978. Fouriertransform multiple quantum nuclear magnetic resonance. Faraday Symp. Chem.Soc. 13, 49.

Ducros, V., Ruffieux, D., Belva-Besnet, H., de Fraipont, F., Berger, F., Favier, A., 2009.Determination of dansylated polyamines in red blood cells by liquidchromatography–tandem mass spectrometry. Anal. Biochem. 390, 46–51.

Fernández, P., Alder, A.C., Suter, M.J.-F., Giger, W., 1996. Determination of thequaternary ammonium surfactant ditallowdimethylammonium in digestedsludges and marine sediments by supercritical fluid extraction and liquidchromatography with postcolumn ion-pair formation. Anal. Chem. 68, 921–929.

González-Mariño, I., Quintana, J.B., Rodríguez, I., Cela, R., 2009. Simultaneousdetermination of parabens, triclosan and triclocarban in water by liquidchromatography/electrospray ionisation tandem mass spectrometry. RapidCommun. Mass Spectrom. 23, 1756–1766.

Halsey, G., Burkin, A.R., 1962. Adsorption of long-chain aliphatic amines on glassvessels. Nature 193, 1177–1178.

Henneken, H., Assink, L., de Wit, J., Vogel, M., Karst, U., 2006. Passive sampling ofairborne peroxyacetic acid. Anal. Chem. 78, 6547–6555.

Hua, W., Bennett, E.R., Letcher, R.J., 2005. Triclosan in waste and surface waters fromthe upper Detroit River by liquid chromatography–electrospray-tandemquadrupole mass spectrometry. Environ. Int. 31, 621–630.

Lundén, J., Autio, T., Markkula, A., Hellström, S., Korkeala, H., 2003. Adaptive andcross-adaptive responses of persistent and non-persistent Listeriamonocytogenes strains to disinfectants. Int. J. Food Microbiol. 82, 265–272.

Martínez Bueno, M.J., Agüera, A., Gómez, M.J., Hernando, M.D., García-Reyes, J.F.,Fernández-Alba, A.R., 2007. Application of liquid chromatography/quadrupole-linear ion trap mass spectrometry and time-of-flight mass spectrometry to thedetermination of pharmaceuticals and related contaminants in wastewater.Anal. Chem. 79, 9372–9384.

Martínez-Carballo, E., Sitka, A., González-Barreiro, C., Kreuzinger, N., Fürhacker, M.,Scharf, S., Gans, O., 2007. Determination of selected quaternary ammoniumcompounds by liquid chromatography with mass spectrometry. Part I.Application to surface, waste and indirect discharge water samples in Austria.Environ. Pollut. 145, 489–496.

Matulis, D., Bloomfield, V.A., 2001. Thermodynamics of the hydrophobic effect. I.Coupling of aggregation and pKa shifts in solutions of aliphatic amines. Biophys.Chem. 93, 37–51.

Mayer, H.K., Fiechter, G., Fischer, E., 2010. A new ultra-pressure liquidchromatography method for the determination of biogenic amines in cheese.J. Chromatogr. A 1217, 3251–3257.

Otting, G., Wüthrich, K., 1988. Efficient purging scheme for proton-detectedheteronuclear two-dimensional NMR. Journal of Magnetic Resonance (1969)76, 569–574.

Sánchez-López, J., Camañes, G., Flors, V., Vicent, C., Pastor, V., Vicedo, B., Cerezo, M.,García-Agustín, P., 2009. Underivatized polyamine analysis in plant samples byion pair LC coupled with electrospray tandem mass spectrometry. Plant Physiol.Biochem. 47, 592–598.

Schliemann, S., Zahlten, A., Krautheim, A., Elsner, P., 2010. Occupational allergiccontact dermatitis caused by N-(3-aminopropyl)-N-dodecylpropane-1,3-diamine in a surface disinfectant. Contact Dermatitis 63, 290–291.

Singer, H., Müller, S., Tixier, C., Pillonel, L., 2002. Triclosan: occurrence and fate of awidely used biocide in the aquatic environment: field measurements inwastewater treatment plants, surface waters, and lake sediments. Environ.Sci. Technol. 36, 4998–5004.

Speksnijder, P., van Ravestijn, J., de Voogt, P., 2010. Trace analysis ofisothiazolinones in water samples by large-volume direct injection liquidchromatography tandem mass spectrometry. J. Chromatogr. A 1217, 5184–5189.

Triki, M., Jiménez-Colmenero, F., Herrero, A.M., Ruiz-Capillas, C., 2012. Optimisationof a chromatographic procedure for determining biogenic amine concentrationsin meat and meat products employing a cation-exchange column with a post-column system. Food Chem. 130, 1066–1073.

Vincent, G., Kopferschmitt-Kubler, M., Mirabel, P., Pauli, G., Millet, M., 2007.Sampling and analysis of quaternary ammonium compounds (QACs) traces inindoor atmosphere. Environ. Monit. Assess. 133, 25–30.

Witek, M.A., Swain, G.M., 2001. Aliphatic polyamine oxidation response variabilityand stability at boron-doped diamond thin-film electrodes as studied by flow-injection analysis. Anal. Chim. Acta 440, 119–129.

http://www.pesticideinfo.org [WWW Document]. 2013. URL <http://www.pesticideinfo.org/Detail_Chemical.jsp?Rec_Id=PC39670#Water>(accessed 25.05.13).