Accepted Manuscript Cationic Gemini Surfactants with Cleavable Spacer: Chemical Hydrolysis, Biodegradation, and Toxicity A.R. Tehrani-Bagha, K. Holmberg, C.G. van Ginkel, M. Kean PII: DOI: Reference:

S0021-9797(14)00721-8 http://dx.doi.org/10.1016/j.jcis.2014.09.072 YJCIS 19880

To appear in:

Journal of Colloid and Interface Science

Received Date: Accepted Date:

25 August 2014 29 September 2014

Please cite this article as: A.R. Tehrani-Bagha, K. Holmberg, C.G. van Ginkel, M. Kean, Cationic Gemini Surfactants with Cleavable Spacer: Chemical Hydrolysis, Biodegradation, and Toxicity, Journal of Colloid and Interface Science (2014), doi: http://dx.doi.org/10.1016/j.jcis.2014.09.072

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Cationic Gemini Surfactants with Cleavable Spacer : Chemical Hydrolysis, Biodegradation, and Toxicity A.R. Tehrani-Bagha ΨΦ, K. Holmberg Φ*, C.G. van Ginkel Ωand M. Kean Ω Φ-Department

of Chemical and Biological Engineering, Chalmers University of Technology, SE-412 96 Göteborg, Sweden Ψ- Chemical Engineering Department, Faculty of Engineering and Architecture, American University of Beirut (AUB), Beirut 1107 2020, Lebanon Ω- AkzoNobel, Ecotoxicology & Environmental Testing Department, PO Box 9300, 6800 SB Arnhem, The Netherlands

* Corresponding author: Krister Holmberg, email: [email protected]; tel: +46 31 772 2969; fax: +46 31 16 0062

Keywords: Gemini surfactant, cleavable, hydrolysis, biodegradation, toxicity

1

Abstract The paper describes synthesis and characterization of a new type of cationic gemini surfactant, which has dodecyl tails and a spacer that contains an ester bond. The nomenclature used to describe the structure is 12Q2OCO1Q12, with Q being a quaternary ammonium group and the numbers indicating the number of methylene or methyl groups. Due to the close proximity to the two quaternary ammonium groups, the ester bond is very stable on the acid side and very labile already at slightly alkaline conditions. The hydrolysis products are two single chain surfactants (i.e. 12Q2OH and 12Q1COOH) which are less surface active than the intact gemini surfactant. 12Q2OCO1Q12 was found to be readily biodegradable, i.e. it gave more than 60 % biodegradation after 28 days. This is interesting because similar gemini surfactants but with ester bonds in the tails instead of the spacer, have previously been found not to be readily biodegradable. The gemini surfactant was found to be toxic to aquatic organisms (ErC50 value of 0.27 mg/l), although less toxic than the two hydrolysis products.

2

Introduction Gemini (dimeric) surfactants have been subject to considerable scientific interest during the last two decades. They are usually much more efficient than the corresponding monomeric surfactants. The critical micelle concentration (CMC) of a gemini surfactant is typically 10-30 times lower than that of a regular monomeric surfactant with the same hydrocarbon chain length [1-3]. This means that a gemini surfactant is likely to give good coverage of a surface already at very low bulk concentration. This is practically interesting because it enables the use of unusually low amounts of surfactant to accomplish a specific effect. For instance, we have shown that a cationic gemini surfactant containing dodecyl tails, a tetramethylene spacer and bromide as counterion -

(12-4-12, 2Br in the gemini surfactant nomenclature) retains its ability to protect mild steel from corroding in acidic media down to concentrations several orders of magnitude lower than is required for the corresponding monomeric surfactant, dodecyltrimethylammonium bromide [4-6].

There has been a particular interest in cationic gemini surfactants, partly because their synthesis from readily available starting materials is straight-forward. The physicochemical properties are generally attractive for many applications but it has been found that these gemini surfactants do not comply with the current environmental regulations. In an effort to increase the rate of biodegradation we have previously synthesized and characterized a large number of gemini surfactants with ester bonds inserted in the hydrocarbon tails. Two series of such amphiphiles have been investigated, one with the ether oxygen of the ester bond turned towards the quaternary ammonium group and one with the reverse orientation of the ester bond [7-9]. These two structures are referred to as ester quat type and betaine ester type, respectively, and they are shown in Figure 1 together with the structures of the corresponding monomeric ester-containing surfactants. We have reported a range of physicochemical properties of the estercontaining surfactants (e.g. micelle aggregation numbers, shape and size of micelles, adsorption at solid surfaces, dye solubilization) in previous papers [9-12].

3

s

s

Figure 1. Monomeric and dimeric ester quat surfactants (left) and betaine ester surfactants (right). R is C9H19 or C11H23 and R′ is C12H25. The number of methylene groups in the spacer unit of the gemini surfactants, s, is 2, 3, 4 or 6.

The rate of chemical hydrolysis of these two types of ester-containing surfactants was studied in some detail and it was found that (i) the dimeric surfactants degraded faster than their monomeric counterparts and (ii) the betaine ester type amphiphiles (geminis as well as monomeric surfactants) were more unstable than the amphiphiles of ester quat type [7, 8]. Insertion of the ester bonds in the structure did not result in readily biodegradable surfactants (i.e. biodegradability over 60% in 28 days), regardless of the orientation of the ester bond. Whereas the monomeric surfactants of ester quat type, as well as of betaine ester type, could be classified as readily biodegradable, none of the ester-containing geminis fulfilled the requirements for this classification. It was demonstrated that this slow biodegradation rate was due to the dicationic species, formed from hydrolysis of the two ester bonds, being very resistant to degradation. This was true for both the diquat diol generated from the ester quat gemini and the diquat dicarboxylate formed through hydrolysis of the betaine ester gemini.

Since the approach of inserting ester bonds in the hydrophobic tails of gemini surfactants proved unsuccessful as a way to achieve rapid biodegradation, we instead turned our attention to manipulation of the spacer unit. In this paper, we present synthesis and

4

physicochemical properties of a gemini surfactant with an ester bond inserted into a short spacer. To the best of our knowledge this type of gemini structure has not been investigated before. We also present the chemical hydrolysis characteristics, as well as the biodegradation profile and the aquatic toxicity, of this novel surfactant.

1. Experimental Section 2.1 Materials. 2-Bromoethanol (95%), bromoacetic acid (97%), bromoacetyl bromide (>98%), N,N-dimethyldodecylamine (97%), dichloromethane, sodium bicarbonate, magnesium sulfate, ethanol, acetone and diethyl ether were all purchased from SigmaAldrich. Deuterium oxide (99.8 atom% D) and 1 M DCl in D2O were purchased from Dr. Glaser AG (Basel, Switzerland).

2.2 Syntheses. 2.2.1. N-dodecyl-N-(2-hydroxyethyl)-N,N-dimetylammonium bromide (12Q2OH). 2Bromoethanol (110 mmol, 13.74 g) in acetone was added drop-wise to a stirred solution of N,N-dimethyldodecylamine (100 mmol, 21.34 g) in acetone under reflux. The reaction mixture was stirred for 1 day under reflux. The mixture was then cooled and the solvent was removed by filtration. The recovered product was recrystallized twice from ethanol/diethyl ether, giving a yield of 73%. 1H NMR (400 MHz, CDCl3): δ 0.88 (t, 3H), 1.2-1.45 (m, 18H), 1.76 (m, 2H), 3.38 (s, 6H), 3.54 (t, 2H), 3.76 (t, 2H), 4.16 (t, 2H), 5.01 (t, 1H).

2.2.2. N-carboxymethyl-N-dodecyl-N,N-dimethylammonium bromide (12Q1COOH). Bromoacetic acid (55 mmol, 7.64 g) in acetone was added drop-wise to a stirred solution of N,N-dimethyldodecylamine (50 mmol, 10.67 g) in acetone under reflux. The reaction mixture was stirred for 1 day under reflux. The mixture was then cooled and addition of diethyl ether resulted in a white precipitate. The solvent was removed by filtration and the recovered product was recrystallized twice from ethanol/diethyl ether, giving a yield of 35%. 1H NMR (400 MHz, CDCl3): δ 0.88 (t, 3H), 1.1-1.4 (m, 18H), 1.86 (m, 2H), 2.79 (s, 3H), 2.80 (s,3H), 3.00 (m, 2H), 4.75-5 (d, 2H), 11.49 (s,1H)

5

2.2.3. 2-Bromoethyl bromoacetate. 2-Bromoethanol (100 mmol, 12.49 g) was dissolved in dichloromethane (30 ml) and charged into a round-bottomed flask. Potassium carbonate (130 mmol, 17.96 g) in water (30 ml) was subsequently added. The flask was cooled in an ice bath and bromoacetyl bromide (130 mmol, 26.24 g) in dichloromethane (30 ml) was added drop-wise to the flask. The reaction was terminated after 2 h and the organic phase was separated, dried over magnesium sulfate, filtered and evaporated to give the product in 81% yield. 1H NMR (400 MHz, CDCl3): δ 3.54 (t, 2H), 3.88 (s, 2H), 4.49 (t, 2H).

2.2.4. N-dodecyl-N-[2-(N’-dodecyl-N’,N’-dimethylammonio)acetyloxyethyl]-N,Ndimethylammonium dibromide (12Q2OCO1Q12). 2-Bromoethyl bromoacetate (80 mmol, 19.67 g) in acetone was added drop-wise to a stirred solution of N,Ndimethyldodecylamine (176 mmol, 37.55 g) in acetone under reflux. The reaction mixture was stirred for 1 day under reflux. The mixture was then cooled and the solvent was removed by filtration. The recovered product was recrystallized twice from ethanol/diethyl ether to give a yield of 53%. 1H NMR (400 MHz, CDCl3): δ 0.88 (t, 6H), 1.1-1.4 (m, 36H), 1.75 (m, 4H), 3.49 (s, 6H), 3.53 (s, 6H), 3.67 (t,2H), 3.77 (t,2H), 4,14 (t,2H), 4.95 (t,2H), 5.41 (s,2H).

1.3. Physicochemical characterization 2.3.1. Determination of critical micelle concentration (CMC). The CMC values for the surfactants were determined by tensiometry and conductometry in Millipore water. A Sigma 70 tensiometer (KSV Instrument LDT, Helsinki, Finland), equipped with an automatic dispenser model Metrohm® dosimat 765, was used for the surface tension determinations and the measurements were repeated for each concentration until a maximum deviation of 0.10 mN/m was obtained.

The conductometry measurements were performed with a CDM 210 conductometer (Radiometer®, France). An exact volume of 10 ml Millipore water was introduced into the vessel and the specific conductivity was measured. The solution was then titrated

6

with a surfactant solution with known concentration and the conductivity was measured after each addition. Plots of specific conductivity (κ) vs. surfactant concentration (C) were constructed and the concentration at which there was a break on the curve was taken as the CMC [13-15]. It should be noted that the conductometry method could not be used for the amphoteric surfactant 12Q1COOH. 2.3.2. Preparation of phosphate buffer in D2O. 100 mM phosphate buffer was prepared by mixing 50 mL of 200 mM potassium dihydrogen phosphate, 30 mL of 200 mM sodium hydroxide and 20 ml of D2O. The glass electrode pH meter (744 Metrohm ® , Switzerland), which was calibrated in H2O-based buffers, showed a pH of 7.1, which means that the effective pD value was approximately 7.5 [22].

1.4. Chemical hydrolysis The chemical hydrolysis of the gemini surfactant with cleavable spacer (12Q2OCO1Q12) was monitored by 1H NMR. Surfactant solutions of different concentrations in phosphate D2O buffer were prepared and held in glass vials at 25 °C under constant stirring. For quenching the reaction, a 550 µl sample was taken from the surfactant solution at different times and added to a NMR tube containing 50 µL 1 M DCl. The NMR spectra were recorded on Agilent 500 MHz NMR spectrometer equipped with auto-sampler. The degree of hydrolysis at different times was calculated from the integrals originating from the N-methyl protons of the intact gemini surfactant and of the hydrolyzed product. The N-methyl protons of the gemini surfactant appear as singlets at δ = 3.24 ppm and δ= 3.11 ppm. During the chemical hydrolysis these peaks gradually disappear and new peaks at δ=3.02 ppm δ=3.09 emerge.

1.5.Biodegradation test The Closed Bottle test was used for evaluating the ready biodegradability of the gemini surfactant (12Q2OCO1Q12) and its hydrolysis products. The test method has been described in detail before [16, 17] and will be discussed only briefly here. The test employed 2.0 mg/l of test substance present in river water spiked with mineral salts.

7

Ammonium chloride was omitted from the mineral salts prescribed to prevent nitrification. The river water was aerated for 7 days before use and particles were removed by sedimentation. The tests were performed in 0.3 l BOD (biological oxygen demand) bottles with glass stoppers, using 3 bottles containing only river water, and series of 3 bottles containing the respective test substances and river water. The bottles were incubated in the dark at a temperature of 23±1˚C. The biodegradation was assessed by following the course of the dissolved oxygen decrease in the bottles with a special funnel [17]. The amount of dissolved oxygen was measured electrochemically. The biodegradability was calculated as the ratio of the BOD to the theoretical oxygen demand.

1.6.Toxicity test The toxicity of 12Q2OCO1Q12, its hydrolysis products 12Q1COOH and 12Q2OH, and a 50:50 mixture of 12Q1COOH and 12Q2OH, were tested on exponentially growing cultures of Pseudokirchneriella subcapitata over a 72 hour period. The test methods were based on the OECD 201 guideline (OECD 2006) [18]. Endpoints were based on nominal concentrations of the test materials. Toxicity endpoints were expressed as ErC50 values, which are the concentrations showing 50% reduction in the growth rate of the test organisms. The details of the method used are provided in Supporting Information section.

2. Results and Discussion 3.1.Syntheses and physicochemical characterization The synthesis routes for the two monomeric surfactants (12Q2OH and 12Q1COOH) and for the gemini surfactant with an ester bond in the spacer (12Q2OCO1Q12) are shown in Scheme 1. The preparation of 12Q2OH and 12Q1COOH from the fatty amine was straight-forward. The synthesis of the spacer unit of the gemini surfactant, i.e. 2bromoethyl bromoacetate, was more challenging because the generated ester bond is very -

susceptible to attack by a nucleophile, such as OH , leading to hydrolysis. The literature proposes the use of pyridine or trimethylamine to catalyze the reaction and to neutralize the HBr formed [18]. That procedure involves a multi-step work-up procedure that may lower the final yield. Instead we explored the use of an aqueous solution of K2CO3 to 8

neutralize the resulting HBr. This lead to a much simplified work-up without compromising the yield compared to the conventional procedure.

12Q1COOH 12Q2OH

-HBr 12Q2OCO1Q12

Scheme 1. Synthesis routes for preparation of the surfactants used in this study. Thermogravimetric analysis showed that 12Q2OH, 12Q1COOH, and 12Q2OCO1Q12 have melting/degration point 258, 244, and 252 °C, respectively. The CMC values for the surfactants were determined by tensiometry and conductometry and the results are reported in Table 1. The gemini surfactant had the lowest surface tension at the CMC among the three surfactants investigated. Also this is in accordance with many previous comparisons between gemini surfactants and the corresponding monomeric surfactants [2].

The CMC value for the gemini surfactant (~ 1 mM) is around 12 times lower than the CMC value for the corresponding monomeric cationic surfactant (12Q2OH) and very similar to those of regular cationic gemini surfactants (without an ester bond in the spacer) with the same alkyl chain length (12-s-12, s=2-6) [19, 20]. The area per molecule when aligned at the air-water interface was calculated from Gibbs adsorption equation using Equation 1:

9

1Γ×ܰ‫=ܣ‬−ܴ݊ܶܰ‫ܶ )ܥ݈݊ ݀ߛ ݀(× ܣ‬−1

(Eq. 1)

R is the gas constant, T the absolute temperature, NA Avogadro’s number, and n a constant that takes the values 1 for nonionic surfactants, 2 for univalent-univalent ionic surfactants, and 3 for divalent-univalent ionic surfactants. We used n=2 for the monomeric surfactants 12Q2OH and 12Q1COOH and n=3 for the gemini surfactant 12Q2OCO1Q12. When γ is in mN/m and R =8.31 J mol-1 K-1, Γis in 10-3 mol.m-2 [2, 21].

The n value assigned to an ionic surfactant is not obvious, however. If the counterion is not fully dissociated from the surfactant ion, the value becomes smaller than the theoretical value. (If the counterion for a univalent-univalent ionic surfactant is fully bound, the value of n becomes 1, i.e. the surfactant is seen as a nonionic amphiphile.) Neutron reflectivity is a useful technique to determine area per molecule for surfactants aligned at an interface and from these results n values can be calculated. Using this procedure a value of n=2 was found for regular cationic gemini surfactants with a short spacer and a value of n=3 was obtained for cationic gemini surfactants with a longer spacer unit [22]. The gemini surfactant studied in this work has five atoms between the quaternary ammonium groups, which is a relatively long spacer. A value of n=3 therefore seems reasonable. The cross-sectional areas of 12Q2OH and 12Q1COOH were found to be 53.8 Å2 and 75.8 Å2, respectively. These are reasonable values compared to values reported for other ionic surfactants. The area occupied by the gemini surfactant, 12Q2OCO1Q12, was surprisingly small, 78.6 Å2. (Using a value of n=2 instead would give a cross-sectional area of 52.4 Å2, which is totally unrealistic.) The low value indicates (i) that the gemini surfactant packs tightly at the air-water interface and (ii) that the bromide ions are fully dissociated. The presence of an ester bond in the spacer may also contribute to the small area per molecule. An ester bond imparts hydrophilicity to the spacer. Whereas spacer units with only methylene groups are known to align at the interface up to at least 5 or 6 carbon atoms and then, when the number of methylene groups increases further, form a loop in the air, an ester-containing spacer is likely to make a turn into the aqueous sub-

10

phase instead. It has been shown before that geminis with a hydrophilic spacer occupies a smaller area per molecule at the air-water interface than geminis with a hydrophobic spacer [23]. Thus, the presence of the ester bond in the spacer of the gemini surfactant 12Q2OCO1Q12 may force the two quaternary ammonium groups closer together.

The ionization degree of surfactant micelles can be studied by conductometry. As can be seen from Table 1, the values obtained for 12Q2OH and 12Q2OCO1Q12 are close to the values from tensiometry. This is in accordance with previous results from monomeric and gemini surfactants with ester bonds situated not in the spacer but in the tails [7, 8, 10]. The fact that the ionization degree is virtually the same for the gemini surfactant and for the monomeric surfactant 12Q2OH is an indication that the micelles at the CMC have the same geometry. The degree of ionization is related to the curvature of aggregates, the higher the curvature, the higher is the degree of ionization. No attempts to characterize the micelles further have been made in this work, however.

The gemini surfactant had the lowest surface tension at the CMC among the three surfactants investigated. Also this is in accordance with many previous comparisons between gemini surfactants and the corresponding monomeric surfactants [2, 19, 20, 24].

Table 1. Physicochemical properties of the surfactants obtained from surface tension and conductometry measurements at 25°C. γCMC is the surface tension at the CMC, dγ/d lnC the slope of the surface tension curve just before the CMC, a the area occupied by a surfactant molecule at the air-water interface calculated from Gibbs adsorption equation, and αthe ionization degree of the surfactant micelles. tensiometry Surfactant 12Q1COOH 12Q2OH 12Q2OCO1Q12

CMC (mM) 3.7 11.3 0.92

γCMC (mN/m) 35.2 32.4 30.8

10.75 15.16 15.57

a െ (Å2) 75.8 53.8 78.6

conductometry CMC α (mM) 13.9 1.04

0.25 0.26

* The standard errors for the CMC and the γCMC values are ± 0.04 and ± 0.3 mN/m, respectively.

11

3.2.Chemical hydrolysis Hydrolysis of the gemini surfactant 12Q2OCO1Q12 yields the monomeric amphiphiles 12Q2OH and 12Q1COOH. This reaction was studied in D2O buffer and monitored by 1H NMR (Figure 2). The figure shows a typical example of stacked 1H NMR spectra at various time intervals. The N-methyl protons of the starting gemini surfactant appear as singlets at δ= 3.11 ppm and δ= 3.24 ppm. During the chemical hydrolysis these peaks gradually disappear and new peaks at δ=3.02 and δ= 3.09 appear. These represent Nmethyl protons from 12Q2OH and 12Q1COOH, respectively. The sum of the integrals of the peaks remains constant during the course of the chemical hydrolysis. The degree of hydrolysis was calculated from the relative integrals of the signal at δ= 3.02 ppm and the signal at δ= 3.24 ppm.

The hydrolysis of an ester-containing surfactant in buffered solution can be expected to obey first-order kinetics (Equation 2), as we have discussed in previous papers [7, 8]. -

Hydrolysis rate = k × [OH ] × [Surf.] = kobs× [Surf.]

(Eq. 2)

[Surf.] is the total or stoichiometric concentration of surfactant, kobs the observed firstorder rate constant, and k the apparent second-order rate constant. The concentration of hydroxyl ions, [OH-], is constant in a buffered solution.

The hydrolysis experiments were carried out in D2O instead of H2O and it is known that due to isotope effects the hydrolysis is likely to proceed somewhat faster in D2O than in H2O [25]; thus, the absolute values of the rate constants obtained in D2O may be different from the values in H2O. However, the relative hydrolysis rates and the trend should be the same in H2O as in D2O.

12

60 min 45 min 30 min 20 min 15 min 10 min 8 min 6 min 4 2 min

Figure 2. Stacked 1H NMR spectra of hydrolysis of 12Q2OCO1Q12 in phosphate buffer in D2O (pD 7.5) at various time intervals. The appearing peak at δ= 3.02 ppm and the disappearing peak at δ= 3.24 ppm were used for monitoring the reaction.

The effect of pH on the hydrolysis rate constant was also investigated. As can be seen from Figure 3, above a pH of 3 the rate constant increases exponentially. We have studied hydrolysis of cationic ester-containing gemini surfactants before but the ester bonds were then situated in the tails, not in the spacer [7]. It was then found that the hydrolysis rate was considerably higher for such ester-containing gemini surfactants than for estercontaining monomeric surfactants, for instance so-called ester quats. We then discussed the mechanism for the hydrolysis reaction in some detail [7, 8]. Also the ester-containing

13

gemini surfactant of this work, with the ester bond situated in the spacer unit, undergoes rapid alkaline hydrolysis, as can be seen from the figure. This is the expected behavior since the two quaternary ammonium groups, and in particular the group on the carbonyl side of the ester bond, which is only one carbon atom away, withdraw electrons from the ester group. This makes the carbonyl carbon unusually electron deficient, which means that it becomes prone to nucleophilic attack by hydroxyl ions. Acid hydrolysis, on the other hand, would involve protonation of the carbonyl oxygen, which would be a very unfavorable intermediate because of the close proximity between this positive charge and the positive charge at the adjacent nitrogen atom. As a result, the ester bond in the gemini surfactant 12Q2OCO1Q12 is very stable in acid and very labile under alkaline conditions. The proposed hydrolysis mechanisms are shown in Scheme 2.

The gemini surfactant 12Q2OCO1Q12 is consequently a good example of a cleavable surfactant for which the degradation rate can be controlled by only small variations of the pH of the surfactant solution [26]. The hydrolysis generates a 1:1 molar mixture of a cationic and an amphoteric surfactant. Thus, the degradation products are also amphiphilic although not as surface-active as the starting gemini surfactant. The CMC of the 1:1 mixture is 5.4 mM and the surface tension at the CMC is 33 mN/m, which can be compared with the values 0.92 mM and 31 mN/m, respectively, for the intact gemini surfactant.

14

k (1/min)

00.00188 00.00166 00.00144 00.00122 0.011 00.00088 00.00066 00.00044 00.00022 0 0

2

4 H pH

6

8

Figurre 33. E Effect oof ppH oon thee ratte cconsstannt, kk, ffor ccheemical hyydroolyssis oof tthe gem minni suurfactannt 112Q Q2O OCO O1Q Q122. OH+ N

δ+

O

N

N

+

O

C122H255

δ−

+

N

+

O

C122H255

C112H25

H+

+

unfav vora able e in nterrme ediate e

C12H25

O

O OH

N

Q OC CO Q

–

N

+

O OH

C12 H25

+

C122 H25 2 O

N

+

N O H + HO

C12H25 2

Q O OH

+

C12 H25

Q CO OOH H

Scchem me 2. Acid aandd baase cataalyzedd hyydroolysis of thee geeminni ssurffacttantt 122Q22OC CO1Q12. Thhe ddottted bluue aarroowss illuustrratee a pulll off ellecttronns bby thhe qquaaterrnarry nnitroogeen aatom ms aw wayy froom thee caarboonyyl grrouup oof thhe eesteer liinkaagee.

15

The plot of chemical hydrolysis of the gemini surfactant 12Q2OCO1Q12 as a function of surfactant concentration at pH 7.1 and 25 °C is shown in Figure 4. As can be seen, the curve has a maximum at the CMC. This is the expected behavior and similar concentration dependence has been encountered before for other ester-containing surfactants [7, 27]. The positively charged micelles attract negative ions, including hydroxyl ions, and as a result the pH will be higher close to the surface than in the bulk. This acceleration of hydrolysis of ester-containing cationic surfactants is often referred to as micellar catalysis [27-29].

The rate of hydrolysis of the gemini surfactant studied here is higher than the hydrolysis rates for the gemini surfactants with ester bonds inserted in the tail regions of the surfactant (see the structures in Figure 1) [7, 8]. There are probably two reasons for this. Firstly, as discussed above, the ester bond of surfactant 12Q2OCO1Q12 is destabilized by a pull of electrons from two sides. The gemini surfactants with ester bonds in the tails had only one quaternary ammonium group adjacent to the ester bond. Secondly, the ester bond in 12Q2OCO1Q12 is situated in the head-group region while the ester bonds in the previously studied gemini surfactants were within the hydrophobic part of the molecule, a few atoms away from the head-group. It is reasonable to assume a better access of hydroxyl ions to the carbonyl carbon of the ester bond when that bond is situated in the spacer unit than in the tails.

As the surfactant concentration is increased above the CMC the increasing concentration of surfactant counterions (bromide in this case) will compete more and more favorably with the hydroxyl ions for a place around the micelle surface. Bromide ions are more polarizable and thus more chaotropic than hydroxyl ions. This means that they will interact more strongly with the micelle. The net effect is that the bromide ions will gradually expel the hydroxyl ions from the micelle surface layer, which means that the micellar catalysis will be less prominent [30]. With chloride or acetate as counterion instead of bromide one would expect the hydrolysis rate to be faster because these anions are more kosmotropic, which means that they do not compete so well with hydroxyl ions for a place at the micelle surface [31-33].

16

k (1/min)

0.002 00.0115 0.001 00.0005 0 0

1

2

3

4 5 C C/C CM MC

6

7

8

9

Figurre 44. E Effect oof ssurffacttantt coonceentrration, noorm malizzedd to thee CM MC C, oon thhe ppseeudoo-fiirst-orrderr ratte cconstannt rratee in phoospphatte bbufffer (pH H 7..1) at 225°C C.

CO1Q12 genneraatess tw wo m monnom merric Sincee hyydroolysis of thee geeminni ssurffacttantt 122Q22OC mphhiphhilees, oonee peermaneentlly ccatioonicc, 112Q Q2O OH, annd oone am mphooterric,, 122Q11CO OOH H, am miixed m miceellees com mpossedd of all thrree speeciees w willl forrm if tthe totaal ssurffacttantt coonceentrratiion is aaboove thee CM MC C off thee m mixtturee. A As thhe hhyddrollysiis pproggresssess, thhe rrelaative am mouunt of tthe two m monnom meriic aampphipphilles in tthe micellles willl inncreease. T This m meanns tthatt thee m miceelle surrfacce w willl grraduuallly bbecoome leess possitivvelyy chharrgedd, ggoinng ffrom m oone poositiive chaargge pper hhyddropphoobicc taiil w wheen thhe ggem minni suurfaactaant iis inntaact tto oone nett poositiive chaargge pper ttwoo taiils whhen thee deegraadattionn is comppleteed, asssum mingg a 1:11 occcuppanncy in tthe miicellles of thee tw wo ddeggraddatioon prooduccts.. Onne ccann exxpecct thhat -

122Q11CO OOH Hw willl bee coomppletely deeprootonnateed iin thhe m mixxedd miicellle, i.e. bee 122Q11CO OO . Thhis graaduaal ‘diluutioon’ of thee poositiive chaarges oon tthe miicellle ssurffacee m makees tthe efffectt of thee m miceellaar caatallysiis ddeclinee as thee reeacttionn prrogrressses..

17

3.3.Biodegradation Cationic surfactants are known for their toxicity to microorganisms mainly targeting cell membranes. Possible toxic effects by the test substances are usually detected prior to the onset of the biodegradation through suppression of the endogenous oxygen consumption [17]. Inhibition of the endogenous respiration by the cationic gemini surfactant or by its hydrolysis products was not noted. All three cationic amphiphiles were therefore considered to be non-toxic in the Closed Bottle test.

Both the starting gemini surfactant and the two degradation products, 12Q2OH and 12Q1COOH, gave more than 60 percent biodegradation after 28 days (Figure 5). According to OECD’s guidelines 60 % biodegradation, calculated on the theoretical amount, should be reached within 28 days under aerobic conditions in order for a surfactant to be classified as readily biodegradable [33]. All three substances can therefore be classified as readily biodegradable. Ready biodegradability of betaines with alkyl chains ranging from 12 to 16 carbons has previously been demonstrated in the Sturm test, another ready biodegradability test [34]. The half-life of the gemini surfactant in water is less than an hour under neutral or slightly alkaline conditions (this study) and the pH values in the Closed Bottle tests were 8.0±0.1. Thus, the gemini surfactant is most likely first hydrolysed and the hydrolysis products are then subject to microbial oxidation.

As mentioned above, we have seen in a previous study that ester linkages inserted in the hydrophobic tails of cationic gemini surfactants resulted in less than 60% biodegradation at day 28 [7]. The degradation products (i.e. N,N’-bis(2carboxymethyl)-N,N,N’,N’-tetramethyl-1,2-ethane ammonium and N,N’-bis(2hydroxyethyl)-N,N,N’,N’-tetramethyl-1,2-ethane ammonium) were found to be recalcitrant diquats [7]. Alkyldiamines with only four substituents like ethylendiaminetetraacetate (EDTA) are also not readily degraded [35, 36]. When striving towards ultimately biodegradable cationic gemini surfactants, a structural modification in the bridge between the two entities therefore seems to be key.

18

Inntrooduuctioon oof a hyydroolyyticaallyy instabble estter bbonnd iin thhe spaacerr is succh aan aappproaach buut ootheer cchem miccal moodifficaationns aare also cconcceivvabble. 1 0 100 900

Biodegradation (%)

800 700 600 500 400 300 200 100 0 0

7

1 14

21 1

28

daays Q2O OH, annd Bioddeggraddatiion proofile of thhe ssurffacttantts : () 12Q1COO OH, ( ) 112Q Figurre 55. B ( withhin 28 ) 112Q Q2O OCO O1Q Q12. Alll thrree surrfacctannts passs thhe 660 % bbiodeggraddatiion linne w daays,, whhichh m meanns tthatt they cann bee classifieed aas reeaddily bioodeegraadabble.

3.4.T Toxxiciity In generral tterm ErC550 ms, cattionnic surrfacctannts aare verry ttoxiic too aqquaaticc orggannism ms w withh E w 1 mgg/L.. Allthoouggh vaaluees foor ddapphniia aand alggae speeciees oofteen bbeinng ssignnificcanntly bellow fissh sspecciess arre inn geeneeral thee leeast sennsittivee ouut oof thhe sstanndarrd tthreee inndicator sspeecies, matiion efffectt leevells arre oofteen in thhe 11- 110 m mg//l raangge [337]. Thherre iss litttle conncluusivve iinfoorm wevver, it iis reggarrdinng thhe m mechaanissmss off thee tooxiccityy off cattionnicss to aqquattic oorganisms. H How hyypotthesizeed tthatt thhe pposiitivee chhargge oof tthe heaadggrouups leaads to acccum mulaatioon oof suurfactannt m mollecuuless att thee suurfaace of celll m mem mbraanes. P Prevvenntionn oof trranssporrt aacrooss thee ceell m mem mbrranee, w wheetheer itt is by phyysiccal bloockiingg or by speeciffic binndinng tto trranspoort prroteeins, leeadss to deeathh off thee ceell. In aaddditioon, thee hyydroophhobic ssurffacttantt taiils aare

19

35

expected to be capable of disrupting the lipid bilayer of cell membranes, also leading to cell death [38, 39].

Toxicity measurements were carried out as described in the Supporting Information section. The control replicates gave satisfactory results according to the guideline criteria. ErC50 values with confidence limits were calculated for all the substances.

Table 2 shows toxicity data for the gemini surfactant and Figure 6 illustrates how the ErC50 value has been obtained. In the Supporting Information section, corresponding tables and graphs for the other surfactants can be found (S1-S3).

Table 2. Toxicity results for the gemini surfactant 12Q2OCO1Q12. Surfactant Concentration mg/l 0 0.0095 0.031 0.098 0.31 1

Specific growth rate

CV%

Normalized Average

0.0696 0.0709 0.0612 0.0535 0.0276 0.0193

5.297 1.247 2.575 4.268 2.927 0.000

0.0000 -0.0182 0.1203 0.2309 0.6035 0.7234

20

Response p

1.0 0 0.9 0 0.8 0 0.7 0 0.6 0 0.5 0 0.4 0 0.3 0 0.2 0 0.1 0 0.0 1E-0 09

ErrC50 = 0.2 27m mg//l

0 000 0.0 001 1

0..001

1

10 000 0

10 000 000 00

1E+ +09 9

Surfa Su acttan onc centrrattion (m (mg/l) nt Co Figurre 66. T Toxiicityy ddataa poointss foor thhe ggem Q2O OCO1Q12 w withhin 95% % mini suurfaactaant 12Q coonfiidenntiaal innterrvall lim mitss acccorrdinng tto a maxi methhodd baasedd m imuum likeelihhood pprobbabiilityy m onn the grow wth ratte eendppoinnt.

Taablee 3 shoowss toxiccity datta ffor tthe thrree surrfacctannts aandd foor thhe 11:1 mixxture oof tthe deegraadattionn prrodductts, eexpressedd as ErrC50 vvaluues. Thhe vvaluues aree low w aand alll thee suubsttancces shoouldd bee coonssideeredd veery toxxic to aalgaae aandd woouldd be cllasssifieed as A Acuute 1 acccorrdinng tto thhe G Glooballly Haarmoniizedd Syysteem (G GHS S) of cllasssificcatiion andd laabellingg off chhem micaals.

Taablee 3.. ErC550 vvaluuess forr thhe surffactantts. Teest Subbstaancce 122Q11CO OOH H 122Q11CO OOH H+ +12Q Q2O OH H (11:1 mixxtuure) 122Q22OH H 122Q22OC CO1Q12

ErC500 m L mg/L 0.0522 0.0577 0.15 0.27

Evvenn if aall thee ErrC50 vvaluues aree low w, iindicattingg coonssideerabble ttoxxicitty, ttherre aare cleear diffferrenccess beetweeenn thee suubsttancess in thiis reespect. It is ppartticuularrly iinteeresstinng thhat thee

21

gemini surfactant, 12Q2OCO1Q12, is more than four times less toxic than the mixture of the two degradation products. This can be practically important and may be taken advantage of in terms of classification. One may speculate that a better hydrolytic stability of the gemini surfactant may improve the toxicity profile further.

Finally it is important to keep in mind that the toxicity data, and the comparison between different surfactants classes, relate to tests performed with equal weight or molar concentrations. The fact that gemini surfactants can normally be used in one order of magnitude lower concentration than normal monomeric surfactants needs to be considered in order to make a relevant comparison.

22

4. Conclusions In this paper we have shown that inserting an ester bond in the spacer unit of a cationic gemini surfactant is advantageous from a biodegradation point of view in comparison to inserting ester bonds in the two hydrophobic tails. A gemini surfactant based on dodecyl tails and with an ester bond in the spacer can be classified as readily biodegradable, which is important from a regulation point of view. Similar geminis but with ester bonds in the tails instead have previously been found not to be readily biodegradable [7, 8]. The aquatic toxicity of the gemini surfactant with an ester bond in the spacer is quite high but still considerably lower than the toxicity of the mixture of the degradation products, a single chain cationic surfactant and an amphoteric amphiphile.

The ester bond in the new gemini surfactant is extremely pH sensitive because of its proximity to two cationic charges. The molecule is very stable on the acidic side and very labile on the alkaline side. This opens for possibilities to use the compound as a ‘cleavable surfactant’: formulating and employing it at low pH, where it is hydrolytically stable and then raising the pH so that it immediately decomposes into two less surface active species.

Acknowledgements The authors wish to thank Mr. Hans Oskarsson at AkzoNobel Surface Chemistry for valuable discussions. References [1] F.M. Menger, J.S. Keiper, Angew. Chem. Int. Ed. 39 (2000) 1906. [2] R. Zana, J. Xia, Gemini Surfactants. Marcel Dekker, New York, 2004. [3] R. Zana, M. Benrraou, R. Rueff, Langmuir 7 (1991) 1072. [4] M. Mahdavian, A.R. Tehrani-Bagha, K. Holmberg, J. Surfactants Deterg. 14 (2011) 605. [5] M. Motamedi, A.R. Tehrani-Bagha, M. Mahdavian, Corros. Sci. 70 (2013) 46. [6] M. Motamedi, A.R. Tehrani-Bagha, M. Mahdavian, Electrochimica Acta 58 (2011) 488. [7] A.R. Tehrani-Bagha, H. Oskarsson, C.G. van Ginkel, K. Holmberg, J. Colloid Interf. Sci. 312 (2007) 444. 23

[8] A.R. Tehrani-Bagha, K. Holmberg, Langmuir 26 (2010) 9276. [9] A.R. Tehrani-Bagha, K. Holmberg, Langmuir 24 (2008) 6140. [10] A.R. Tehrani-Bagha, J. Kärnbratt, J.-E. Löfroth, K. Holmberg, J. Colloid Interf. Sci. 376 (2012) 126. [11] A.R. Tehrani-Bagha, R.G. Singh, K. Holmberg, J. Colloid Interf. Sci. 376 (2012) 112. [12] A.R. Tehrani-Bagha, K. Holmberg, M. Nyden, L. Nordstierna, J. Colloid Interf. Sci. 405 (2012) 145. [13] M. Fujiwara, T. Okano, T.-H. Nakashima, A.A. Nakamura, G. Sugihara, Colloid Polym. Sci. 275 (1997) 474. [14] R. Zana, J. Colloid Interf. Sci. 246 (2002) 182. [15] G. Sugihara, A.A. Nakamura, T.-H. Nakashima, Y.-I. Araki, T. Okano, M. Fujiwara, Colloid Polym. Sci. 275 (1997) 790. [16] OECD, Guidlines for Testing of Chemicals, Degradation and Accumulation, No. 301, Ready Biodegradability; Paris Cedex France, 1992. [17] C.G. van Ginkel, C.A. Stroo, Ecotox. Environ. Saf. 24 (1992) 319. [18] OECD, Guideline 201. Freshwater Alga and Cyanobacteria, Growth Inhibition Test Paris Cedex France, 2006. [19] A.R. Tehrani-Bagha, H. Bahrami, B. Movassagh, M. Arami, S.H. Amirshahi, F.M. Menger, Colloids and Surfaces A. 307 (2007) 121. [20] A.R. Tehrani-Bagha, H. Bahrami, B. Movassagh, M. Arami, F.M. Menger, Dyes Pigments 72 (2007) 331. [21] M.J. Rosen, Surfactants and Interfacial Phenomena. 3rd ed., Wiley, New York, 2004. [22] Z.X. Li, C.C. Dong, R.K. Thomas, Langmuir 15 (1999) 4392. [23] K. Holmberg, in: Novel Surfactants: Preparation, Applications, and Biodegradability; Marcel Dekker, Inc., New York, 2003. [24] L. Liu, M.J. Rosen, J. Colloid Interf. Sci. 179 (1996) 454. [25] T.C. Bruice, T.H. Fife, J.J. Bruno, P. Benkovic, J. Am. Chem. Soc. 84 (1962) 3012. [26] A.R. Tehrani-Bagha, K. Holmberg, Curr. Opin. Colloid Interface Sci. 12 (2007) 81. [27] R.A. Thompson, S. Allenmark, J. Colloid Interface Sci. 148 (1992) 241. [28] C.A. Bunton, L.B. Robinson, J. Schaak, M.F. Stam, J. Org. Chem. 36 (1971) 2346. [29] D. Lundberg, K. Holmberg, J. Surfactants Deterg. 7 (2004) 239. [30] E. Leontidis, Curr. Opin. Colloid Interface Sci. 7 (2002) 81. [31] K.D. Collins, G.W. Neilson, J.E. Enderby, Biophys. Chem. 128 (2007) 95. [32] N. Vlachy, B. Jagoda-Cwiklik, R. Vácha, D. Touraud, P. Jungwirth, W. Kunz, Adv. Colloid Interface Sci. 146 (2009) 42. [33] B. Kronberg, K. Holmberg, B. Lindman, Surface Chemistry of Surfactants and Polymers. Wiley, Chichester, 2014. [34] G.W. Fernley, J. Am. Oil. Chem. Soc. 55 (1978) 98. [35] P. Pitter, V. Sykora, Chemosphere 44 (2001) 823.

24

[36] M. Buecheli-Witschel, T. Egli, FEMS Microbiol. Rev . 25 (2001) 69. [37] G.-G. Ying, in: U. Zoller, Handbook of Detergents Part B: Environmental Impact Marcel Dekker Inc., 2004. [38] G. McDonell, A.D. Russel, Clin. Microbiol. Rev. 12 (1999) 147. [39] T. Cserhati, Environ. Health Persp. 103 (1995) 358. [40] M.A. Partearroyo, S.J. Pilling, M.N. Jones, Comp. Biochem. Phys. A 101 (1992) 653.

25

CMC=11.3 mM

OHHydrolysis

CMC=0.92 mM CMC=3.7 mM

Biodegradation (%)

100 90 80 70 60 50 40 30 20 10 0

All three Surfactants readily biodegradable

0

7

14

21 days

28

Highlights: - A cationic gemini surfactant with an ester bond in the spacer has been synthesized and characterized - The new surfactant was readily biodegradable but the aquatic toxicity was relatively high - The new surfactant is very surface active and breaks down into two less surface active amphiphiles

26