Introduction

Surfactant-free colloidal syntheses of nanoparticles (NPs) are promising synthetic strategies to develop more active materials, e.g. for electrocatalysis [1,2,3]. The use of simple and safe chemicals facilitates the successful supporting of the NPs to develop heterogeneous catalysis and alleviate washing and/or steps of surfactant removal [4]. In particular, the recent development of the green synthesis of surfactant-free NPs in low viscosity media with low boiling point solvents obtained by a low temperature process [5, 6] bears promising features to simply yet efficiently develop active electrocatalyst [6, 7]. The term surfactant-free can be challenging to define, and we here adopt the definition proposed recently, where a colloidal synthesis is considered surfactant-free if no other chemical than the metal precursor required has a molar mass superior to 100 g mol−1 [4].

Following this definition, NP synthesized using plant extracts or other nature-derived extracts are excluded. We previously commented on the benefits of avoiding plant extracts to ensure more reproducible results and green syntheses, for instance in the synthesis of gold (Au) NPs [5]. Indeed, obtaining plant extracts often require organic chemicals and/or solvents; therefore, the common approach of using plant based of food-based extracts to prepare Au NPs [8] might not be as green as expected in the first place. Moreover, plants/fruits from different parts of the world will have different properties which can make the syntheses challenging to reproduce. In addition, the more chemicals used, the more likely the presence of impurities detrimental to optimize the reproducibility of NP syntheses [9].

These drawbacks are at least partially addressed in surfactant-free colloidal syntheses where only simple chemicals such as water, a base and an alcohol are used [5]. However, reproducibility in the chemical sciences remains a general concern and challenge [10]. Often, seemingly irrelevant experimental details actually lead to significantly different results, if they are not properly controlled and addressed. An example is Au NPs and the Turkevich-Frens synthesis using citrate as reducing agent and stabilizer [11], where the order of the addition of the chemicals strongly influences the size and size distribution of the resulting NPs [12]. Detailed protocols are certainly necessary to alleviate this possible source of irreproducibility [13]. Another source of irreproducibility comes from the purity of the chemicals used [9]. The latter can affect for instance the structural and/or catalytic properties of the NPs [9, 14]. The effect of impurities can be especially important in surfactant-free colloidal syntheses where mainly electrostatic interactions and stabilization by small molecules are observed [2, 15]. If fundamental research will prefer the use of well-defined and high purity chemicals, scaling up and real-life applications might benefit from lower purity alternatives, typically cheaper, provided the quality and performances of the NPs are not detrimentally affected.

We reported a simple approach to obtain small size (5–10 nm) Au NPs using water as solvent, HAuCl4 as metal precursor and ethanol, or alternative alcohols, as the source of alkoxides that will play the role of reducing agents under slightly alkaline conditions [5, 16]. We studied the effect of different alcohols, alcohol contents, nature of the base, base concentration and temperature as well as other variables such as the type of containers used, stirring, volumes, order of addition of the chemicals and effect of the concentration of the stock solution of HAuCl4 [5, 7]. The optimal synthesis is performed at room or low (< 40 °C) temperature with 0.5 mM HAuCl4, 2 mM NaOH and ca. 20–30 v.% ethanol, and HAuCl4 is preferentially added last to the reaction mixture from a concentrated stock solution. Various nanomaterials including bimetallic NPs can be obtained by this approach [5] and the use of a low viscosity solvents such as ethanol is directly relevant to more easily process the NPs to obtain active electrocatalysts [7].

We previously observed that using relative fresh solutions base in water, ideally stored in plastic containers, lead to more reproducible syntheses [17]. Here, we investigate the influence of the water conductivity and the grade of the ethanol source used on the resulting properties of the Au NPs. For the influence of the water purity, experiments performed using glycerol [18, 19] or ethylene glycol (EG) [20] are also considered. An overview of the experimental parameters investigated is given in Table 1 and detailed in Supporting Information (SI); the correlation between UV-vis and STEM data is also detailed in SI. These parameters were selected based on the recent development of this synthetic approach [5, 7]. The use of the three different bases, LiOH, NaOH or KOH, is here considered to assess the robustness of the syntheses and because the NPs tend to be more stable as the size of the cation decreases in the order Li+ > Na+ > K+, due to the expected stronger interactions between the smaller cation and metal surfaces [5, 21, 22].

Table 1 Overview of the experimental parameters. PS stands for polystyrene
Full size table

Results and discussion

Effect of water purity

A first clear conclusion from the parametric studied performed over 100 samples (×6 alcohol sources, ×2 water source, ×3 bases, at least triplicates) is that when deionized water (DI) with a lower resistivity than high purity water (mQ) is used, see SI for more details, no stable colloidal Au NP dispersions are obtained, regardless of the source of reducing agents used: ethanol, EG or glycerol. The stable colloidal Au NPs obtained using mQ water show a red colour as illustrated in Fig. 1a–c, indicative of small size Au NPs, whereas the dispersions obtained using DI water lead to unstable darker materials, indicative of larger materials. For the syntheses performed using ethanol and EG, the black deposit is formed on the wall of the reactor. This in contrast to our previous results, for which using DI water led to the successful synthesis of Au NPs [5]. We recently observed that the success of the synthesis depends on the origin of the DI water used: DI water from different building lead in some cases to stable colloids and in other cases not. This stresses the need to carefully control water purity. This discrepancy is attributed to different water quality from different laboratories, in agreement with a reported fact that water purity can have a strong impact on Au nanomaterial synthesis [10, 13]. The importance of high purity water (e.g. mQ) is to be related to the electrostatic stabilization of the surfactant-free NPs [2, 5].

Fig. 1
figure 1

Illustrative (a–c) pictures and (d) UV-vis data for Au nanomaterials prepared using mQ or DI water, as indicated, for a synthesis using (a) ethanol, (b) EG or (c) glycerol, as indicated. In (a–c), the UV-vis cuvette is 1 cm wide. e Illustrative bright field STEM micrograph of the Au NPs prepared using ≥ 99.8% grade ethanol and mQ water. See also Figs. S1–S2

Full size image

The UV-vis characterization of the dispersions reported in Fig. 1d shows that using mQ water leads to Au NPs with a characteristic surface plasmon resonance around 520 nm. The use of DI water leads to rather large materials without pronounced plasmonic properties. The red dispersions are indeed made of Au NPs ca. 5–10 nm as illustrated in Fig. 1e and Fig. S1 in SI. The interpretation of the different values retrieved from UV-vis spectroscopy is detailed elsewhere [5] and in SI; see Table S1. The size estimation is in agreement with previous reports [5] and with the evaluation of the size of the NPs after 7 months of storage in a fridge at ca. 5 °C reported in Figs. S2–S3.

These results stress the need to control water purity for the successful synthesis of surfactant-free Au NPs at low temperature. This is also true for the viscous media obtained using glycerol and accounts for the use of mQ water also preferred in previous reports [18, 19, 23]. It must be commented that using high purity water, more expensive to obtain, might go against the principles of Green Chemistry for NPs synthesis [24], but remains a greener approach compared to using hazardous chemicals or solvents.

Effect of ethanol grade

Having established that mQ water is more likely to lead to stable Au NPs in a reproducible way, we now turn to the effect of different grades of ethanol sources. Four different grades of ethanol are used as detailed in Table 1, referred to as E99.8%, E99.5%, E96% and E70%, where the under-scripted values refer to the grade of the stock solution of ethanol used (see details in SI). The price of the solvent tends to decrease as the ethanol content decreases and it is therefore relevant for larger scale applications, and ultimately for solvent recycling, to investigate the actual need for absolute ethanol (≥ 99.8%) as starting ethanol source.

Based on the ʎspr values (see details in SI) reported in Fig. 2, no significant difference between the different ethanol grades used is observed. Using glycerol, the lower ʎspr values (below 515 nm) indicate that slightly smaller NPs are obtained, probably due to the high viscosity of this solvent, in agreement with previous work [18, 25] and also in agreement with STEM analysis in Figs. S1–S2. Using other parameters such as Aspr/A450, A650/Aspr, A380/A800 or A400 values, defined and detailed in Table S1, a similar conclusion is reached; see Fig. S4. The grade of the ethanol source does not influence much the properties of the Au NPs that all show ʎspr values around 518–521 nm, corresponding to ca. 10 nm NPs [5]; see also Fig. S2. It can be added, in agreement with previous reports [5, 21], that using KOH leads to slightly larger NPs, probably due to a poorer stabilization of the NPs. Note that the possible presence of different impurities in LiOH, NaOH and KOH cannot be excluded at this stage. The results nevertheless confirm that NaOH can be substituted to LiOH, to lead to greener syntheses and yet yield small size Au NPs [7].

Fig. 2
figure 2

ʎspr values for the parametric study performed when different alcohols and bases are used, as indicated. See experimental conditions in Table 1. mQ was used as water. The grey circles correspond to standard deviations retrieved from at least ×3 experiments for each data point; see Table S2

Full size image

Comment on stability

It is worth pointing that despite being surfactant-free, the Au NPs are stable to centrifugation and sonication; see Fig. S5. We also previously stressed that despite the rather higher concentration of 0.5 mM of HAuCl4 precursor used (i.e. a rather high concentration for colloidal syntheses of Au NPs [26]), the as-prepared NPs are stable for long periods of time covering months or even years [5]. Figure S2 reports the UV-vis spectra and STEM micrographs of various colloidal dispersions measured 7 months after synthesis, for Au NPs obtained using different alcohols as the source of reducing agent and with a focus on those obtained using NaOH. NaOH is preferred to develop greener syntheses using a base cheaper and less toxic than LiOH. The results detailed in SI show the stability of the NPs that remain ca. 10 nm in size (closer to 5 nm for glycerol mediated synthesis) and do not agglomerate over time.

Comment on scalability

Although it is not the goal of this work to demonstrate the scalability of this approach, achieved elsewhere in 1 L scale [5], the findings suggest that lower purity alcohol (e.g. E70%) could be used to reduce the cost related to production and yet obtained small size and stable Au NPs as illustrated in Fig. S6. Although a 20–30 v.% ethanol content in water is less hazardous than 100% ethanol, there will certainly be some risk associated with handling alcohol-water mixture at large scale. Nevertheless, the promising combination of the approach presented using low viscosity media with flow systems could allow high throughput and increased safety [27, 28].

Conclusion

To successfully perform the low temperature (ca. 30 °C) synthesis of surfactant-free Au NPs, the use of high resistivity water (> 18.2 MΩ•cm) is to be preferred when ethanol, EG or glycerol are used as source of reducing agents. The grade of the ethanol source does not influence much the resulting properties of the Au NPs (plasmonic properties and therefore size and stability). This is an important finding since the Au NPs are best obtained using ca. 20–30 v.% ethanol, and therefore a cheaper stock solution of ethanol at 70 v.% can be used. This finding has direct relevance to increase the safety, cost and therefore large-scale implementation of this synthesis method leading to surfactant-free yet stable ca. 10 nm colloidal Au NPs.