Introduction

Chemical and physical behaviors of metal nanostructures are different from bulk materials because of quantum size effect resulting in specific electronic structures and these properties are dependent of their shape and size [1, 2]. The types and ratios of stabilizing agent, reducing agent and other chemicals determine these tunable properties in chemical synthesis of nanostructures. Among metal nanoparticles, gold nanoparticles have special interest and wide range of applications including biological applications like sensing, delivering and labeling [3–5]. Due to tunability of localized plasmon resonances of gold nanorods (AuNR), the synthesis of monodisperse novel AuNR is an appealing field of research [6–8]. Several methods were reported to synthesize AuNR with controllable size and aspect ratio utilizing seed – mediated growth [9]. Especially stabilizing agents that surrounds gold nanoparticles are essential in chemical and biological applications [4, 10]. One of the most used stabilizing agent in gold nanoparticle synthesis is cetyltrimethyl ammoniumbromide (CTAB) which yields stable gold nanoparticles [11]. In general, use of an acidic and weak reducing agent on a spherical gold nanoparticle (seed) in the presence of CTAB and silver ions leads AuNR formation with different aspect ratio. It is well documented that CTAB is a shape inducing agent and provides colloidal stability via shielding to prevent aggregation [12]. The exact mechanism involved in the anisotropic growth of gold nanorods in the seeded growth method is still lacking at a molecular level. However, silver plays important role during growth by formation of AgBr (bromide comes from CTAB) and getting deposited on 110 facet of gold surface. This promotes directed growth of gold nanorod from a facet which is less densely covered [13, 14]. Surfactant has also contribution to this unequal deposition on different facets. In this surfactant contribution, known as a zipping mechanism, the van der Waals interactions between hydrophobic tails within the CTAB bilayer would favor the longitudinal growth of the nanorod [6, 10].

However, cytotoxicity of CTAB capped gold nanoparticles is a concern in biological applications because CTAB has acute toxicity in most human cells [15, 16]. In this study, we offer an alternative cationic surfactant in gold nanorod synthesis: benzalkonium chloride (BAC) which is cheaper and less toxic than CTAB. BAC has been frequently used as a preservative in drugs and disinfectant in various applications [17]. Especially in ophthalmic solutions, it is the most frequently used preservative because it efficiently fights microbial contamination in bottles and also permits antimicrobials and antihypertensives to enter the anterior chamber by breaking cell-cell junction in the corneal epithelium [18, 19]. In the AuNR synthesis, concentrated CTAB has ability to form elongated rod-like micellar structures [20] and BAC has also similar tendency to form rod-like micelles [21]. The reported zipping mechanism as proposed the longitudinal growth of AuNR would favor identical nanorod growth [22]. Like CTAB, BAC has quaternary ammonium group through which BAC might interact with gold surfaces. Long alkyl chain and a phenyl group connected to quaternary ammonium in BAC can help stabilization of colloids.

Surface Enhanced Raman Spectroscopy (SERS) has great potential in biological applications due to high sensitivity which is sufficient for detection of trace biomolecules in complex biological matrix or even single molecule detection [23]. Fabrication of novel gold nanostructures is also important in SERS since SERS amplification strongly depends on plasmons on the surface of gold nanoparticles [24]. Recent studies have explicitly demonstrated that gold nanoparticles possess different optical properties and localized surface plasmon resonance that is essential in SERS [25]. In particular, anisotropic nanoparticles, AuNR, have strong electromagnetic field enhancements due to sharp corners and SERS intensity of a molecule at these sites has substantially increased [26]. The contributions from both electromagnetic and chemical enhancement may promote stronger SERS activity of AuNR than that of spherical gold nanoparticles [27]. Furthermore, the probe molecules can be easily attached to the gold surface via the gold-thiolate bonding, [28] which is essentially identical to the strong covalent bonding. Novel gold particles as SERS substrates can be used to maximize signals for intended analytes and makes the ultra-sensitive analysis possible with a Raman spectrometer. However sensitivity is not only concern for the fabrication of a new nanomaterials in SERS if it is aimed to use them in a quantitative analysis, different issues from colloidal stability to reproducibility should be taken into account [23]. A major advance is demonstrated herein the growth of BAC capped gold nanorods (BAC-AuNR) was achieved using single – component surfactant system are essentially identical to CTAB capped gold nanorods (CTAB-AuNR). Replacement of CTAB by BAC surfactant during AuNR synthesis results in monodisperse AuNR formation with anisotropic optical properties. We also studied analytical performance of the synthesized particle in the quantitative analysis of potato virus X (PVX) by SERS. PVX is the type member of the genus potexvirus and it is one of the most useful model system studying different aspects of virus infections in plant virus research [29]. Surface modification efficiencies of CTAB-AuNR and BAC-AuNR by thiol replacement on the surfaces were also studied.

Material & method

Chemicals and instrumentation

N-Ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), N- Hydroxysuccinimide (NHS), ascorbic acid, NaOH, HClO4, FeSO4, FeCl3, AgNO3, HAuCl4, phosphate buffered saline (PBS) tablets, ethanol, hydroxylamine hydrochloride, CTAB, 5,5′-Dithiobis(2-nitrobenzoic acid) (DTNB) were ordered from Sigma Aldrich (Germany, http://www.sigmaaldrich.com). Benzalkonium chloride solution (50%) was obtained from Tekkim (Bursa, Turkey, http://www.tekkim.com.tr). PVX and PVX antibody were purchased from Loewe Biochemica GmbH (Sauerlach, Germany, http://www.loewe-info.com). Water was ultrapurified by Mili-Q, Milipore system (18 MΩ cm).

Raman spectra were measured with DeltaNu Examiner Raman Microscopy system (Deltanu Inc., Laramie, WY, USA, https://www.sciaps.com) with a 785-nm laser source, a motorized microscope stage sample holder, and a cooled charge-coupled device (at 0 °C) detector. Instrument parameters were as follows: 150-mW laser power, 30-s acquisition time, 2-μm laser spot size and 20 × objective. UV-Vis spectra were measured by Agilent Technologies Cary 60 spectrophotometer (Santa Clara, CA, USA, http://www.agilent.com). Transmission electron microscopy (TEM) measurements were conducted with a JEOL TEM instrument (Peabody, MA, USA, http://www.jeolusa.com). Gold and sulfur amounts were measured with a THERMO ELECTRON X7 inductively coupled plasma-mass spectrometer (ICP-MS) (Waltham, MA USA, http://www.thermofisher.com). IR spectra were measured with Perkin Elmer Spectrum 400 FTIR/FTNIR (Waltham, MA USA, http://www.perkinelmer.com)

Seed-growth synthesis of gold nanoparticles

The procedure for AuNR synthesis started with seed formation reducing 250 μL HAuCl4 (0.01 M) by freshly prepared ice cold 600 μL NaBH4 (0.01 M) in the presence of 7.5 mL CTAB in an ice bath. Then, 1.15 mL HAuCl4.3H2O (0.01 M) was added to 4.75 mL of BAC (10%) to get orange growth solution in another vial. Then, 60 μL 0.01 M AgNO3 and 250 μL 0.1 M ascorbic acid were added to the growth solution subsequently. The resulting solution was colorless. Finally, 100 μL seed solution was added on growth solution and gently shaken for 10 s. After three hours of waiting, blue colored solutions. Solutions were centrifuged for 10 min at 12,500 rpm (12,453 rcf) to remove excess BAC. After centrifugation, AuNR were suspended in deionized water or buffer depending upon the intended purposes.

Synthesis of iron oxide core gold Shell magnetic nanoparticle

Iron oxide core – gold shell magnetic nanoparticles (MGNP) was used for magnetic nano-extraction of PVX in SERS application of AuNR. For this aim, MGNP were prepared according to our previous procedure [30]. Briefly, iron oxide nanoparticles were formed by co-reduction of 5 mL FeCl3 (1.28 M) and FeSO4 (0.64 M) after dropwise addition of 1.0 M NaOH (125 mL). Iron oxide particles were collected by a magnet and washed with deionized water. Then they were kept in 2 M HClO4 for 24 h and they were centrifuged at 10,000 rpm (8609 rcf) for 10 min. Iron oxide particles were dried at room temperature and 10 mg of them was dissolved in 5 mL water. Then 0.5 g ethylenediaminetetraacetic acid in 5 mL NaOH was added. After centrifugation at 8000 rpm (6887 rcf), 7 ml CTAB (0.1 M), 3 mL HAuCl4 (0.01 M), 0.35 mL NaOH (1 M), and 0.15 g hydroxylamine hydrochloride were added to iron oxide particles respectively. Wine-red color formation indicated surface covering of iron oxide particle with gold shell.

Characterization of nanoparticles

Nanoparticles were characterized according to their colors, UV-Vis absorption spectra and transmission electron microscope (TEM) images. UV-Vis absorption measurements of gold nanoparticles were measured between 400 and 800 nm. Gold nanoparticle samples were prepared on carbon/formvar coated copper grids and they were recorded on a TEM operated at 80 kV. Sulfur and gold amount in the nanoparticles after modification with some mercapto-compounds were measured with an ICP-MS. Silver content of the nanorod was also measured with the ICP-MS. Zeta potentials were measured for testing stability of AuNR. IR spectra were measured to observe BAC on the surface of AuNR.

Surface modification with mercapto-compounds

Surfaces of BAC-AuNR and CTAB-AuNR were modified with different mercapto-compounds; mercaptoundecanoic acid, lipoic acid and mercaptophenyl boronic acid: 1 mL BAC-AuNR and 1 mL CTAB-AuNR were centrifuged at 10000 rpm (14,000 rcf) for 10 min and washed with distilled water for three times separately. After three times washing, they were suspended in ethanolic solution of 20 mM 1 mL mercapto compounds separately and they were incubated for one night on a shaker. Then they were centrifuged and washed for three times with ethanol. After suspending in 5 mL water, finally Au and S amounts were measured with an ICP-MS.

SERS applications

Signal amplification of BAC-AuNR was firstly tested by using a Raman tag, DTNB. Signal enhancements by BAC-AuNR and CTAB-AuNR on DTNB were also compared. For this aim, same concentrations of AuNR were washed with water for three times after centrifugation at 1000 rpm for 10 min. Then they were suspended in ethanol containing 100 mM DTNB. After 1 h incubation at room temperature, AuNR were washed with ethanol for three times for removing excess DTNB and they were resuspended in ethanol. SERS spectra were obtained after evaporation of 3 μL samples on glass surface at room temperature. After signal enhancement study, BAC-AuNR was used for determination of PVX in a sandwich assay with MGNP. BAC-AuNR was modified with DTNB and PVX antibody for Raman labeling, while MGNP was modified with PVX-antibody for extraction.

BAC-AuNR solution was washed with water for three times after centrifugations at 1000 rpm for 10 min. They were suspended in ethanol and incubated for 1 h with a Raman label, DTNB (100 mM). After washing excess DTNB, carboxylate groups of DTNB were activated with EDC/NHS (0.1 M) for modification with PVX-antibody. Then they were incubated at 4 °C for one night with PVX-antibody. Excess PVX-antibody was washed with PBS for three times.

For magnetic nanoextraction, the surface of 1 mL MGNP was modified with a linker; 20 mM 0.1 mL mercaptoundecaonic acid (MUA) in ethanol. The mixture was incubated for 30 min and excess MUA was removed by washing with ethanol. Then, magnetic particles were re-suspended in 1 mL PBS, and carboxylate group of MUA on the surface of MGNP was activated using 0.2 mL EDC/NHS (0.1 M) for 30 min for antibody conjugation. After activation, they were incubated with 0.1 mL 100 μg/mL PVX-antibody at 4 °C for one night. Finally, they were washed with PBS for three times to remove free antibodies and they were re-suspended in 1 mL PBS.

PVX samples were prepared in PBS in different concentrations between 10 and 750 ppb. 50 μL of PVX-antibody modified MGNP were added to 100 μL PVX samples and the solutions were incubated for one hour at 4 °C. After three times washing with PBS, the sandwich assay procedure with DTNB labeled BAC-AuNR was applied: Firstly 100 μL 1% ethanolamine (v/v) were added to PVX captured MGNP solutions and incubated for three hours to prevent non-specific interaction. Then 50 μL DTNB labeled and PVX antibody modified AuNR were added to the solutions. After one night incubation at 4 °C, particles were collected with a magnet and washed for three times. Finally, they were suspended in 100 μL PBS and 3 μL samples were dropped on aluminum foil covered glass slides for Raman measurement. Three replicates of PVX at 100, 300 and 500 ppb were prepared in PBS and analyzed according to above procedure for three days for analytical validation studies.

Results & discussion

It is well documented that the seed mediated growth procedure is the most practical method for the synthesis of AuNR. Generally, CTAB and gemini surfactants were used as shape – directing agents in seed mediated growth procedure. The resulting aspect ratio can be controlled by changing the concentration of HAuCl4 and seed. In the present study, we describe the use of BAC as the directing agent for seed mediated growth of AuNR. BAC growth medium has some advantages over traditional CTAB based growth systems like BAC is less toxic and cheaper than CTAB. Although CTAB was used in seed synthesis and at later steps it was removed from particles so BAC is the primary surfactant for seeded growth of gold nanorods. The FTIR spectrum of BAC-AuNR shows that BAC exist on the surface of AuNR and CTAB which was used in seed synthesis was removed in seed mediated growth and washing procedures (Fig. S1–3). Although BAC and CTAB have similar molecular groups in their structures, it is possible to differentiate them in IR spectrum especially considering aromatic group of BAC and some other small differences in their spectra. Therefore, we employed FTIR spectroscopy to confirm only BAC exist on the surface of AuNR. Bands of BAC-AuNR mostly match with BAC while it differs more with CTAB. First and the most important IR absorption is at 1628 cm−1 assigned to partial double bonds of aromatic group in BAC and BAC-AuNR have also weaker form of that band. Also IR absorption at 3002 cm−1 is seen in BAC and BAC-AuNR spectra that is well known band for phenyl group which does not exist in CTAB FTIR spectrum. Asymmetric vibrations of CH3–N+ in BAC exactly match with BAC-AuNR 1457, 1467, 1476 cm−1 (νasym (CH3–N+)) while CTAB shows small differences having those peaks at 1462, 1473, 1486 cm−1. The peaks at 1430, 1407 and 1396 cm−1 of CTAB are assigned to symmetric CH3–N+ vibrations and those vibrations can be seen at 1447 and 1378 for BAC and BAC-AuNR. Stretching of hydrocarbon chain (sp3 C-H) around 2900 cm−1 also shows small differences between CTAB and BAC. This region of BAC-AuNR mostly matches with BAC while it is a little more different from CTAB. Finally, CTAB has strong bands at 912 and 961 cm−1 that do not exist in BAC-AuNR spectrum. While most of the bands of BAC can be clearly seen in BAC-AuNR, some bands were weaker in BAC-AuNR FTIR spectrum and it might be due to AuNR limits some vibrational and rotational modes of molecules attached on its surface. On the other hand, bands of CTAB do not coincide with BAC-AuNR bands. Consequently, FTIR spectrum of BAC-AuNR displays only BAC exist on the surface of AuNR. Therefore, as a surface protecting agent, BAC determines some characteristics of AuNR such as aggregation tendency, toxicity, etc.

Our approach to the formation of AuNR is to deposit Au+ ions on the surface of gold seed in a fashion similar to seeded growth used in CTAB as a surface protecting agent. In this study, the surfactant solution used is BAC and the addition of gold chloride salt solution to a mixture of silver and BAC leads to the formation of a complex involving both benzalkonium ions (BA+) and Au3+ which appeared as a pale – yellow. In consequence of reduction of Au3+ to Au+ by ascorbic acid, Au+ ions were deposited on to the small gold seeds in the presence of directing surfactant BAC and silver ions. The resulting aspect ratio can be controlled by changing the concentration of HAuCl4, seed, BAC, silver and reducing agent.

Optimizations

The effect of BAC, HAuCl4, seed, AgNO3 and ascorbic acid concentration on gold nanoparticle size and shape were studied. In the first set of experiments, we investigated the effect of BAC concentration on aspect ratio. We observed remarkable changes in size and shape when the amount of BAC changed between 1 and 20% (Fig. S4). The change of shape in gold nanoparticles depending on BAC concentration could be originated from the fact that BAC might have different binding affinities to facets of gold nanoparticles. This may change the order of free energies for different planes and it can cause different growth rate in planes [31]. As a result, different shapes can be observed rather than spherical particles. We think that the increase of the amount of BAC can affect the difference of growth rate in planes resulting different particle shapes from spheres to rods. When BAC concentration increased from 1% to 6%, size of anistropic spheres and cubes increased from 20 nm to 40 nm (Table 1). After 8% BAC concentration, we observed rod formation. Anisotropic triangles, isocahedra and cubes were observed at 20% BAC concentration. Among these concentrations, 10% BAC concentration was selected for further studies due to higher yield, homogeneity, and repeatability.

Table 1 Benzalkonium chloride optimization (100 μL seed, 1000 μL HAuCl4 (1.86 × 10−3 M), 60 μL AgNO3 (0.01 M), 250 μL ascorbic acid (0.1 M))
Full size table

We performed another set of experiment for gold concentration optimization keeping other parameters unaltered. The effect of gold concentration was investigated between 5.67 × 10−4 M – 2.20 × 10−3 M (Table 2). As expected, low gold concentrations yield spherical particles while rod-shaped particles were observed at higher concentrations of gold (Fig. S5). UV-Vis absorption spectra (Fig. S9) confirms TEM images by formation of a new band around 700 nm that belongs to longitudinal band on rod. The transverse band is around 530 nm. TEM image and a small shoulder of the transverse peak show that small amount of spherical particles still exists with rod shaped particles. Gold concentration was hold at 1.85 × 10−3 M for further optimization experiments.

Table 2 Gold concentration optimization (100 μL seed, 10% benzalkonium chloride, 60 μL AgNO3, 250 μL ascorbic acid)
Full size table

Seed concentration also affected particle shape and sizes (Table 3). Seed volumes from 50 μL to 600 μL creates spherical, rod and mixture of these particles (Fig. S6). According to TEM images and UV-Vis absorption spectra (Fig. S10), we decided to use 100 μL seed for further experiments.

Table 3 Seed volume optimization (1.85 × 10−3 M Au, % 10 benzalkonium chloride, 60 μL Ag, 250 μL ascorbic acid)
Full size table

AgNO3 amount, which was used to enhance yields in gold nanorod synthesis, was optimized between 15 and 100 μL (Table 4). After 30 μL of AgNO3, rod formation was observed in TEM images (Fig. S7) with formation of a new UV-Vis absorption band (Fig. S11). 60 μL of AgNO3 which yields more homogenous particle profile was selected as optimum amount.

Table 4 Silver nitrate (0.01 M) volume optimization (1.85 × 10−3 M Au, % 10 benzalkonium chloride, 100 μL seed, 250 μL ascorbic acid
Full size table

Finally, we studied the effect of reducing agent, ascorbic acid in a range of 100–250 μL (Table 5). When 100 μL ascorbic acid was used, anisotropic spheres about 20 nm were observed. After 125 μL ascorbic acid, we observed rod formations with aspect ratio from 1.5 to 3 observed (Fig. S8 and Fig. S12). The final AuNR solution has blue color and zeta potential of 27.8 mV.

Table 5 Ascorbic acid (0.1 M) volume optimization (1.85 × 10−3 M Au, % 10 benzalkonium chloride, 100 μL seed, 60 μL Ag)
Full size table

Optimized AuNR contain a little silver in it and silver content was found as 1.4 μg·mL−1 by ICP-MS measurements.

Modification with mercapto-compounds

After the synthesis, the surface of AuNR should be modified for intended application. For this purpose, the capping agent has to be replaced effectively with a new agent that has higher affinity to gold surfaces. We compared surface modification of CTAB-AuNR and BAC-AuNR with frequently used surface modifying agents, mercaptoundecanoic acid, lipoic acid and mercaptophenylboronic acid. According to S/Au ratio, BAC-AuNR has higher modification yield in mercaptoundecanoic acid modification while CTAB-AuNR show higher yields in lipoic acid and mercaptophenyl boronic acid modification according to ICP-MS measurements (Table S1). These results indicate that BAC-AuNR can be easily modified like CTAB-AuNR.

SERS applications of BAC-AuNR

Amplification of the Raman signals by metal nanostructures depends on hot spots created on sites of metal particles. These intensified electric fields that is responsible for surface enhanced light scattering effect depends on fabrication techniques of metal particles [26, 32]. Metal particles with different synthesis routes can have different signal enhancement effects in spectroscopy. We compared signal amplification of BAC-AuNR and CTAB-AuNR on a Raman dye, DTNB. SERS spectra (Fig. 1) show that BAC-AuNR has higher enhancement effect on DTNB than that of CTAB-AuNR. The calculated enhancement factor (EF) for BAC-AuNR is ~1.7 × 104 (see Electronic Supplementary Material for EF calculation). The results show that BAC-AuNR can be used for SERS analysis of samples which requires high sensitivity.

Fig. 1
figure 1

Enhancements of DTNB Raman spectrum by CTAB-AuNR and BAC-AuNR

Full size image

We also tested analytical performance of BAC-AuNR in a quantitative SERS analysis of a biologically important analyte, a virus. Fabrication of homogenous and stable nanostructures affects analytical parameters in quantitative SERS analysis because aggregated or heterogeneous particles have low repeatability and reproducibility in SERS. We tested the analytical performance of BAC-AuNR (Fig. 2) on PVX determination by sandwich assay with using raman labeled BAC-AuNR and MGNP. The strong peak at 1332 cm−1 in SERS spectra (Fig. 3a) were used in the quantitative analysis. Range was determined by observation of linear responses between PVX concentration and peak intensities. Linear relationship in the range of 10–750 ppb of PVX has high coefficient of determination in the calibration graph (Fig. 3b). Detection and quantitation limits were calculated as 2.2 ppb and 6.5 ppb respectively from slope and intercept values of calibration sets. Samples at three different concentration levels were analyzed to test accuracy and precision. High recoveries between 97.7–103.3% shows that the assay has high accuracy. Intra-day and inter-day precision were assessed by relative standard deviations of samples (Table S2). Validation results prove that BAC-AuNR can produce precise and accurate signals in SERS analysis.

Fig. 2
figure 2

TEM image of BAC-AuNR used in PVX determination

Full size image
Fig. 3
figure 3

SERS study of PVX determination based on Raman label (DTNB) measurements a calibration spectra b calibration graph

Full size image

Table 6 shows comparison of BAC-AuNR with some other SERS active particles in terms of application and enhancement factors. It is clear that the developed BAC-AuNR is not one of the best material in terms of enhancement factor but still it provides high-sensitivity comparing other detection techniques used in PVX detection such as UV absorbance in ELISA [33] since SERS itself is a high sensitive technique that even single molecule detection is possible with 107–108 enhancement factors. The enhancement factors of silver aggregates can reach to level of 1010 [34] however precision is generally low if they are used in quantitative analysis due to non-uniform structure of aggregates. In consequence, BAC-AuNR provides low toxicity, acceptable accuracy and precision for quantitative analysis and moderate sensitivity comparing other SERS active particles.

Table 6 Recent examples of SERS active nanomaterials
Full size table

Conclusion

In this study, we developed a SERS active novel AuNR by using BAC as a stabilizing agent. BAC was used as a shape – directing agent in seed mediated growth procedure. The effects of chemicals-reducing agent, stabilizing agent, gold, and silver-were observed for homogenous and stable AuNR fabrication. TEM images indicated that the monodisperse AuNR were synthesized with a high yield. The high–yield anisotropic growth and monodispersity of AuNR can be attributed to the use of BAC as a surfactant. Synthesized BAC-AuNR not only shows high signal amplification in Raman spectroscopy but also serves high analytical performance for quantitative analysis in SERS. High sensitivity was achieved in SERS measurements of DTNB. We also showed that the surface of BAC-AuNR can easily be modified with thiol containing compounds for different applications. Finally, we used BAC-AuNR in a quantitative SERS application for PVX determination. BAC can be a cheaper and safer alternative to CTAB in AuNR synthesis. This study concentrates on introduction of a new and less toxic surface protection agent in AuNR synthesis and exploring analytical performance of the AuNR in quantitative analysis. Further studies such as facet control study might be necessary to if higher enhancement factors desired.