1.

Introduction

Human malaria disease is a worldwide disease with estimated 225 million cases, accounting for 800,000 deaths per year.¹ The disease is caused by a parasite protozoan, in which the parasite infects blood cells of the host and hemozoin biocrystals are disposed as byproducts after the ingestion of hemoglobins.² Since the malaria disease can aggravate into a fatal illness within hours upon development of the first symptom,³ the early diagnosis of malaria infection is important, which requires the detection of hemozoin at low concentrations in infected blood cells.^1,2 In malaria diagnosis, microscopic examination of blood smears remains the “gold standard” for the detection of malaria parasites, but this method is labor-intensive and time-consuming; moreover, special expertise from operators is required for reliable data interpretation.⁴ Therefore, the development of a sensitive technique, which requires minimal labor and expertise for hemozoin detection, is warranted in early malaria diagnosis.

Recently, several other malaria diagnosis techniques, such as the quantitative buffy coat method, the molecular diagnostic method, flow cytometry technique, serological tests, light scattering measurement, optical tweezing and laser desorption mass spectrometry,^4–8 have been developed to overcome the shortcomings of the traditional method. Among these methods, resonance Raman spectroscopy (RRS) has been reported to amplify the Raman signal of hemozoin in malaria parasite-infected blood cells by the close Raman shift matching between the laser source and electronic transition of hemozoin.^6,9 Moreover, surface-enhanced Raman scattering (SERS) effect has also been shown on a silver tip to enhance the Raman spectrum of hemozoin in infected cells via the augmented electromagnetic coupling between hematin and gold or silver nanoparticles.¹⁰ The success in detecting the RRS and SERS signal of hemozoin shows the potential for further augmentation by combining the two effects, known as surface enhanced resonance Raman spectroscopy (SERRS). This SERRS technique has been demonstrated on other test molecules [rhodamine 6G (R6G)] adsorbed on silver nanoparticles for further enhancement in the Raman signal, as silver gives a higher enhancement than other metals (e.g. gold).¹¹ The SERRS signal could be further enhanced by employing a magnetic field to enrich the hemozoin concentration. A similar strategy has been employed in other applications such as magnetic purification.^12,13 Moreover, magnetic nanoparticles, e.g. nanoparticles made of iron oxide that show to be effective magnets at room temperature,¹⁴ could also attract and attach to hemozoin in a magnetic field because hemozoin is paramagnetic in nature,¹² in a magnetic field. This approach is similar to the strategy reported for capturing and enriching bacteria.¹⁵

In this work, we report a novel magnetic field enrichment strategy on SERRS by using magnetic nanoparticles to augment Raman signals from $\beta$-hematin crystals, similar to hemozoin¹⁶ in molecular, magnetic and Raman properties. The SERRS effect is further enhanced by the magnetic enrichment of $\beta$-hematin crystals and magnetic SERS-active nanoparticles with iron oxide core and silver shell. The performance of magnetic field-enriched SERRS quantified experimentally is compared with that of the SERRS without the influence of magnetic field, and the ordinary RRS on $\beta$-hematin crystals. Furthermore, the analytical enhancement factor and sensitivity of the proposed magnetic field-enriched SERRS technique are investigated.

2.

Materials and Methods

3.

Results

Figure 1(a) gives the FESEM image of the raw ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles. Individual ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles have a mean diameter of about 50 nm ($\pm 5\mathrm{nm}$). Figure 1(b) shows the FESEM images of the ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles coated with silver shells. The ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles were well dispersed in the image. Each ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticle has a mean diameter of about 140 nm [Fig. 1(c)], with a size range $\pm 20\mathrm{nm}$ characterized by zetasizer measurements. The core-shell geometry is reconfirmed by the TEM image [Fig. 1(d)] with an EDX spectrum[Fig. 1(e)] that reveals the elemental composition of the nanoparticle. Fe, Ag, O, Cu and C can be observed in the EDX graph. Fe, O and Ag signals are originated from the ${\mathrm{Fe}}_3{\mathrm{O}}_4$ core and Ag shell, while Cu, and C are attributed to the copper grid.

Fig. 1

(a) FESEM image of raw ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles. (b) FESEM and (c) zoomed in FESEM image of Ag nanoparticles. (d) Representative TEM image of ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles (Other structures in the image are attributed to surfactant). (e) EDX of the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles.

Figure 2 compares the SERS spectra of aqueous R6G solution adsorbed on the fabricated ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles [Fig. 2(a), concentrations varying from ${10}^{-6}\mathrm{M}$ to ${10}^{-8}\mathrm{M}$] with the ordinary Raman spectra of R6G solution [Fig. 2(b), concentrations varying from ${10}^{-2}\mathrm{M}$ to ${10}^{-3}\mathrm{M}$]. Most prominent Raman peaks, such as $\mathrm{C}─\mathrm{C}─\mathrm{C}$ ring in-plane bending ($615{\mathrm{cm}}^{-1}$), CH out-of-plane bending ($775{\mathrm{cm}}^{-1}$), $\mathrm{C}─\mathrm{O}─\mathrm{C}$ stretching ($1185{\mathrm{cm}}^{-1}$), $\mathrm{C}─\mathrm{C}/\mathrm{C}─\mathrm{N}$ stretching ($1310{\mathrm{cm}}^{-1}$ and $1365{\mathrm{cm}}^{-1}$), and aromatic $\mathrm{C}─\mathrm{C}$ stretching ($1508{\mathrm{cm}}^{-1}$),²⁶ can be observed in the SERS spectra of R6G. The minimum detectable concentration of R6G absorbed on the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles is $1\times {10}^{-8}\mathrm{M}$, which is five orders of magnitude more sensitive than that of ${10}^{-3}\mathrm{M}$ detected in the ordinary Raman spectrum without enhancement. We estimated²⁷ that the analytical enhancement factor (AEF) of the SERS signals relative to the ordinary Raman measurement (${AEF}_{\mathrm{SERS}/\text{Raman},\mathrm{R}6\mathrm{G}}$) is about $5.77\times {10}^6$, which is comparable to the AEF values (around ${10}^3$ to ${10}^6$) of nanoparticle colloids stated in the literature.^27,28 These results suggested the feasibility of using the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles for enhancing the Raman signal of $\beta$-hematin crystals.

Fig. 2

SERS spectra of R6G solution at concentrations of ${10}^{-6}\mathrm{M}$, ${10}^{-7}\mathrm{M}$, and ${10}^{-8}\mathrm{M}$, with ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles, at laser excitation power of 0.1 mW. (b) Ordinary Raman spectra of R6G at concentrations of $1\times {10}^{-2}\mathrm{M}$, $5\times {10}^{-3}\mathrm{M}$, and $1\times {10}^{-3}\mathrm{M}$, at laser excitation power of 10 mW. The acronyms in the legends mean the following. ip: in-plane; op: out-off-plane; str: stretching; Arom: Aromatic.

Figure 3 shows the FESEM image of $\beta$-hematin crystals fabricated using the acid-catalyzed method [Fig. 3(a) and 3(b)] and the size distribution of crystals [Fig. 3(c)]. The fabricated $\beta$-hematin crystals are comparable to the size of hemozoin biocrystals found in the ring stage parasites (estimated from the concentration per cell and density of hemozoin)^29,30 that dominate over other stages¹² in the bloodstream for detection. Close resemblance in the spatial dimensions between the two types of crystals presumably would result in similar SERRS enhancement effect. The minimization of fabricated hemozoin size is to avoid artificially higher enhancement in magnetic field enrichment since larger crystals will have higher magnetic field enrichment as shown in the result later. Hence, the acid-catalyzed method is preferred to fabricate smaller crystals over the anhydrously synthesized²⁰ $\beta$-hematin or in the biochemically cloned³¹ hemozoin, although the resulted crystals may have lower crystallinity and smaller sizes as reported in Ref. 19.

Fig. 3

(a) FESEM image of $\beta$-hematin crystals. (b) Zoomed in FESEM image of $\beta$-hematin crystals. (c) Population distribution of size (area) $\beta$-hematin crystals obtained by using Matlab software.

Figure 4 compares resonance Raman spectra of hematin with [Fig. 4(a)–4(c)] and without [Fig. 4(d)–4(f)] magnetic field-enriched strategy at concentrations ranging from ${10}^{-3}\mathrm{M}$ to $5\times {10}^{-9}\mathrm{M}$. Prominent vibrational features, such as ${\nu }_8$ (based on the electron spin and crystallographic coordination notation tetragonal $D_{4h}$ system for resonance Raman peaks studies on myoglobin)³² at $345{\mathrm{cm}}^{-1}$, ${\gamma }_6$ at $367{\mathrm{cm}}^{-1}$, ${\nu }_{15}$ at $754{\mathrm{cm}}^{-1}$, ${\nu }_{22}$ at $1120{\mathrm{cm}}^{-1}$, ${\nu }_{11}$ at $1551{\mathrm{cm}}^{-1}$, ${\nu }_2$ at $1570{\mathrm{cm}}^{-1}$, and ${\nu }_{10}$ at $1628{\mathrm{cm}}^{-1}$,³³ are noted in most of these spectra. The locations of these peaks are equal to those reported Raman peaks⁶ for hemozoin biocrystals, confirming that the spectral features of $\beta$-hematin crystals are equivalent to hemozoin in Raman spectroscopy. The effect of surface enhancement can be clearly distinguished when the SERRS [Fig. 4(a) and 4(d)] are compared with the RRS measurements [Fig. 4(b) and 4(e)] and the RRS measurements by using ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles [Fig. 4(c) and 4(f)]. We also note that the lowest detectable concentrations of $\beta$-hematin for SERRS with ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles, RRS, and RRS with ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles under magnetic field enrichment are $5\times {10}^{-9}\mathrm{M}$, $5\times {10}^{-7}\mathrm{M}$, and $5\times {10}^{-7}\mathrm{M}$, respectively, and those without magnetic field enrichment are $5\times {10}^{-6}\mathrm{M}$, $5\times {10}^{-4}\mathrm{M}$, and $5\times {10}^{-4}\mathrm{M}$, respectively, where the excitation power in the SERRS measurements was 0.1 mW and that in the RRS measurements was 10 mW.

Fig. 4

Magnetic field-enriched spectra of $\beta$-hematin crystals in a magnetic field (a) with ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles (Concentrations: $5\times {10}^{-6}\mathrm{M}$ to $5\times {10}^{-9}\mathrm{M}$; $P_{\text{Ex}}$: 0.1 mW), (b) without and (c) with ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles (Concentrations: $5\times {10}^{-6}\mathrm{M}$, and $5\times {10}^{-7}\mathrm{M}$; $P_{\text{Ex}}$: 10 mW). Spectra of $\beta$-hematin crystals without magnetic field (d) with ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles (Concentrations: $5\times {10}^{-5}\mathrm{M}$ to $5\times {10}^{-6}\mathrm{M}$; $P_{\text{Ex}}$: 0.1 mW), (e) without and (f) with ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles (Concentrations: $5\times {10}^{-3}\mathrm{M}$, and $5\times {10}^{-4}\mathrm{M}$; $P_{\text{Ex}}$: 10 mW). $P_{\text{Ex}}$ means the excitation power.

4.

Discussion

We have demonstrated the feasibility and significant improvement of magnetic field-enriched SERRS over conventional SERRS for detecting $\beta$-hematin crystals at low concentrations. To gain additional insight into Raman enhancement in this technique, we calculated the analytical enhancement factor (AEF) in each of the following techniques relative to the RRS measurement of $\beta$-hematin crystals without ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles [Fig. 4(e)]: 1. magnetic field-enriched SERRS (${AEF}_{\text{mag}\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$), 2. SERRS without magnetic field (${AEF}_{\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$), and 3. magnetic field-enriched RRS (${AEF}_{\text{mag}\mathrm{RRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$). These AEFs have been calculated by applying Eq. (1):

The enhancement mechanism behind the addition of ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles is studied. SERRS signal of $\beta$-hematin [Fig. 4(d)] is only exhibited by mixing nanoparticles with the SERS-active silver shell and $\beta$-hematin. RRS is resulted for ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles and $\beta$-hematin mixture [Fig. 4(f)], with RRS signal comparable to that of $\beta$-hematin without any nanoparticles [Fig. 4(e)]. This observation also applies to the magnetic field-enriched measurement, with only SERRS enhancement noted in the mixture of ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles and $\beta$-hematin [Fig. 4(a)], while RRS is exhibited in the $\beta$-hematin mixture with [Fig. 4(b)] and without [Fig. 4(c)] ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles under a magnetic field. RRS intensities of $\beta$-hematin with ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles under magnetic field is 1.4 times higher than that without ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles, probably attributed to the enhanced aggregation of $\beta$-hematin due to the ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles. The spectral shapes are similar in the SERRS and RRS spectra despite the fact that higher intensity is noted in the SERRS spectra. The similarity may be explained by the unchanged chemical structure and symmetry of $\beta$-hematin crystals that are magnetically held to the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles. Compared to many other molecules such as R6G, the adsorption of $\beta$-hematin crystals onto the Ag surface is weaker thus its SERRS spectrum is less influenced by the adsorption. The similar phenomenon is also observed in other chemicals³⁴ that have weak interactions with Ag.

We have compared ${AEF}_{\mathrm{SERRS}/\mathrm{RRS}}$ for $\beta$-hematin with ${AEF}_{\mathrm{SERS}/\text{Raman}}$ for R6G, since the two quantities are considered equivalent.³⁵ The enhancement factor of SERS relative to ordinary Raman for R6G (${AEF}_{\mathrm{SERS}/\text{Raman},\mathrm{R}6\mathrm{G}}\approx 5.77\times {10}^6$) is higher than that of the SERRS relative to RRS for $\beta$-hematin (${AEF}_{\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}\approx 1.54\times {10}^3$), which can be attributed to the larger size of the $\beta$-hematin compared R6G molecules, with size (area) at least greater than roughly $30\mathrm{nm}\times 30\mathrm{nm}$ (Fig. 3). Nevertheless, aggregates formed between magnetic $\beta$-hematin and the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles can still lead to effective SERS activities, similar to the configurations reported in the literature (e.g. localized AEF of about ${10}^9$ in configuration such as dimers and trimers).³⁶ In addition, SERS can be observed in $\beta$-hematin at a distance from the Ag surface in the aggregation configuration ($<40\mathrm{nm}$), similar to that in other SERS nanoparticles.³⁷ The ${AEF}_{\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ is compensated by the further augmentation in the magnetic field-enriched SERRS and RRS measurements, which can be explained by the following effects induced by the magnetic field. First, $\beta$-hematin is enriched. Second, more ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles are attached to each $\beta$-hematin crystal and thus, leading to higher SERRS intensity. The potential mechanisms responsible for these effects are elaborated as follows.

The enrichment of $\beta$-hematin concentrations due to a magnetic field can be interpreted by the fact that paramagnetic¹⁶ $\beta$-hematin are attracted much faster to the bottom of the vial by the magnet than unmagnetized crystals without the influence of a magnetic field. Consequently, the concentration of $\beta$-hematin will be higher at the laser spot than that without the magnetic field, which has been demonstrated by a value of 68 in ${AEF}_{\text{mag}\mathrm{RRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ for RRS measurements under the influence of a magnetic field in the absence of nanoparticles [Fig. 4(c)]. The magnetic field enrichment effect is more significant in larger $\beta$-hematin crystals, as confirmed by our Raman experiment in the $\beta$-hematin supernatant after centrifuging (Fig. 5). Higher ${AEF}_{\text{mag}\mathrm{RRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ is observed in the magnetic field-enriched RRS for $\beta$-hematin mixture without centrifuging (68 in Fig. 4) than that calculated²⁷ in the $\beta$-hematin supernatant (4 in Fig. 5). Hence, the magnetic field can effectively enrich $\beta$-hematin crystals to give rise to strong enhancement.

Fig. 5

(a) Magnetic field-enriched SERRS and (b) SERRS spectra of $\beta$-hematin supernatant obtained by centrifuging ($\beta$-hematin at concentration of ${10}^{-4}\mathrm{M}$) at an excitation power of 0.1 mW. (c) Magnetic field-enriched RRS and (d) RRS spectra of the same $\beta$-hematin supernatant at an excitation power of 10 mW.

More nanoparticle-hematin aggregates are formed by an external magnetic field. Figure 6 shows that more ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles are bound to $\beta$-hematin crystals in a magnetic field [Fig. 6(a)] compared to the case without a magnetic field [Fig. 6(b) and 6(c)]. Since the ${\mathrm{Fe}}_3{\mathrm{O}}_4$ core of the nanoparticles is ferromagnetic,¹⁴ the magnetic field produced by each ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticle attracts adjacent $\beta$-hematin crystals [Fig. 6(b) and 6(c)]. More ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles are close or tightly bound to $\beta$-hematin crystals under an external magnetic field [Fig. 6(a)]. Figure 6(d) shows a schematic physical model for the configurations of ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles and $\beta$-hematin crystals that leads to surface enhancement. For each crystal attached at the contact or in close vicinity ($<40\mathrm{nm}$) in the gap to the ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticle, known as “hot spots,” intense SERS activities occur.^11,36,37Additional hot spots are formed with more aggregations with the application of an external magnetic field. The increased number of nanoparticles-hematin aggregates may be responsible for the further improvement in the ${AEF}_{\text{mag}\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ that is 149 times larger than the ${AEF}_{\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$, in comparison with that in RRS without the involvement of nanoparticles as in ${AEF}_{\text{mag}\mathrm{RRS}/\mathrm{RRS},\beta \text{-}\text{hema}}\approx 68$. In contrast, the ratio of $E{F}_{\text{mag}\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ to $E{F}_{\mathrm{SERRS}/\mathrm{RRS},\beta \text{-}\text{hema}}$ in $\beta$-hematin supernatant ($3540/1200\approx 3$) is similar in magnitude to the AEF in the magnetic RRS measurement (${AEF}_{\text{mag}\mathrm{RRS}/\mathrm{RRS},\beta \text{-}\text{hema}}\approx 4$),²⁷ since the $\beta$-hematin supernatant contained mostly small crystals that are already attached on the nanoparticles. Therefore, more nanoparticle-hematin aggregates can give further augmentation in Raman signals, in addition to the SERRS and $\beta$-hematin enrichment effects.

Fig. 6

FESEM images of ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles and $\beta$-hematin crystals (a) with and (b) without the external magnetic field.

With the two aforesaid magnetic field induced effects, we evaluate the detection limit of $\beta$-hematin using magnetic field-enriched SERRS by converting $\beta$-hematin concentration to the concentration of malaria parasites in blood for practical evaluation. The detection limit of $\beta$-hematin concentration at $5\times {10}^{-9}\mathrm{M}$ in the magnetic field-enriched SERRS measurement obtained in this study is equivalent to roughly $30\text{parasites}/\mathrm{\mu l}$ (considering a hemozoin concentration of about $0.22\mathrm{pg}/\text{cell}$ in the earlier malaria infection at the ring stage and a molecular weight of $1229\mathrm{g}/\text{mol}$ for hemozoin)^29,38 at the early stage. More importantly, the sensitivity is comparable to other rapid malaria detection techniques for hemozoin detection at later malaria stages, e.g. $10\text{parasites}/\mathrm{\mu l}$ (with hemozoin concentration at about $0.6\mathrm{pg}/\text{cell}$),¹² for laser desorption mass spectrometry and automated blood cell counters.^33,39,40 With the high sensitivity in the detection of $\beta$-hematin in the current configuration without optimization, the magnetic field-enhanced SERRS technique has demonstrated great potential for early malaria diagnosis. The detection sensitivity of our technique could be further improved by optimizing the configuration of the magnetic field and the physical geometry of SERS-active nanoparticles.

5.

Conclusions

We report the detection of $\beta$-hematin crystals using magnetic field-enriched SERRS enabled by ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles. The method enriches $\beta$-hematin crystals and ${\mathrm{Fe}}_3{\mathrm{O}}_4@\mathrm{Ag}$ nanoparticles by applying an external magnetic field and synergizes with the enhancement capability of SERRS, thereby promoting further augmentation in the Raman signal of $\beta$-hematin crystals. A parasitemia level of $30\text{parasites}/\mathrm{\mu l}$ in blood in the early stages of infection can be achieved by using this method with the current setup, which demonstrates the potential of employing magnetic field-enriched SERRS in early malaria diagnosis.

6.

Appendix