Smartphone based LSPR sensing platform for bio-conjugation detection and quantification

Abstract

We report here the working of a localized surface plasmon resonance (LSPR) sensor using the camera of a smartphone. Integrating light weight and simple laboratory optical components with the camera module of the phone, we first designed a visible spectrophotometer with a pixel resolution of 0.336 nm per pixel. The LSPR peak absorption wavelength shift due to size variation of gold nanoparticles (AuNPs) and analyte (protein and enzyme) conjugation with AuNPs have been successfully recorded by our smartphone spectrophotometer. The shift in LSPR peak absorption wavelength can be correlated with the size of the AuNPs and concentration of biomolecules attached to it. The limit of detection (LOD) of the designed sensor for quantification of BSA protein and trypsin enzyme was estimated to be 19.2 μg mL⁻¹ (equivalent to 0.28 μM) and 25.7 μg mL⁻¹ (equivalent to 1.10 μM) respectively. We compare the results with a laboratory grade standard UV-VIS spectrophotometer and observe high reliability of our designed sensor. Owing to its compact size, simple optics design and involvement of low-cost optical components we envision that the proposed sensing system could emerge as an alternative inexpensive handheld LSPR sensing tool that can be suitable for different in-field applications.

---

Introduction

Smartphones are considered to be a ubiquitous tool in the day-to-day lives of humans. Smartphones are equipped with Wi-fi, bluetooth, high mega pixel cameras, multimedia services along with a variety of inbuilt sensors such as gyroscopes, accelerometers, ambient light sensors (ALS), GPS etc. With the unprecedented growth in hardware and software realms, smartphones are being synonymised as portable computers having computational power similar to that of desktop based computers. With the growth of mobile software technology it is feasible to develop flexible mobile applications in accordance to the need of the user. Now-a-days a smartphone with 5 MP camera can be procured at an affordable price which subsequently can be developed as a microscope,^1,2 spectrometer^3–6 and different sensing tools.^7–12 Further, smartphones have been integrated with microfluidics for identification of pathogenic nucleic acids¹³ and red wine testing.¹⁴ Preechaburana et al.¹⁵ have developed a smartphone based lab-on-chip (LOC) device for chemical sensing. Label-free detection is one of the most promising technique in the field of bio-sensing research.¹⁶ Gallegos et al.¹⁷ have demonstrated photonic crystal integrated smartphone sensor for label-free detection of biomolecules. Amongst various types of label-free biosensing techniques, Surface Plasmon Resonance (SPR) and Localized Surface Plasmon Resonance (LSPR) have enormous credentials to their name for serving a wide variety of analytical and bio-analytical sensing investigations.^18–24 In spite of its promising advantages, both SPR and LSPR sensors suffer from certain limitations such as portability, cost-involvement, lack of field-deployable configuration and absence of data sharing capabilities from remote area to centralized laboratories. The instrumentation required for performing both types of sensing involves sophisticated, bulky and expensive devices such as SPR biosensors and spectrophotometers (for LSPR biosensing). These issues related to SPR sensors have been successfully mitigated by implementation of optical fibres in combination with smartphones as reported by Liu et al.²⁵ and Bremer et al.²⁶ Both groups have demonstrated similar line of work by integrating smartphones with optical fibres for possible sensing of biomolecular concentration (IgG antibody) and refractive index variation. Although SPR sensing exhibits better sensitivity¹⁸ (∼10⁶) than that of LSPR sensing techniques (∼10²) SPR sensing requires uniform metal film of nano-dimensional thickness (∼50–60 nm) which is not always feasible and requires sophisticated and expensive instruments for thin film deposition. Moreover, silver (Ag) films used as thin film are easily oxidized and hence brings instability in the detection process. On the other hand LSPR based sensing can be realized by simple laboratory spectrophotometer and does not requires thin film deposition. One of the most remarkable feature of LSPR based sensing is its marginal bulk effect as compared to SPR sensing. The electromagnetic field of surface plasmons extends up to 300–400 nm at the metal-dielectric interface whereas in LSPR the coupled electromagnetic field is localized within 20–40 nm from the NPs' surface. Due to relatively larger bulk effect of SPR, biomolecules that are not actually binding to the sensor surface may produce false signal after interaction with the evanescent plasmonic field of surface plasmons. While in case of LSPR, biomolecules that are within 40 nm distance of the nanoparticles can be detected due to which the selectivity of LSPR sensors is higher than that of SPR sensors. The LSPR field coupling wavelength for Ag (silver) and Au (gold) nanoparticles falls in the wavelength range of 400–700 nm which coincides with the optical window of imaging sensor of the smartphone camera. With simple optics design and tapping the plasmonic response of nanoparticle–biomolecular interaction, sensing and bio-sensing investigations can be performed using the camera of the smartphone. Till date, Roche et al.²⁷ have reported the possible use of cellphone for LSPR based sensing investigation using AuNPs and AuNRs (gold nanorods) for detection of CCL2 (chemokine ligand 2). In that work they have reported that the sensitivity of their system is 30% less than that of a commercial UV-VIS spectrophotometer for a small incubation period. Also, the optical sensing scheme adopted by them is based on measurement of red and green channels of plasmonic response of AuNPs which cannot be used for average size determination of metal nanoparticles. Till date, no research group have reported a dedicated smartphone based LSPR sensing platform for determining average nanoparticle size as well as detection of biomolecular conjugation with metal nanoparticles in the same platform. In this context we have reported for the first time a smartphone based LSPR sensing platform that can detect the size-variation of metal nanoparticles as well as bio-conjugation induced change in plasmonic response through detection of shift in peak resonance wavelength conditions similar to a commercial spectrophotometer. Recently, we have demonstrated the utilization of smartphone for measuring absorption band of colored dyes²⁸ and pH level of water from different environmental locations.²⁹ The optics design involved for development of such smartphone sensors are simple and relatively inexpensive. In the present work we have primarily utilized the similar optics design and demonstrate the working of a smartphone sensor for LSPR based sensing investigations. As a proof of concept of the designed system, we have shown two important features of LSPR. At first, we synthesize AuNPs of different size and estimate the average size of the nanoparticles by measuring the spectral shift in LSPR wavelength conditions. In the second phase, we made bio-conjugation of AuNPs with BSA protein and trypsin enzyme and measured its spectral shift in peak absorption wavelength that depends on the concentration of the attached biomolecules. A shift in LSPR peak absorption wavelength is a signature of change in size of metal nanoparticles as well as change in immediate environment of the nanoparticles. These two aspects of LSPR coupling conditions have been recorded with our designed sensor and we obtained highly reliable data while comparing the results with the standard spectrophotometer. Further, with the smartphone applications such as ImageJ³⁰ and learnlight spectrometry³¹ which are freely available in Google playstore, it is possible to read the spectral response directly in the phone itself. We envision that the proposed sensor can be developed as a portable, inexpensive LSPR sensing tool, which would be useful for different in-field applications in resource-bound conditions.

Experimental

Materials

The chemicals used for the present experiments have been purchased from the following sources: BSA protein, trypsin enzyme, sodium citrate tribasic dihydrate (C₆H₅Na₃O₇·2H₂O) were procured from SRL, India. Gold chloride trihydrate (HAuCl₄·3H₂O) and sodium borohydride (NaBH₄) were procured from Himedia, India. All chemicals have been used as received without further purification.

Preparation of AuNPs of different size

AuNPs of different sizes have been synthesized following standard procedures.^32–35 Table 1 summarizes the brief procedures followed during synthesis of different sized AuNPs in the present work. 1 mL of 1 wt% HAuCl₄·3H₂O is dissolved in 100 mL of distilled water under constant stirring for 1 minute. Then, 1 mL of 1 wt% C₆H₅Na₃O₇·2H₂O is added to it. Finally, after a period of 1 minute, 1 mL solution containing 0.075 wt% NaBH₄ and 1 wt% C₆H₅Na₃O₇·2H₂O is added under vigorous stirring condition at room temperature. After a period of 5 minutes, AuNP colloids stabilized by negative citrate ions from sodium citrate are formed. Average size of the resultant AuNPs was estimated to be 5 nm. 13 nm AuNPs have been prepared by quickly injecting 10 mL (1.14 wt%) of C₆H₅Na₃O₇·2H₂O into 100 mL (0.039 wt%) boiling solution of HAuCl₄·3H₂O with vigorous stirring. The solution has been heated with continues stirring for 15 minutes and then cooled down to room temperature. The average size of the resultant AuNPs was estimated to be 13 nm. For 20 nm size AuNPs, 0.2 mL (1 wt%) HAuCl₄·3H₂O solution is added to 19.8 mL distilled water and heated to boiling with vigorous stirring. 0.26 mL (1 wt%) of C₆H₅Na₃O₇·2H₂O solution is then added to the boiling solution. After a certain interval of time, the mixture turns wine red indicating the formation of 20 nm AuNPs. 30 nm AuNPs colloids have been synthesized by adding 400 μL of C₆H₅Na₃O₇·2H₂O (1 wt%) to 50 mL boiling solution of HAuCl₄·3H₂O (0.01 wt%) solution. After 10–12 minutes the colour of the AuNPs turns light pink indicating the formation of AuNPs. 360 μL of 1 wt% C₆H₅Na₃O₇·2H₂O is added to a boiling solution of 50 mL (0.01 wt%) HAuCl₄·3H₂O solution. After 10–12 minutes, the colour of the solution turns deep pink indicating the formation of AuNPs of average size 40 nm. The sizes of the synthesized AuNPs have been estimated using 200 keV TEM which has an imaging resolution of 2.4 Å (TECNAI G2 20 S-TWIN, FEI, USA) equipped with Gatan CCD camera. For TEM imaging AuNP samples have been prepared by putting 2 μL of colloidal solution onto a carbon grid and left to dry for a certain period of time. Fig. 1(a–d) shows the TEM images of the AuNPs of size 5 nm, 13 nm, 20 nm and 40 nm respectively.

Table 1 Experimental conditions for synthesis of gold nanoparticles of different sizes

Size (nm)	Volume of HAuCl4·3H2O (mL)	Volume of C6H5Na3O7·2H2O (mL)	Volume of NaBH4 (mL)	Volume of water (mL)	Size confirmation
5	1 (1 wt%)	1 (1 wt%)	1 (0.075 wt%)	100	TEM imaging
13	100 (0.039 wt%)	10 (1.14 wt%)	—	—
20	0.2 (1 wt%)	0.26 (1 wt%)	—	19.8
30	50 (0.01 wt%)	0.4 (1 wt%)	—	—
40	50 (0.01 wt%)	0.36 (1 wt%)	—	—

---

Fig. 1 TEM images of AuNPs of average size (a) 5 nm, (b) 13 nm, (c) 20 nm, (d) 40 nm and (e) photo images of the synthesized AuNPs of size (i) 5 nm, (ii) 13 nm, (iii) 20 nm, (iv) 30 nm and (v) 40 nm.

Fig. 1(e) shows photo images of AuNP colloidal solutions of average particle sizes (i) 5 nm, (ii) 13 nm, (iii) 20 nm, (iv) 30 nm and (v) 40 nm respectively as synthesized for the present investigation. The polydispersity index of the synthesized AuNPs have been measured using dynamic light scattering tool (NanoPlus, zeta/nano particle analyzer, Micromeritics corporation). Table 2 lists the polydispersity index of the AuNPs. The low value of polydispersity index of AuNP samples indicates that the prepared AuNP colloidal samples are highly monodispersive in nature.

Table 2 Polydispersity index of the prepared AuNP colloidal samples

Gold nanoparticle size (nm)	Polydispersity index
5	0.132
13	0.236
20	0.284
30	0.121
40	0.300

---

Conjugation of BSA and trypsin with AuNPs

BSA are known to possess –SH group³⁶ (sulfhydryl group or thiol group) and cysteine residues in their structures.^37,38 The –SH functional group have high affinity towards attachment with gold nanoparticles. Also, trypsin enzyme possesses cysteine residues and lysine residues and the covalent interaction between cysteine residues with AuNP surface causes binding of trypsin with gold nanoparticles.^39,40 Bio-conjugation of BSA protein has been done with 20 nm AuNP colloids. BSA protein solutions of concentration 0.1 mg mL⁻¹, 0.2 mg mL⁻¹ and 0.3 mg mL⁻¹ are prepared by dissolving solid BSA in double distilled water. BSA conjugated AuNPs have been prepared by mixing different concentrations of 100 μl BSA solution to 300 μl of AuNPs. The mixed solutions are then incubated for 2 hours at room temperature. Upon incubation, BSA conjugated AuNP samples are ready for protein quantification. Similarly, trypsin solutions of concentration 0.1 mg mL⁻¹, 0.2 mg mL⁻¹ and 0.3 mg mL⁻¹ have been prepared by dissolving solid trypsin powder in phosphate buffer solution (pH 8). Different concentrations of 100 μl trypsin solution are then treated with 300 μl of AuNPs colloidal solution for conjugation and the mixture is incubated for 2 hours at room temperature. Upon incubation the trypsin conjugated AuNP samples are ready for analysis. The confirmation binding of BSA and trypsin enzyme with AuNPs has been done through standard spectrophotometer (UV 2450, SHIMADZU) and ζ (Zeta) potential analyzer (NanoPlus, zeta/nano particle analyzer, Micromeritics corporation).

ζ-Potential analysis

ζ (Zeta) potential is considered to be a powerful tool in confirming the adsorption of biomolecules with AuNPs.^41,42 The ζ potential analysis has been performed using Zeta potential analyzer (NanoPlus, zeta/nano particle analyzer, Micromeritics corporation). In order to perform the analysis, the ζ potential of bare AuNPs (20 nm) and BSA or trypsin-conjugated AuNPs have been measured to determine the change in their surface charge. Fig. 2 shows the plot for ζ potential values of bare AuNPs and BSA/trypsin conjugated AuNPs. The ζ potential of the unconjugated AuNPs is found to be −24.84 mV. The negative zeta potential is caused by the negative citrate capping over the AuNPs. The ζ potential for BSA and trypsin-conjugated AuNPs has been found to be −16.98 mV and −14.19 mV respectively. The decrease in the zeta potential value of the AuNPs upon conjugation of BSA protein indicates the coverage of BSA protein layer on the surface of AuNPs.⁴³ Similarly, the drop in surface potential due to the trypsin–AuNPs conjugates confirms the attachment of the enzyme with AuNPs.

Fig. 2 Plot for zeta potential values of bare AuNPs and BSA/trypsin conjugated AuNPs.

Working principle of the designed sensor

Fig. 3(a) shows the schematic of our proposed smartphone based LSPR biosensor and Fig. 3(b) shows the handheld photograph of our designed sensor. Light from a broadband optical source (350–1050 nm, Ocean Optics LS-1) is allowed to illuminate a pin hole (50 μm, Edmund Optics) and then gets collimated by a plano-convex lens (FL: 75 mm, Edmund Optics). The collimated beam is then allowed to pass through a quartz cuvette (path length 5 mm, Optiglass U.K) containing the test samples. The transmitted light beam through the cuvette is focused into a thin line beam by a cylindrical lens (focal length 50 mm, Edmund Optics). A plane transmission grating (1200 lines per mm, Edmund Optics) which is attached to the camera of the smartphone is placed at the focal plane of this cylindrical lens that disperses the focused line beam into its constituent colours. The use of cylindrical lens is critical in the present optical set-up. Fig. 3(c and d) shows the dispersed spectrum of the broadband source in the absence and presence of the cylindrical lens respectively. In the absence of cylindrical lens the dispersed spectrum appears as a thin vertical line. The presence of cylindrical lens makes the transmitted modulated beam get focused into a thin line beam which can be effectively coupled with the camera pupil of the smartphone. This would significantly increase the signal to noise (S/N) ratio⁵ for the designed set-up. Owing to the presence of the internal optical filter, the imaging sensor of the camera phone is sensitive to visible spectrum in the wavelength range of 400–700 nm only. For LSPR sensing study we introduce bare (unconjugated) AuNP samples and bio-conjugated AuNP samples in the optical path of our designed sensor. The broadband light signal interacts with the free electrons of the AuNPs due to which a strong wavelength-specific absorption takes place at resonance which can be observed in the dispersed spectrum recorded by the smartphone's camera.

Fig. 3 (a) Schematic of the proposed sensor (b) snapshot of the handheld smartphone integrated LSPR sensing tool and dispersed spectrum of the broadband source in the (c) absence and (d) presence of the cylindrical lens.

Calibration of the designed sensing system

Prior to start the investigation, we calibrate the pixel based visible spectrum information recorded through our smartphone into wavelength scale. At first the dispersed spectrum of the broadband source is recorded by the camera of the smartphone. The broadband optical source is then subsequently replaced by three diode laser sources of known wavelength 405 nm (blue laser), 533 nm (green laser) and 655 nm (red laser). ImageJ software has been used to determine the pixel position for the standard laser sources. The pixel positions of these standard laser sources have been plotted with respect to their corresponding wavelength value. Upon doing the linear fitting of the pixel position with respect to the wavelength values, an excellent goodness of linear fit has been obtained with R² value of 0.99993, shown in Fig. 4(a). Thus, pixel to wavelength calibration follows a strictly linear trend of variation between pixel position against the wavelength values along the dispersive direction. Assuming a linear relationship between the pixel positions and the corresponding wavelength, the linear fitted calibration equation (shown in Fig. 4(a)) has been used to convert the pixel scale into wavelength scale. Fig. 4(b) illustrates the characteristic intensity distribution of all the three laser sources along with the spectrum of the broadband source measured by our designed sensing set-up. The figure shows the plot of intensity distribution along Y-axis with the corresponding wavelength along bottom X-axis and pixels involved along top X-axis. Pixel to wavelength calibration shows that for each pixel shift, the corresponding increment in wavelength is 0.336 nm. The wavelength of the standard laser sources with their corresponding pixel positions are shown in Table S1 in the ESI.†

Fig. 4 (a) Linear fit for pixel versus wavelength variation for three standard laser sources and (b) pixel to wavelength calibrated graph for bare spectrum and the laser sources.

Smartphone camera characterization and image capturing

As optical detector, an iPhone 4 (1 GHz Cortex-A8 processor, 1/3.2′′ sensor size, 1.75 μm pixel) with 5 mega pixel (2592 × 1936 pixels) rear camera has been used. The display resolution of the phone is 330 ppi (640 × 960 pixels). Before taking the image spectra, the camera of the phone is locked at autofocus mode (AF mode). Locking at AF mode ensures that the camera lens can automatically adjust its focus irrespective of the position of the object of interest. For each of the considered samples, 3 consecutive spectrum images have been recorded and subsequently transferred to a desktop for post processing activities.

Results and discussion

Size estimation of gold nanoparticles

At first, we estimate the size of the synthesized AuNP samples using our designed optical set-up. Synthesized AuNPs are placed in the optical path of the set-up for direct light signal coupling from the broadband optical source. Due to LSPR phenomenon, the transmitted optical signal would carry information that contains the characteristic LSPR peak wavelength absorption condition. These peak wavelength absorption conditions would vary with the variation in size of the AuNPs. For smaller sized AuNPs, resonance absorption takes place across lower wavelength but as the average size of the nanoparticles increases, a progressive spectral red-shift of the resonance wavelength occurs which is the signature optical property^44–46 of noble metal nanoparticles undergoing LSPR phenomena. This spectral red-shift with increase in size of AuNPs has been recorded with the designed smartphone sensor. Fig. 5(a) describes the characteristic peak wavelength absorption conditions for the considered AuNP samples recorded by our sensor. With the designed sensor, we observe LSPR peak wavelength value of 514.52 nm, 524.04 nm, 527.29 nm and 531.12 nm for AuNP samples of sizes 5 nm, 20 nm, 30 nm and 40 nm respectively. For reference, the images of the captured spectra of the considered samples as recorded by the smartphone camera have been included. Peak wavelength absorption condition for these samples have been measured with the standard optical spectrophotometer (UV 2450, SHIMADZU) having spectral resolution of 0.1 nm. Fig. 5(b) illustrates the characteristic curves for these samples when measured with the standard tool. The spectrophotometer data reveals that the peak wavelength absorption conditions for 5 nm, 20 nm, 30 nm and 40 nm AuNP samples due to LSPR coupling to be 512.54 nm, 524.22 nm, 527.45 nm and 532.70 nm respectively. By comparing these data we observe that experimental data recorded by the designed smartphone sensor yields fairly accurate result. For clarity we incorporate the plot showing the accuracy of our smartphone sensor in measuring the LSPR peak absorption wavelength values of the considered samples with respect to that of the standard spectrophotometer tool in the ESI section (Fig. S1a†). Standard deviations in measurement of LSPR peak absorption wavelength for 7 identical samples (n = 7) have been indicated by the error bars.

Fig. 5 (a) Characteristic transmission plot of AuNPs of size 5 nm, 20 nm, 30 nm and 40 nm recorded by smartphone sensor and (b) corresponding transmission plot of the AuNPs as measured by standard spectrophotometer.

Resolution and response sensitivity in size estimation

Fig. 6 shows the plot of LSPR peak absorption wavelength conditions for Au nanoparticles of different sizes obtained with our smartphone sensor. The LSPR peak wavelength conditions for 5 nm, 13 nm, 20 nm, 30 nm and 40 nm are measured to be 514.52 nm, 518.18 nm, 524.04 nm, 527.29 nm and 531.12 nm respectively. Based on the LSPR peak wavelength conditions, we infer that peak spectral shift due to LSPR phenomenon varies nearly linearly with increasing size of the AuNPs. The sensor response shows good linear fitted curve with R-squared value = 0.9690. Assuming no background noise present in the sensor set-up, the slope of the linear fitted curve is measured to be 0.4980. Considering a linear relationship between spectral shift with size variation of the AuNPs, the size response sensitivity of the designed sensor was found to be 0.4980. Again, from pixel to wavelength conversion calibration curve (Fig. 4) for our designed smartphone sensor we notice that for each pixel shift in the image spectrum, the wavelength value gets changed by 0.336 nm. Thus, for our designed sensing set-up it can be inferred that the minimum AuNP size variation that can be detected by our sensor is 1.0 nm.

Fig. 6 Linear fitting of LSPR peak absorption wavelength with the increase in AuNPs sizes from 5 nm to 40 nm.

LSPR bio-sensing investigation using the smartphone sensor set-up

Specificity test. In order to demonstrate that the designed sensing platform can perform LSPR bio-sensing investigation, BSA protein and trypsin enzyme conjugated AuNPs have been prepared as discussed in the experimental section. We study the spectral shift in peak resonant wavelength condition due to conjugation of BSA and trypsin with AuNPs. BSA conjugated AuNP samples have been placed in the optical path of the sensing set-up and the corresponding transmitted modulated light signal spectrum is recorded by the camera of the smartphone. The conjugation of different amounts of BSA protein/trypsin enzyme molecules with the AuNPs causes the change in the refractive index of the surrounding dielectric constant of the nanoparticles which causes a shift in its peak resonance wavelength. Fig. 7(a) illustrates the measured transmission characteristic curves of the recorded spectra as measured by our smartphone sensor for different concentrations of BSA conjugated AuNP samples. For reference, we also include the LSPR peak wavelength condition when bare AuNP sample of same particle dimension (20 nm) is placed in the optical set-up. We observe that with reference to bare AuNPs, the resonant wavelength condition is shifted from 524.04 nm to 526.14 nm, 527.29 nm and 528.46 nm for 0.1 mg mL⁻¹, 0.2 mg mL⁻¹ and 0.3 mg mL⁻¹ respectively for BSA conjugated AuNP samples. Inset includes the captured transmission spectra of AuNPs upon conjugation with BSA protein. These resonant wavelength coupling conditions due to LSPR phenomenon for the same samples have been measured using the standard spectrophotometer tool. Fig. 7(b) describes the characteristic transmission curves of the samples as measured by the spectrophotometer. The peak wavelength absorption condition for the same samples recorded by the spectrophotometer are found to be 524.22 nm, 525.31 nm, 526.27 nm and 528.41 nm. For clarity we incorporate the accuracy of our smartphone sensor in measuring the LSPR peak wavelength absorption conditions of the considered BSA–AuNPs conjugates samples with respect to that of the standard spectrophotometer tool in the ESI section (Fig. S1b†). Standard deviations in measurement of LSPR peak absorption wavelength for 7 identical samples (n = 7) have been indicated by the error bars. We notice nearly same amount of resonant wavelength shift for both the devices. We observe that for BSA conjugated AuNP colloids the net resonant spectral shift measured by our smartphone sensor is 4.42 nm when the concentration of BSA is changing from 0 to 0.3 mg mL⁻¹ while for the same samples the standard spectrophotometer measures a net spectral shift of 4.19 nm. The performance of the designed sensor for trypsin-conjugated AuNPs has also been evaluated. Fig. 8(a) shows the characteristic transmission spectroscopic plot of citrate capped AuNPs and corresponding plots of trypsin binded AuNPs as recorded by our designed smartphone sensor. In the absence of trypsin enzyme, citrate capped AuNPs (20 nm) shows peak resonance absorption wavelength at 524.04 nm. After attachment of trypsin enzyme (0–0.3 mg mL⁻¹), the peak resonant coupling wavelength is shifted from 524.04 nm to 525.62 nm, 527.12 nm and 527.50 nm as recorded by the smartphone thus showing a overall shift of 3.46 nm for the considered samples. Inset includes the photo images of captured transmission spectra of AuNPs upon conjugation with trypsin enzyme. Fig. 8(b) shows the characteristic transmission plots for the same samples when measured with the standard spectrophotometer. For these samples we observe a spectral shift from 524.22 nm to 525.02 nm, 527.24 nm and 528.30 nm when measured with the spectrophotometer thus observing a net resonant wavelength shift of 4.08 nm. For clarity we incorporate the accuracy of our smartphone sensor in measuring the LSPR peak wavelength absorption conditions of the considered trypsin–AuNPs conjugates samples with respect to that of the standard spectrophotometer tool in the ESI section (Fig. S1c). Standard deviations in measurement of LSPR peak absorption wavelength for 7 identical samples (n = 7) have been indicated by the error bars. For the above two different biomolecules when attached to the AuNPs, the shift in peak LSPR coupling condition as recorded by our designed sensor is specific for a specific biomolecule and it depends on its concentration.

Fig. 7 (a) Characteristic transmission plot of BSA–AuNPs conjugates (BSA concentration: 0 to 0.3 mg mL⁻¹) recorded by smartphone sensor (b) standard spectrophotometer.

Fig. 8 (a) Characteristic transmission plot of trypsin–AuNPs conjugates (trypsin concentration: 0 to 0.3 mg mL⁻¹) recorded by smartphone sensor and (b) standard spectrophotometer.

Response sensitivity

Fig. 9(a) shows the peak resonance wavelength coupling conditions for BSA samples of different concentration when attached to AuNPs (20 nm). The LSPR peak absorption wavelength conditions were measured to be 524.04 nm, 525.10 nm, 526.14 nm, 527.29 nm and 528.46 nm for 0 mg mL⁻¹, 0.05 mg mL⁻¹, 0.1 mg mL⁻¹, 0.2 mg mL⁻¹ and 0.3 mg mL⁻¹ BSA concentration when attached to 20 nm AuNPs. Again, Fig. 9(b) shows the peak resonance wavelength coupling conditions for trypsin at concentration of 0 mg mL⁻¹, 0.1 mg mL⁻¹, 0.15 mg mL⁻¹, 0.2 mg mL⁻¹ and 0.3 mg mL⁻¹. The corresponding LSPR peak absorption wavelength values were measured to be 524.04 nm, 525.62 nm, 526.16 nm, 527.12 nm and 527.50 nm respectively. An excellent goodness of fit have been observed with R² = 0.965 for BSA and R² = 0.974 for trypsin. The response sensitivity of the sensor can be estimated from the slope of the linear fitted lines in given figures. With 20 nm AuNP colloidal solution the response sensitivity for BSA conjugated and trypsin conjugated AuNPs were found to be 17.466 and 13.063 respectively.

Fig. 9 Linear fitted plot of LSPR peak absorption wavelength vs. BSA/trypsin concentration for (a) BSA (b) trypsin.

Limit of detection (LOD)

Taking into account the operational resolution (R) of the designed smartphone sensor and the sensitivity (S), we evaluate the LOD^47,48 of our system to be

The resolution of our sensor is 0.336 nm. Thus, LOD of our system for BSA and trypsin conjugated AuNPs (20 nm) is calculated to be 19.2 μg mL⁻¹ (equivalent to 0.28 μM) and 25.7 μg mL⁻¹ (equivalent to 1.10 μM) respectively.

The peak wavelength resonant conditions of different sized AuNPs and bio-conjugated AuNP samples measured by the smartphone sensor is slightly different than the readings obtained from the standard spectrophotometer. This is attributed to the fact that our calibration has been done using three diode laser source having finite spectral width. Also, the methodology adopted for pixel to wavelength calibration is different from the one adopted by the commercial spectrophotometer. Hence due to calibration mismatch, deviation in spectral red-shift might have occurred. The technique discussed here for LSPR bio-sensing using a smartphone is relatively simple as compared to commercially available spectrophotometer usually considered for such purpose. The whole set-up can be integrated on a single fixture for attachment to the camera of the smartphone. With the present optical components used for development of the proposed sensor, the dimension of the whole set-up is 9 cm in length, 4 cm in wide 3.5 cm in height. For in-field applications with our designed sensor we can use sun light or any other white light source such as low cost white LED or flash lamp of the smartphone. The overall cost involved in the proposed sensor design excluding the smartphone and broadband optical source is below $250 which includes the cost of pinhole, collimating lens, cylindrical lens, grating and the nylon-made optomechanical attachment and the fabrication charges. Whereas, a standard laboratory grade spectrophotometer (UV 2450, SHIMADZU) costs ∼$ 5547. Further, the miniaturization in device fabrication makes our device truly portable and handy for in-field applications.

Conclusion

In summary, we have demonstrated the working of a prototype LSPR biosensor using the camera of a smartphone. The designed smartphone biosensor meets the criteria of the present generation analytical devices in terms of cost-effectiveness, light weight, portability, good sensitivity, LOD down to μM ranges and data transmission facilities. The additional advantage that our designed sensor offers is simple instrumentation. The performance of our designed sensor has been tested with 20 nm citrate capped AuNPs and the corresponding spectral shift with biomolecular (BSA/trypsin) adsorption on AuNPs has been demonstrated. We have also shown the spectral red-shift of LSPR wavelength of 40 nm AuNPs upon conjugation with BSA and it has been included in the ESI section (Fig. S2†). We observe that with reference to bare AuNP sample, the resonant wavelength condition is shifted from 531.12 nm to 536.64 nm and 544.86 nm after conjugation of 0.1 mg mL⁻¹ and 0.2 mg mL⁻¹ concentration BSA protein respectively. It can be seen from the transmission characteristics that there is a spectral broadening of peak resonant wavelength upon conjugation of 0.2 mg mL⁻¹. The spectral broadening of the resonant wavelength may be due to charge neutralization by interaction of different amino acid residues of the protein with citrate layer leading to agglomeration.⁴⁹ The selectivity of the device in measuring the biomolecular concentration upon conjugation with AuNPs is governed by the affinity of the citrate capped AuNPs towards the biomolecules of interest. The LSPR sensing scheme lacks selectivity features owing to the fact that in complex environment such as biological media, the selective adsorption of different biomolecules having affinity to bind on the bare nanoparticle's surface cannot be selectively controlled. However with the use of specific linkers on the nanoparticle's surface, selective adsorption of biomolecules can be possible. Further, the selectivity can be improved by self assembled monolayer (SAM) on the nanoparticles's surface, by the use of biological scaffolds and size/shape complementarity.⁵⁰ Size dependent spectral red-shift of different sized AuNPs have also been estimated successfully with designed sensor. The shift in LSPR wavelength coupling conditions for BSA & trypsin conjugated AuNP samples have been successfully measured with our smartphone sensor. We anticipate that all types of LSPR based biosensing investigation is possible within the domain of visible spectrum with our designed system. We envision that the proposed sensing design can be developed as an inexpensive handheld smartphone based LSPR biosensor in different fields of applications.

Acknowledgements

P. Nath acknowledges the guidance received from Prof. B. Cunningham on similar line of work during his visit to the University of Illinois at Urbana Champaign. Authors thank to Prof. A. Kumar of Material research laboratory, Dr P. Deb of Advanced functional material laboratory and Dr S. K. Das for allowing us to use their laboratory facilities and to A. Hazarika of Sophisticated and Analytical Instrumentation Centre (SAIC), Tezpur University for taking TEM images of Au samples. Authors would also like to acknowledge Mr P. K. Bora and Mr Rongpi of central workshop, Tezpur University for designing the optical set-up.

References

1. D. N. Breslauer, R. N. Maamari, N. A. Switz, W. A. Lam and D. A. Fletcher, PLoS one, 2009, 4, e6320.
2. D. Tseng, O. Mudanyali, C. Oztoprak, S. O. Isikman, I. Sencan, O. Yaglidere and A. Ozcan, Lab Chip, 2010, 10, 1787–1792.
3. H. Yu, Y. Tan and B. T. Cunningham, Anal. Chem., 2014, 86, 8805–8813.
4. Z. J. Smith, K. Chu, A. R. Espenson, M. Rahimzadeh, A. Gryshuk, M. Molinaro, D. M. Dwyre, S. Lane, D. Matthews and S. Wachsmann-Hogiu, PLoS one, 2011, 6, e17150.
5. M. Arafat Hossain, J. Canning, S. Ast, K. Cook, P. J. Rutledge and A. Jamalipour, Opt. Lett., 2015, 40, 1737–1740.
6. E. K. Grasse, M. H. Torcasio and A. W. Smith, J. Chem. Educ., 2016, 93, 146–151.
7. Q. Wei, R. Nagi, K. Sadeghi, S. Feng, E. Yan, S. J. Ki, R. Caire, D. Tseng and A. Ozcan, ACS Nano, 2014, 8, 1121–1129.
8. S. Lee, V. Oncescu, M. Mancuso, S. Mehta and D. Erickson, Lab Chip, 2014, 14, 1437–1442.
9. V. Oncescu, M. Mancuso and D. Erickson, Lab Chip, 2014, 14, 759–763.
10. M. Zangheri, L. Cevenini, L. Anfossi, C. Baggiani, P. Simoni, F. Di Nardo and A. Roda, Biosens. Bioelectron., 2015, 64, 63–68.
11. I. Hussain and P. Nath, Appl. Opt., 2015, 54, 5739–5742.
12. E. Lebiga, R. E. Fernandez and A. Beskok, Analyst, 2015, 140, 5006–5011.
13. C. F. Fronczek, T. San Park, D. K. Harshman, A. M. Nicolini and J.-Y. Yoon, RSC Adv., 2014, 4, 11103–11110.
14. T. San Park, C. Baynes, S.-I. Cho and J.-Y. Yoon, RSC Adv., 2014, 4, 24356–24362.
15. P. Preechaburana, M. C. Gonzalez, A. Suska and D. Filippini, Angew. Chem., 2012, 124, 11753–11756.
16. H. K. Hunt and A. M. Armani, Nanoscale, 2010, 2, 1544–1559.
17. D. Gallegos, K. D. Long, H. Yu, P. P. Clark, Y. Lin, S. George, P. Nath and B. T. Cunningham, Lab Chip, 2013, 13, 2124–2132.
18. A. J. Haes and R. P. Van Duyne, Anal. Bioanal. Chem., 2004, 379, 920–930.
19. K. A. Willets and R. P. Van Duyne, Annu. Rev. Phys. Chem., 2007, 58, 267–297.
20. C. R. Yonzon, E. Jeoung, S. Zou, G. C. Schatz, M. Mrksich and R. P. Van Duyne, J. Am. Chem. Soc., 2004, 126, 12669–12676.
21. M. Svedendahl, S. Chen, A. Dmitriev and M. Kall, Nano Lett., 2009, 9, 4428–4433.
22. Y. Chen and H. Ming, Photonic Sens., 2012, 2, 37–49.
23. K. M. Mayer and J. H. Hafner, Chem. Rev., 2011, 111, 3828–3857.
24. J. Homola, S. S. Yee and G. Gauglitz, Sens. Actuators, B, 1999, 54, 3–15.
25. Y. Liu, Q. Liu, S. Chen, F. Cheng, H. Wang and W. Peng, Sci. Rep., 2015, 5, 12864.
26. K. Bremer and B. Roth, Opt. Express, 2015, 23, 17179–17184.
27. P. J. Roche, S. Filion-Côté, M. C.-K. Cheung, V. P. Chodavarapu and A. G. Kirk, J. Sens., 2011, 2011, 406425.
28. S. Dutta, A. Choudhury and P. Nath, IEEE Photonics Technol. Lett., 2014, 26, 568–570.
29. S. Dutta, D. Sarma, A. Patel and P. Nath, IEEE Photonics Technol. Lett., 2015, 27, 2363–2366.
30. IJ_Mobile, https://play.google.com/store/apps/developer?id=Antonio%20Garcia%26hl=en, (accessed May 2015).
31. LearnLight Spectrometry, https://play.google.com/store/apps/details?id=com.flappit.learnlight%26hl=en, (accessed August 2014).
32. S. Darwich, PhD thesis, University of HAUTE-ALSACE, 2011.
33. J. H. Kim, R. A. Estabrook, G. Braun, B. R. Lee and N. O. Reich, Chem. Commun., 2007, 4342–4344.
34. Y. Qu and X. Lü, Biomed. Mater., 2009, 4, 025007.
35. N. N. Long, C. D. Kiem, S. C. Doanh, C. T. Nguyet, P. T. Hang, N. D. Thien and L. M. Quynh, Synthesis and Optical Properties of Colloidal Gold Nanoparticles, J. Phys.: Conf. Ser., 2009, 187, 012026.
36. S. Dominguez-Medina, J. Blankenburg, J. Olson, C. F. Landes and S. Link, ACS Sustainable Chem. Eng., 2013, 1, 833–842.
37. S. H. Brewer, W. R. Glomm, M. C. Johnson, M. K. Knag and S. Franzen, Langmuir, 2005, 21, 9303–9307.
38. X. Shi, D. Li, J. Xie, S. Wang, Z. Wu and H. Chen, Chin. Sci. Bull., 2012, 57, 1109–1115.
39. H. Zhang, J. Cao, S. Wu and Y. Wang, Mol. Biol. Rep., 2014, 41, 4911–4918.
40. M. Lv, E. Zhu, Y. Su, Q. Li, W. Li, Y. Zhao and Q. Huang, Prep. Biochem. Biotechnol., 2009, 39, 429–438.
41. S. Zhang, Y. Moustafa and Q. Huo, ACS Appl. Mater. Interfaces, 2014, 6, 21184–21192.
42. M. Schäffler, M. Semmler-Behnke, H. Sarioglu, S. Takenaka, A. Wenk, C. Schleh, S. M. Hauck, B. D. Johnston and W. G. Kreyling, Nanotechnology, 2013, 24, 265103.
43. Â. Barreto, L. G. Luis, A. V. Girão, T. Trindade, A. M. Soares and M. Oliveira, J. Nanopart. Res., 2015, 17, 1–13.
44. S. Link and M. A. El-Sayed, Int. Rev. Phys. Chem., 2000, 19, 409–453.
45. E. Saion, E. Gharibshahi and K. Naghavi, Int. J. Mol. Sci., 2013, 14, 7880–7896.
46. S. Link and M. A. El-Sayed, J. Phys. Chem. B, 1999, 103, 8410–8426.
47. I. M. White and X. Fan, Opt. Express, 2008, 16, 1020–1028.
48. W. Ma, H. Yang, W. Wang, P. Gao and J. Yao, Sensors, 2011, 11, 8643–8653.
49. J. Deka, A. Paul and A. Chattopadhyay, J. Phys. Chem. C, 2009, 113, 6936–6947.
50. S. Unser, I. Bruzas, J. He and L. Sagle, Sensors, 2015, 15, 15684–15716.

---

Footnote

† Electronic supplementary information (ESI) available: Tables S2, S3 and S4 containing the resonant wavelength values as measured by the smartphone sensor and spectrophotometer is included. Also, the price of the optical components involved is included in the supplementary section. See DOI: 10.1039/c6ra01113f

---