Setting up TCA readers with CW vs. pulsed lasers
To achieve ultra-high signal amplification fold on the GNP labels, the TCA system can be improved by increasing the laser energy fluence. During laser irradiation, the heat generation of a GNP,(dot{{Q}_{GNS}}), can be estimated as
$$dot{{Q}_{GNP}}={C}_{abs}bullet {I}_{0},$$
(1)
where ({C}_{abs}) is GNPs absorption cross section (unit: ({mathrm{mm}}^{2})), and ({I}_{0}) is the energy fluence of laser irradiation (unit: (mathrm{W}cdot{mathrm{mm}}^{-2})). Increasing ({I}_{0}) creates a higher photothermal response from GNPs ((dot{{Q}_{GNS}})), which could help lower TCAs detection limit of GNPs in LFA. In most previous studies, a CW laser at 532nm was used in TCA and the regular irradiation power on LFAs was set as~25mW10,11,13,15. The measured diameter of the laser spot on LFA was about 0.1mm13, whose average input energy fluence, ({I}_{0}), was estimated as 3.2(mathrm{W}cdot{mathrm{mm}}^{-2}) (Table 1).
To maximize the photothermal response of GNPs, the traditional CW laser was upgraded to a pulsed laser with higher energy fluence. Here, a 1064nm Nd:YAG laser (iWeld 980 Series, 120J, LaserStar Technologies, FL, USA) was used to provide a high-energy singular millisecond pulse, as shown in Supplementary Fig. S1a. As calibrated, the highest laser pulse energy was 60.64J within 20ms16. For a 2mm spot, the energy fluence from the pulsed laser was up to 955.4(mathrm{W}cdot{mathrm{mm}}^{-2}), about 300-fold higher than that in previous studies10,11,13,15. To maximize (dot{{Q}_{GNP}}) under the same laser irradiation, the GNS was chosen over other GNPs, such as gold nanorods (about 90nm in length and 15nm in width) which also absorb strongly at 1064nm, because GNS has larger ({C}_{abs}) than other GNPs as characterized in a previous study17. Table 1 compares the (dot{{Q}_{GNP}}) of different GNP-laser settings. The GNS-pulsed laser (400V) setting has the highest heat generation which can be as high as 2080-fold of that for the 30nm gold nanosphere (GNSp)-CW laser (25mW) setting. Thus, it was chosen to test the limit of TCA. However, less than maximum pulsed laser intensity (22.3(mathrm{W}cdot{mathrm{mm}}^{-2})) was used to test GNS-loaded NC membrane (model LFA) since it was prone to burn under more intensive irradiation.
To test the limit of TCA, both TCA readers equipped with CW laser and pulsed laser were set up to compare their limits of detection (LoDs) for GNPs precoated in NC membrane and on coverslips as immunoassay models. Their schematic setup is shown in Fig.1a,c. More details on CW laser TCA can be referred to our previous work9,13. Details of ultrafast TCA setup and characterization are provided in Supplementary Sect. S1. As compared between Fig.1b,d, different lasers enable different heating intensity and speed. When heating a GNP spot with an ms pulsed laser, the heating energy from pulsed laser was confined within the laser spot which, in turn, enabled a much higher temperature increase than CW laser heating (detailed in Supplementary Sect. S4). The temperature increase of a GNP spot can be done within ms by pulsed laser heating while CW laser would need many seconds to heat the spot. As summarized in Fig.1e, faster reading can be achieved with the pulsed laser ultrafast TCA (seconds) than CW laser TCA either with discrete or continuous reading algorithms (115min) as detailed in previous work9,13,14. Additionally, different temperature measurement products (IR camera vs. sensor) were used to fit with the lasers as summarized in Fig.1e.
TCA readers equipped with continuous-wave (CW) laser vs. pulsed laser. (a) Arrangement of the laser path, IR camera, and testing platform, such as a substrate coated with gold nanoparticle (GNP) spot, in the CW laser TCA reader. (b) Schematic record of temperature response of a GNP spot heated by CW laser. (c) Arrangement of the laser path, IR sensor, and testing platform in the ultrafast TCA reader equipped with an ms pulsed laser. The gray area was the field of view of the IR sensor, which depends on the alignment parameters, d and (theta) (detailed in Supplementary Sect. S1). (d) Schematic record of temperature response of a GNP spot read by pulsed laser. (e) Comparison of reading time, laser heating time scale, and temperature measurement products in ultrafast TCA vs. CW laser TCA with continuous and discrete reading algorithms.
In addition to lasers, GNP-loaded substrates being irradiated also impact the thermal responses. In general, substrates with lower thermal mass and higher tolerance for laser intensity against thermal damages will achieve higher thermal signals. Table 2 lists three substrates (NC membrane, plastics, and coverslip) that can be potentially used for immunoassays and testing the limit of TCA. NC membrane (widely used in LFAs) and coverslip were chosen as substrates to be tested in this study since they had significant differences in both thermal mass normalized by volume and maximum temperature without thermal damage.
To test the limit of TCA, we compared thermal signals of the (pulsed laser) ultrafast TCA with CW laser TCA when reading the same model LFAs (GNS-loaded NC membrane) as seen in Fig.2. The UVvis-NIR extinction spectrum of the GNS is shown in Supplementary Fig. S3. The intensity output of the pulsed laser was set at 22.3(mathrm{W}cdot{mathrm{mm}}^{-2}) (Table 1) to avoid thermal damage to NC membrane, whose thermal signals are shown in Fig.2a. For CW laser TCA, both traditional discrete reading and continuous reading (i.e., fast reading) were applied and results were plotted in Fig.2b,c, respectively. The CW laser intensity was set at 12.7(mathrm{W}cdot{mathrm{mm}}^{-2}) (100mW, Table 1), nearly twofold lower than that from ultrafast TCA. Compared with visual reading of model LFAs, TCA readings showed a 10- to 20-fold reduction in LoD for GNSs loaded in NC membrane, as shown in Fig.2ac. The ultrafast TCA had higher thermal signals than CW laser TCA for the controlled GNS concentrations, as compared in Fig.2d. However, it also had much higher background noise for the blank NC membrane (i.e., without GNSs). We speculate that this is due to the limitation of the IR sensor. Ideally, the acquisition time of the IR sensor should be at least tenfold smaller than the pulse width (3ms) to ensure the accuracy and consistency of the signal acquisition. Unfortunately, in our case, the IR sensor, which was chosen based on its small size to fit into laser chamber and price consideration, had a comparable acquisition time of 3ms (Fig.1e) despite the claim that it could show interpolated temperature at 1ms interval; this may contribute to some noise or inconsistency in the reading. In contrast, the CW laser TCA had a much faster temperature acquisition (16.7ms) than the laser heating time scale (seconds), thus with high reading consistency. Perhaps, as a result, the current ultrafast TCA setup did not show an apparent benefit in signal amplification to read model LFA compared to the fast TCA. The lowest LoD was achieved by the fast TCA reading (i.e., using CW laser and continuous reading algorithm), and was twofold lower than those from ultrafast TCA and the other discrete reading algorithm. Future optimization may consider a more advanced IR sensor, although a higher cost is expected. Alternatively, increasing lasers pulse width can reduce the impact of IR sensors inadequate sampling, which can also enhance thermal signals with a significant increase in laser energy fluence. Since NC membrane was prone to pyrolysis and burn under intensive laser heating (Table 2), another assay substrate (i.e., glass) was considered to test the limit of TCA in the next section.
Reading gold nanoparticles in nitrocellulose (NC) membrane as model lateral flow immunoassays (LFAs) by TCAs with continuous wave (CW) laser vs. pulsed laser. NC membrane was precoated with diluted silica-cored gold nanoshells (GNSs) as model test regions in lateral flow immunoassays. (a) Thermal signals from ultrafast TCA reading with a pulsed laser (22.3(mathrm{W}cdot{mathrm{mm}}^{-2}), 170V, 1.41J, 3ms) (red). (b) Thermal signals from CW laser TCA reading with a discrete reading algorithm (yellow). (c) Thermal signals from fast TCA reading with CW laser and continuous reading algorithm (blue). (d) Comparison of these thermal signals from different TCA readings. Round shadows: limits of detection (LoDs) for GNSs. Square shadow (gray): visual cutoff to read GNS spot in NC membrane (model LFA). Statistical significance is indicated with asterisks: ns: p>0.05; *p<0.05; **p<0.01. The GNS concentration in NC membrane was the projected surface concentration=volumetric concentration (times) membrane thickness.
For even higher signal amplification, proof-of-concept measurement was conducted by TCA reading of GNSs pre-coated on a glass coverslip as a model MIA, which can tolerate much higher irradiation intensity than either paper or plastic (see Table 2 and Fig.3a). To maximize the thermal signals in measurement, the maximal energy output of the pulsed laser (400V, 60.64J, 20ms pulse width, and 2mm spot size, ({I}_{0}=) 955.4(mathrm{W}cdot{mathrm{mm}}^{-2})) in ultrafast TCA was applied to detect GNSs on the coverslips in Fig.3b. A stricter metric (IUPAC, see Methods) was applied to extrapolate the LoD for GNSs on coverslip by ultrafast TCA reading rather than ANOVA analysis which was used as default for other measurements. To understand the benefit of coverslip, its thermal signals were compared with those of model LFAs with NC membrane read by ultrafast TCA but at lower pulse energy (1.41J) to avoid thermal damage (Fig.3c). Unlike model LFAs, the GNS-coverslips in Fig.3b were all subvisual due to poor visual contrast, while the visual cutoff of model LFAs was shown in Fig.3c. Regarding ultrafast TCA reading as compared in Fig.3d, the coverslips had higher thermal responses than model LFAs for the same GNS concentrations. The thermal LoD for GNSs on coverslip was also lower (~57-fold) than the visual LoD for model LFAs. This suggests that increasing laser pulse energy enabled higher thermal responses, which compensated for the large thermal mass of coverslip. Since coverslip has better thermal tolerance, 20ms pulse was applied, which was~6.7-fold longer than the acquisition time of IR sensor (3ms). Thus, the sensor sampling issue that may have influenced readings in the model LFA case (Fig.3c) was likely not an issue here (Fig.3b). Further modeling and discussion on substrate comparison for TCA are provided in Supplementary Sect. S4 to potentially achieve even higher thermal signals and thus better signal amplification from TCA reading. Certainly, finding a sensor that can operate under even shorter pulses with improved signal-to-noise will also help.
Testing the limit of thermal contrast amplification (TCA) by improving substrates for ultrafast TCA reading. Thermal signals were measured through ultrafast TCA reading silica-cored gold nanoshells (GNSs) precoated in nitrocellulose (NC) membrane and on coverslips at the same projected surface concentrations as model test regions in immunoassays. (a) Experimental tolerance of laser pulse energy by the tested GNS-NC membrane and GNS-coverslip systems. (b) Thermal signals from GNS-coverslips with maximal laser pulse energy (60.64J) over 20ms. (c) Thermal signals from GNS-NC membrane with laser pulse energy at 1.41J over 3ms to avoid thermal damage. (d) Comparison of these thermal signals from different substrates. Blue round shadow: limits of detection (LoDs) for GNSs in NC membrane. Square shadow (gray): visual cutoff to read GNS spot in NC membrane. Dashed line: extrapolated LoD for GNSs on coverslip by IUPAC metric. All the coverslip cases were subvisual. Statistical significance is indicated with asterisks: ns: p>0.05; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. The GNS concentration in NC membrane was the projected surface concentration=volumetric concentration (times) membrane thickness.
Figure4a compared thermal signals from GNS-coverslip and GNS-NC membrane (or model LFA) when being read by their respective optimal TCAs. The LoD for GNSs in the coverslip case (1.24E3 GNSs/mm2) was still about 2.85-fold lower than that of the NC membrane case. This further proved that increasing the laser fluence can improve thermal response and signal amplification fold via TCA reading, and thus the sensitivity of immunoassays. Figure4a also showed that the background noise of blank samples for ultrafast reading of GNS-coverslip was around 1C, much higher than GNS-NC membrane with fast TCA reading, which may set the major limit to an even lower LoD. This noise might be due to the system error of the ultrafast TCA, absorption of laser energy by glass, etc. For even greater MIA sensitivity enhancement by TCA, future efforts would be needed to reduce the background noise.
(a) Comparison of thermal signals from diluted silica-cored gold nanoshells (GNSs) precoated in nitrocellulose (NC) membrane (model LAF) and on coverslips as model test regions in immunoassays when being read by their respective optimal thermal contrast amplification (TCA) systems. Model LFA was read by fast TCA (i.e., continuous-wave (CW) laser with a continuous reading algorithm) while coverslips were read by ultrafast TCA at maximal energy output. Blue round shadow: limits of detection (LoDs) for GNSs in NC membrane. Square shadow (gray): visual cutoff to read GNS spots in NC membrane. All the coverslip cases were subvisual. Dashed line: extrapolated LoD for GNSs on coverslip by IUPAC metric. (b) Summary of the LoDs for GNSs precoated in/on different substrates (i.e., NC membrane or coverslip) and read by different TCA systems. Their corresponding amplification folds were calculated by comparing them with visual cutoff for reading GNSs in NC membrane. For NC membrane, GNS concentration was the projected surface concentration=volumetric concentration (times) membrane thickness.
To summarize, Fig.4b shows the LoDs for GNSs measured on various substrates (NC membrane vs. coverslip) when being read by different TCA systems (CW laser vs. pulsed laser). The signal amplification folds were normalized by the visual cutoff of reading model LFAs, which is a conventional readout format for commercial LFAs. The coverslip and ultrafast TCA with maximal pulsed laser energy output had the maximal signal amplification (57-fold), followed by the model LFA with fast TCA reading (20-fold). When reading model LFAs, the discrete reading by CW laser TCA showed a similar amplification fold (tenfold) to the ultrafast TCA. It is also expected that the amplification fold by ultrafast TCA could be further improved by reducing the background noise and/or using a better IR sensor (faster response), despite the higher cost and other changes in TCA setup. For future ultrafast TCA-MIA applications, the consideration of assay kinetics and design was also discussed in Supplementary Sect. S5 apart from signal amplification. Overall, TCA is able to enhance signals for both LFAs and MIAs. MIA with TCA is promising for future ultrasensitive POC diagnostics, although further improvement in reducing background noise will be needed if further signal amplification is needed or required.
Here is the original post:
Fast and ultrafast thermal contrast amplification of gold nanoparticle-based immunoassays | Scientific Reports - Nature.com