Complementary Photothermal Heating Effects Observed between Gold Nanorods and Conjugated Infrared-Absorbing Dye Molecules

: Due to their biocompatibility, ease of surface modiﬁcation, and heating capabilities, gold nanomaterials are considered excellent candidates for the advancement of photothermal therapy techniques and related applications in cancer treatment. Various morphologies of gold nanomaterials have been shown to heat when exposed to high-powered laser irradiation, especially that which is from the near-infrared (NIR) region. While these lasers work well and are effective, light-emitting diodes (LEDs) may offer a safe and low-powered alternative to these high energy lasers. We investigated the heating capability of NIR-dye conjugated gold nanorods when exposed to an 808 nm LED light source using polyethylene glycol (PEG)-coated gold nanorods as the control. In this way, since the rods exhibited a surface plasmon resonance peak between 795 and 825 nm for both the PEG-coated rods and the dye-conjugated rods, which are fairly close to the frequency of the 530 mW, 850 nm LED light source, we were able to reveal the heating effect of the dye modiﬁcation. While both morphologies heat when irradiated with the LED light, we demonstrated that the addition of an NIR dye increases the rate of heating and cooling, compared to the PEGylated counterpart. To our knowledge, the complementary effect given by the conjugated NIR-dye has not been previously reported in the literature. The targeting abilities of the NIR-dye combined with the increased heating rate of the modiﬁed particles used in this proof-of-concept work suggests that these particles may be exceptional candidates for theranostic applications


Introduction
Cancer is a leading cause of death worldwide, and according to the CDC [1], in 2018, nearly 600,000 people died of cancer in the US alone. Traditional cancer therapies, such as radiation, chemotherapy, and surgery are not only uncomfortable for patients, but can also be costly for both the hospital and those undergoing treatment. The Mayo clinic [2] reports that late effects of radiation and chemotherapy include dental problems, heart problems, infertility, lung disease, and an increased risk of other cancers. Moveover, Yabroff [3] suggests that these potential side effects can be difficult for patients to physically handle and may add secondary costs to their treatment, putting further financial liability on people already bearing the bodily demands of therapy and recovery. It is clear that safe, low-cost, effective alternatives are necessary to mitigate the burden of cancer on our healthcare systems.
Alternative therapies, presented by Lee [4], Wenger [5], Rosenberg [6], and Costello [7], such as laser ablation, have been explored, but are currently still invasive and considered to be a surgical procedure. This can be difficult for patients, particularly those who are elderly or immunocompromised. Additionally, the American Cancer Society [8] reports that laser treatment has inherent limitations including lack of trained personnel, intense safety precautions, and the need for repeated treatments. Plasmonic nanomaterials, particularly gold nanomaterials, offer a safe, biocompatible, and non-invasive addition to improve current tumor ablation methods by allowing a tumor to be heated via NIR light of wavelengths that can penetrate the skin. Of particular interest are those materials with surface plasmon resonance (SPR) peaks that fall in the near-infrared (NIR) transparency windows (650-1300 nm), as suggested by Pramanik [9], Maestro [10], Zhou [11] and Quintanilla [12], this range is ideal for safely penetrating through biological tissues.
Various morphologies of gold nanomaterials have been shown to heat when irradiated with NIR light. Most recently, Jiang [13] and Alrahili [14,15] explored the heating capabilities and photothermal conversion efficiencies of different morphologies, from size and shape to different surface modifications of gold nanomaterials, when exposed to a 4 W continuous wave 808 nm laser. It was found that morphologies with SPR peaks closer to the frequency of the laser (808 nm) heated quicker and more efficiently than those with off-resonance SPRs. Specifically, we showed that rods with a geometry of 10 × 41 nm with an inherent SPR peak at 808 nm heated well without modification. We also showed that by conjugating 20 nm gold nanospheres (SPR at 524 nm) with an infrared (IR) dye that absorbed at 808 nm their heating efficiency increased by nearly a factor of three. Moreover, IR dyes have been shown to bind to tumors preferentially without the aid of special ligands or other targeting moieties, as shown by Usama [16], Tan [17], Luo [18], C. Zhang [19], E. Zhang [20] and Zhao [21]. By surface functionalizing the gold nanorods with an NIR dye, we can take advantage of these properties to both target and heat tumors effectively. Ultimately, this will allow for the selective heating of malignant tumor cells, while avoiding healthy cells and leaving them intact.
While laser ablation in conjunction with gold nanomaterials is a feasible therapeutic route, we may also consider LED light sources, as they may offer a safe alternative to lasers in photothermal therapy, according to Schüppert [22] and You [23]. Not only are they lower-powered and eye-safe, NIR-LEDs will not require highly trained personnel and are considerably more cost-effective than high-powered lasers. In this work, we aimed to explore how 10 × 41 nm gold nanorods (GNRs) which have been conjugated with a proprietary IR dye (GNRCs), heat when irradiated by a low power (530 mW), 850 nm LED light source. This wavelength was selected as it is within the aforementioned biological transparency window and any observed heating will provide support for future experiments beyond this proof-of-concept work. We explored how the addition of the dye affected the heating rate and the photothermal conversion efficiency when compared to their PEGylated counterparts, used as a control.

Materials and Methods
Gold nanorods with dimensions of 10 × 41 nm were purchased from Nanopartz (Loveland, CO, USA). The rods were purchased with functionalized amine-terminated polyethylene glycol (PEG) surfaces, which had a 10 kDa molecular weight PEG group, to allow for the IR 808 nm dye conjugation. The PEGylated gold nanorods were conjugated with a proprietary hepta-cyanine based IR-dye by Lahjavida (Colorado Springs, CO, USA). The PEG groups added approximately 40 nm to each side of the rod, while the dye (approximately 1.1 nm in length) did not contribute significantly to particle size. The presence of the dye molecule on the nanorods after synthesis and purification was confirmed through comparative UV-Vis spectra. N,N-Dimethylformamide (DMF, 99.8%) was purchased from Fisher Scientific. The solutions were irradiated with a 530 mW, 850 nm LED light source (LEDSupply, Randolph, VT, USA) in all experiments presented. The emission spectrum of the LED was measured, and a peak can be observed around 860 nm as shown in Figure A1, found in Appendix A. The temperature of the solution was monitored by an Agilent 34970 A data acquisition system equipped with a k-type thermocouple and using Agilent Benchlink Data Logger 3 software. VWR brand glass cuvettes (path length 1 cm) used for sample measurement were purchased from Fisher Scientific. A power and energy meter console (PM100D console, ThorLabs) coupled with a photodiode power sensor (S121C, ThorLabs) was used to measure the incident and transmitted power was purchased from ThorLabs.
Three milliliters (3 mL) of the gold nanorod (GNR and GNRC for the PEGylated and IR-conjugated rods, respectively) solution was placed in a glass cuvette and positioned directly in front of the LED, about 5 mm away from the diode. A heatsink was attached to the LED to ensure that residual heat from its operation was not contributing to the heating of the solutions. The solution was continuously stirred to achieve uniform temperature distribution during LED irradiation. The thermocouple sensor was placed in a central position in the cuvette, keeping the surface area in contact with the solution constant between experiments. To show that the solvent did not contribute to the heating of our system, a control experiment was performed. In this, the pure solvents of the GNRs and GNRCs, ultrapure water (UPH 2 O) and N,N-Dimethylformamide (DMF), respectively, were each added to the cuvette individually and their temperature profiles measured under identical irradiation conditions. During LED irradiation, a negligible temperature increase in the pure solvent was observed, and therefore, we concluded that the gold nanorods were responsible for the temperature changes of the solution ( Figure A2).
For the temperature profile measurements, the stock solutions (optical density of 0.8) as well as the solvents, water and DMF, were investigated. The solutions were irradiated until an equilibrium was reached (T max ) before the LED was turned off (~2500 s) and then were allowed to cool to room temperature. For the concentration dependence measurements, we explored three different dilutions, including 1:2.5, 1:5, and 1:10. Each solution was irradiated for a total of 1000 s and the temperature was monitored for the duration of the experiment. Each experiment was run in triplicate to ensure reproducibility of the results. A schematic of the experimental set up is shown in Figure 1.
Agilent Benchlink Data Logger 3 software. VWR brand glass cuvettes (path lengt used for sample measurement were purchased from Fisher Scientific. A power and meter console (PM100D console, ThorLabs) coupled with a photodiode power (S121C, ThorLabs) was used to measure the incident and transmitted power w chased from ThorLabs.
Three milliliters (3 mL) of the gold nanorod (GNR and GNRC for the PEGyla IR-conjugated rods, respectively) solution was placed in a glass cuvette and pos directly in front of the LED, about 5 mm away from the diode. A heatsink was atta the LED to ensure that residual heat from its operation was not contributing to the of the solutions. The solution was continuously stirred to achieve uniform temp distribution during LED irradiation. The thermocouple sensor was placed in a cen sition in the cuvette, keeping the surface area in contact with the solution const tween experiments. To show that the solvent did not contribute to the heating of o tem, a control experiment was performed. In this, the pure solvents of the GN GNRCs, ultrapure water (UPH2O) and N,N-Dimethylformamide (DMF), respe were each added to the cuvette individually and their temperature profiles measu der identical irradiation conditions. During LED irradiation, a negligible tempera crease in the pure solvent was observed, and therefore, we concluded that the gold rods were responsible for the temperature changes of the solution ( Figure A2).
For the temperature profile measurements, the stock solutions (optical density as well as the solvents, water and DMF, were investigated. The solutions were irr until an equilibrium was reached (Tmax) before the LED was turned off (~2500 s) an were allowed to cool to room temperature. For the concentration dependence m ments, we explored three different dilutions, including 1:2.5, 1:5, and 1:10. Each s was irradiated for a total of 1000 s and the temperature was monitored for the dura the experiment. Each experiment was run in triplicate to ensure reproducibility of sults. A schematic of the experimental set up is shown in Figure 1.

Photothermal Conversion Efficiency Theory and Calculations
Due to the plasmonic nature of the gold nanorod solutions, irradiation of these solutions with LED light leads to a change in the thermal energy. In general, this can be rationalized by considering Joule heating, or similar effects, in the metal nanoparticle during plasmon oscillations. Following from Jiang [13], Alrahili [14] and Jauffred [24], the thermal energy and temperature change can be expressed from the energy balance between the heat given to the system by the light absorbing nanorods, (Q in ), and the heat dissipated from the system to the surrounding environment, (Q out ), where m i and C i are the mass and specific heat capacity of the individual components in the system, respectively, T is the temperature of the solution, and t is time. For the gold nanorod solutions used in the experiments presented here, the mass of GNRs (about 0.000126 g) is significantly less than that of solvent (2.999 g for water and 2.847 g for DMF), and the heat capacity of gold (0.129 J g −1 K −1 ) is also much smaller than that of the solvents (4.18 J g −1 K −1 for water and 2.03 J g −1 K −1 ) [13]; therefore, when considering the summation of all individual components, the solvent is the primary contributor to solution's heat capacity. The input heat that is absorbed by the gold nanorods can be described as, where I 0 is the incident LED intensity, I tr is the intensity of LED light transmitted through the gold nanorod solution, and η is the photothermal conversion efficiency. The heat dissipating from the system to the surroundings can be defined as, where h is the heat transfer coefficient, S is the surface area of the cuvette, T(t) is the temperature at time t, and T 0 is the initial room temperature at ambient conditions. Since the volume fraction of the GNRs and GNRCs is small compared to the volume of solvent, we can assume that only the mass and heat capacity of solvent will play a role in our calculations. From consideration of Equations (2) and (3), along with employment of the mass and heat capacity of the solvent, Equation (1) can be written as, where m s and C s are the mass and specific heat capacity of the solvent, respectively. Here, we define ∆T as the change in temperature (T(t) − T 0 ). Additionally, B = hS m s C s defines the rate constant of heat dissipation, which can be determined by exponentially fitting the curve of the temperature profile after the LED has been turned off. In this regime, where the LED is off, (I 0 − I tr ) = 0, and thus Equation (4) can be reduced to where the initial condition of the LED turned off corresponded to t = 0, and at this point, we have the maximum temperature, T max . At thermal equilibrium, the temperature is constant and Q in = Q out . Therefore, we can now rearrange Equation (4) to find the photothermal conversion efficiency, η, Finally, using the η found in Equation (6), we can determine the theoretical temperature change of the gold nanorod solution when the LED is on (during heating experiments),

Experimental Determination of Photothermal Conversion Efficiencies
The heat dissipation coefficient, B, must be determined experimentally for each system under investigation. Figure 2 shows a fit of the experimental cooling data (after the LED was turned off), where the slope of this exponential was used to determine the B-values for our samples.

Experimental Determination of Photothermal Conversion Efficiencies
The heat dissipation coefficient, B, must be determined experimentally for ea tem under investigation. Figure 2 shows a fit of the experimental cooling data (a LED was turned off), where the slope of this exponential was used to determine values for our samples. Using these experimentally determined values for the heat dissipation coef along with other parameters matched to experiment, a theoretical heating profile w structed for a modeled system of particles (Figure 3). Specifically, we plotted the t ical heating profile for the GNR solution, as these particles' absorption cross sect be determined form Mie theory and can be used to establish the value of . Sin theory is typically employed for spherical particles, we used an equivalent radius, the 10 × 41 nm particles to estimate the volume of a spheroid [25,26]. Using these experimentally determined values for the heat dissipation coefficient, along with other parameters matched to experiment, a theoretical heating profile was constructed for a modeled system of particles ( Figure 3). Specifically, we plotted the theoretical heating profile for the GNR solution, as these particles' absorption cross section can be determined form Mie theory and can be used to establish the value of Q in . Since Mie theory is typically employed for spherical particles, we used an equivalent radius, R eq , for the 10 × 41 nm particles to estimate the volume of a spheroid [25,26].
However, for the dye-conjugated counterparts, Mie theory is not appropriate to use and, therefore, not presented in this work. This was due to a number of assumptions that would need to be made, including the size of the dye, distance of the dye from the rod, as well as number of dye molecules present on the surface of the rod which are beyond the scope of this work. The theoretical temperature profile was, therefore, only constructed for the GNRs, yet it still fits quite well with our experimental data, as seen in Figure 3.
The experimental photothermal conversion efficiency was also calculated by using Equation (6) shown in the previous section. The volume of solution used in the temperature profile experiments was 3 mL and the reported heat capacities of 4.184 J/g • C and 2.030 J/g • C for water and DMF, respectively, were employed for calculations. Using the ThorLabs power and energy meter console with the photodiode power sensor, the incident power density of the LED was measured to be 0.500 W at the front face of the cuvette and the transmitted power density was measured to be 0.100 W after interacting with the gold nanorod solutions, meaning that the solution absorbed 0.400 W of the incident power. The B values (heat dissipation), calculated from the slope of the cooling portion of the temperature profile curve, were found to be 2.87 × 10 −3 s −1 and 4.75 × 10 −3 s −1 for the GNRs and GNRCs, respectively, as shown in Figure 2. From these values, the experimental photothermal conversion efficiency was determined through Equation (6) and calculated to be 50% for the GNRCs and 55% for their unconjugated counterparts. Although it is apparent that the addition of the IR dye did not have much effect on the photothermal conversion efficiency, we must also consider the solvent's characteristics. As reported by Jiang et al. [13], it is important to note that the solvent may play a role in the heat dissipation of the GNRCs, which may in turn impact the B values found from the slope of the line in Figure 2. For comparison, the B value determined for water was 3.26 × 10 −3 s −1 , and for DMF, it was 3.73 × 10 −3 s −1 . In contrast to these relatively similar B values fit from the pure solvent data, the GNRC sample exhibited a significantly faster rate of heat dissipation. Moreover, the heating rate over the first 20 s was found to be 0.016 and 0.023 • C/s for the GNRs and GNRCs, respectively, which is also a substantial increase due to the conjugation of the dye. Analysis of the specific mechanisms involved with this behavior was beyond the scope of the work presented here and is a subject of continued investigation.
OR PEER REVIEW 6 Figure 3. Comparison of theoretically (solid, from Equation (7)) and experimentally (dashed) determined temperature changes for the GNRs. Experimental and theoretical temperature change was determined to be 6.12 and 5.14 °C, respectively.
However, for the dye-conjugated counterparts, Mie theory is not appropriate to use and, therefore, not presented in this work. This was due to a number of assumptions that would need to be made, including the size of the dye, distance of the dye from the rod, as well as number of dye molecules present on the surface of the rod which are beyond the scope of this work. The theoretical temperature profile was, therefore, only constructed for the GNRs, yet it still fits quite well with our experimental data, as seen in Figure 3.  (7)) and experimentally (dashed) determined temperature changes for the GNRs. Experimental and theoretical temperature change was determined to be 6.12 and 5.14 • C, respectively.

LED Heating of the Gold Nanorods and Gold Nanorod Conjugates
For this study, we investigated 10 × 41 nm PEGylated gold nanorods, as well as their IR dye conjugated counterparts, GNRs and GNRCs, respectively. The gold nanorods were characterized by UV-Vis spectrometry (PerkinElmer Lamba 1050) to determine the location of the SPR peaks of the solutions as shown in Figure 4. The GNRs exhibit two SPR peaks, which arise from the transverse and longitudinal collective coherent oscillations of the conduction electrons, resulting in a weak peak in the short wavelength range and a strong peak in the longer wavelength range. Since our LED has a wavelength of 850 nm and the first NIR window is in the 700-900 nm range, we were interested in the strong peak which occurs at 794 nm for the GNRs and 822 nm for the GNRCs. To this end, the OD of the solutions was matched at around 0.8 in this regime.
beyond the scope of the work presented here and is a subject of continued investigation.

LED Heating of the Gold Nanorods and Gold Nanorod Conjugates
For this study, we investigated 10 × 41 nm PEGylated gold nanorods, as well as their IR dye conjugated counterparts, GNRs and GNRCs, respectively. The gold nanorods were characterized by UV-Vis spectrometry (PerkinElmer Lamba 1050) to determine the location of the SPR peaks of the solutions as shown in Figure 4. The GNRs exhibit two SPR peaks, which arise from the transverse and longitudinal collective coherent oscillations of the conduction electrons, resulting in a weak peak in the short wavelength range and a strong peak in the longer wavelength range. Since our LED has a wavelength of 850 nm and the first NIR window is in the 700-900 nm range, we were interested in the strong peak which occurs at 794 nm for the GNRs and 822 nm for the GNRCs. To this end, the OD of the solutions was matched at around 0.8 in this regime. The surface-modification of the gold nanorods was shown to play a role in the heating rate of these solutions, as seen in Figure 5. On average, the temperature change of the GNRs was found to be 6.12 °C, and the temperature change of the GNRCs was found to be 7.28 °C. While this is not a substantial difference between the two solutions, the GNRCs clearly reach equilibrium quicker than their unconjugated counterparts and the rate of heating between the two solutions was found to be 44% higher for the GNRCs than the GNRs alone. As the gold nanorods are identical in composition apart from the addition of the dye in the GNRCs, these observations suggest that the addition of the dye plays a role in the overall heat transfer capabilities of the gold nanorods. For additional temperature analysis, including that of the solvents, see Figures A2 and A3 in Appendix A. The surface-modification of the gold nanorods was shown to play a role in the heating rate of these solutions, as seen in Figure 5. On average, the temperature change of the GNRs was found to be 6.12 • C, and the temperature change of the GNRCs was found to be 7.28 • C. While this is not a substantial difference between the two solutions, the GNRCs clearly reach equilibrium quicker than their unconjugated counterparts and the rate of heating between the two solutions was found to be 44% higher for the GNRCs than the GNRs alone. As the gold nanorods are identical in composition apart from the addition of the dye in the GNRCs, these observations suggest that the addition of the dye plays a role in the overall heat transfer capabilities of the gold nanorods. For additional temperature analysis, including that of the solvents, see Figures A2 and A3 in Appendix A.
As expected, concentration also plays a role in the heating capabilities of the gold nanorod solutions. As we decrease the overall concentration of gold present in the solution through dilution, the temperature change decreases, accordingly, as seen in Figure 6. The stock solution of the GNRs, as provided by Nanohybrids, had a reported concentration of 7.4 × 10 11 rods/mL at OD 1. Since the solutions were matched to have an OD around 0.8, the starting concentration of the solutions was calculated to be 5.92 × 10 11 rods/mL. To investigate the concentration dependence, we diluted by a factor of 2.5, 5, and 10, resulting in concentrations of 2.37 × 10 11 , 1.18 × 10 11 , and 5.92 × 10 10 rods/mL, respectively. When the solutions are diluted by a factor of 2.5, the change in temperature is 3.62 • C for the GNRs and 4.06 • C for the GNRCs. When diluted by 5, the temperature change is 2.68 • C for the GNRs and 3.43 • C for the GNRCs. Finally, when diluted by a factor of 10, the temperature change is 1.3 • C for the GNRs and 2.40 • C for the GNRCs. This trend is expected as with each dilution, the predominant component responsible becomes overwhelmingly the solvent, which was shown to not contribute significantly to heating. We also theoretically calculated the temperature change for the various concentrations of the GNRs (Figure 7) and again, the theory matches fairly well with the experimental values.
, FOR PEER REVIEW 8 Figure 5. Heating profiles of the GNRs (blue) and GNRCs (green). The curves represent the average of 3 measurements. The maximum ΔT for the GNRs and GNRCs is 6.12 and 7.28 °C, respectively.
As expected, concentration also plays a role in the heating capabilities of the gold nanorod solutions. As we decrease the overall concentration of gold present in the solution through dilution, the temperature change decreases, accordingly, as seen in Figure 6. The stock solution of the GNRs, as provided by Nanohybrids, had a reported concentration of 7.4 × 10 11 rods/mL at OD 1. Since the solutions were matched to have an OD around 0.8, the starting concentration of the solutions was calculated to be 5.92 × 10 11 rods/mL. To investigate the concentration dependence, we diluted by a factor of 2.5, 5, and 10, resulting in concentrations of 2.37 × 10 11 , 1.18 × 10 11 , and 5.92 × 10 10 rods/mL, respectively. When the solutions are diluted by a factor of 2.5, the change in temperature is 3.62 °C for the GNRs and 4.06 °C for the GNRCs. When diluted by 5, the temperature change is 2.68 °C for the GNRs and 3.43 °C for the GNRCs. Finally, when diluted by a factor of 10, the temperature change is 1.3 °C for the GNRs and 2.40 °C for the GNRCs. This trend is expected as with each dilution, the predominant component responsible becomes overwhelmingly the solvent, which was shown to not contribute significantly to heating. We also theoretically calculated the temperature change for the various concentrations of the GNRs (Figure 7) and again, the theory matches fairly well with the experimental values.

Discussion
In this work, we demonstrated that, not only are LED light sources an effective alternative to high-powered lasers, but also that the addition of an IR dye to the surface of gold nanorods increases their overall effectiveness. The complementary effect presented by such a conjugated IR-dye has not been previously reported and could introduce novel ways of enhancing the heating capabilities of gold nanoparticles. The heating rate of the GNRCs was found to be 63.4% higher than the GNRs, which demonstrates that rods functionalized with the IR dye allow for fast and efficient heating. This matches with previous work from our lab where Alrahili et al. [14] showed that spheres conjugated with this dye also heated faster and more efficiently than their unconjugated counterparts when exposed to a high-power laser. When considering eventual clinical application of low-power LEDs in photothermal therapy, the nanorods did heat a sufficient amount to trigger cell death-which occurs 6 °C above the standard body temperature at 43 °C as reported by Kim [27]-heating about 7 °C overall. To further strengthen this claim, we must also recognize that when allowed to circulate through the body of a patient, the tumor-targeting capabilities of the NIR-dye, shown by Usama [16], Tan [17], Luo [18], C. Zhang [19], E. Zhang [20] and Zhao [21], would allow for locally high concentrations of these nanoparticles to accumulate in and around cancer cells. This, in turn, could potentially increase

Discussion
In this work, we demonstrated that, not only are LED light sources an effective alternative to high-powered lasers, but also that the addition of an IR dye to the surface of gold nanorods increases their overall effectiveness. The complementary effect presented by such a conjugated IR-dye has not been previously reported and could introduce novel ways of enhancing the heating capabilities of gold nanoparticles. The heating rate of the GNRCs was found to be 63.4% higher than the GNRs, which demonstrates that rods functionalized with the IR dye allow for fast and efficient heating. This matches with previous work from our lab where Alrahili et al. [14] showed that spheres conjugated with this dye also heated faster and more efficiently than their unconjugated counterparts when exposed to a high-power laser. When considering eventual clinical application of lowpower LEDs in photothermal therapy, the nanorods did heat a sufficient amount to trigger cell death-which occurs 6 • C above the standard body temperature at 43 • C as reported by Kim [27]-heating about 7 • C overall. To further strengthen this claim, we must also recognize that when allowed to circulate through the body of a patient, the tumor-targeting capabilities of the NIR-dye, shown by Usama [16], Tan [17], Luo [18], C. Zhang [19], E. Zhang [20] and Zhao [21], would allow for locally high concentrations of these nanoparticles to accumulate in and around cancer cells. This, in turn, could potentially increase the magnitude of observed heating, as was demonstrated in our dilution experiments. An additional approach to achieving larger heating effects and better understanding the role of each parameter would be to further optimize experimental conditions by exploring various LED power ranges, optical densities of the solutions, distances from the source, and various solvents. Photothermal therapy for cancer treatment using plasmonic nanomaterials has been hindered by the lack of targeting capabilities, as well as the use of unsafe lasers. However, this work opens the door for continued exploration of alternative cancer therapy options using low-powered, safe, and user-friendly LED light sources.    GNRs and GNRCs were dissolved in water and DMF, respectively. These temperature changes are much smaller than that observed in either of the gold nanorod solutions. Figure A3. Average change in temperature (ΔT) for solvents and samples. The GNR sample was in H2O and the GNRC sample was in DMF. Error bars are presented for the standard deviation between three experiments.