Fig. 1. The portable THz handheld scanner. (a) Handheld system and skin diagram. (b) A single raw THz reference signal reflected from a gold mirror and the result of averaging 20 pulses. (c) Fourier spectra indicating that averaging increases the dynamic range by over 10 dB. (d) Diagram illustrating the two reflections of the THz pulse, one from the air–quartz interface labeled as “A,” and one from the quartz–skin–air interface labeled as “B.” (e) Example of a typical THz-TDS trace reflected from the interface air–quartz (A) and quartz–skin (B). (f) Photograph of the handheld probe scanning the volar forearm of a volunteer. (g) Two force-sensitive resistors are fitted on the tip of the probe to ensure consistent force applied onto the skin for every measurement.

Fig. 2. Comparisons of the three systems and their corresponding measurements of volunteer 1. Photos of the actual systems scanning the volar forearm of a volunteer. (a) Proposed handheld probe, (b) system 1 (Menlo K15) with a fixed window platform, and (c) system 2 (TeraView TPS4000 Gantry system). (d) The refractive index and (e) absorption coefficient results for volunteer 1 from the handheld probe (red dots), system 1 with a fixed window platform (blue dots), and system 2 (black dots). The results shown are the average of 20 measurements (dots) from between 55 and 60 s of occlusion and the corresponding standard deviation (error bars).

Fig. 3. Experimental occlusion curves obtained with the handheld system. (a) Impulse function recorded during 60 s of measurements from the volar forearm of a single volunteer. The pulses have been offset horizontally for clarity. The P2P amplitude decays as a function of recording time due to the occlusion effect. (b) P2P curves (occlusion curves) and their bi-exponential fit (dots and continuous lines, respectively) of seven different volunteers highlighting the variation across volunteers. The definition of ΔP2P is given as the vertical difference between the first and last point on the occlusion curve and is an indicator of hydration. The color bar indicates the occlusion time for the corresponding data plotted.

Fig. 4. (a) The water profile distribution follows a quadratic increase in the SC and a linear increase within the epidermis. The parameters H0, H1, and d characterize this water profile. (b) Simulated water distribution by varying the H0 parameter as a function of time following a bi-exponential curve (inset figure) to simulate the occlusion effect for 60 s. (c) Simulated THz pulses using the hydration profiles shown in panel (b). The pulses have been offset horizontally for clear visualization. (d) Simulated occlusion curves when the parameter H0ini is varied from 10% to 35% to simulate skin with different initial hydration conditions, while H1 and d have been kept constant at 70% and 15μm, respectively. (e) Simulated occlusion curves when the parameter d varies from 15 to 25μm, while H0 and H1 have been kept constant at 15% and 70%, respectively. The main effect of the thickness of SC is the vertical offset of the occlusion curve.

Fig. 5. Scatterplots of the data. (a) Cloud of points of the participants who answered to have normal skin, dry skin, and dry skin being treated with moisturizer. (b) P2P50 versus skin hydration for all the participants. (c) P2P50 versus SC thickness. (d) Skin hydration (H0) versus SC thickness. Different colors have been employed for volunteers with normal skin (green), dry skin (red), and dry skin previously treated with moisturizer (blue). In all the cases, a clear clustering of people with dry skin is observed. Meanwhile, treated dry skin exhibits similar trends as normal skin.

Fig. 6. Histogram showing the number of participants at each skin hydration for dry (red) and normal (green) skin participants (shaded bar chart, axes RHS). The probability density functions (solid lines, axis LHS) are calculated for dry (red) skin and normal (green) skin. The mean hydration values for each group are indicated by the vertical dashed lines.

Fig. 7. Effect of the change of the SC thickness on the occlusion curve. Three regions on the volar forearm of a volunteer are measured in panels (a)–(c) before (blue squares) and after (red squares) the reduction in the SC thickness using the tape stripping method. The reduction of the SC thickness is seen as a vertical down offset in the position of the occlusion (P2P) curve. The three regions are each measured 4 times, the squares represent the mean values, and the error bars show the standard deviation.

Fig. 8. Accumulation of water as a function of depth into the skin and occlusion time in the skin for nine different volunteers. (a)–(c) Three with normal skin, (d)–(f) three exhibiting dry skin, and (g)–(j) three with dry skin previously treated with moisturizer. Higher hydration levels are found closer to the surface for normal skin at the beginning of the measurement starting at values close to 20%, while for dry skin the initial values are below 20%.

Fig. 9. Numerical modeling of skin. Hydration profile as a function of the skin depth in which the first point, H0 is the skin surface hydration, H1 is the epidermis hydration, and d is the SC thickness.