**2. Materials and Methods**

The films were deposited simultaneously on quartz glass, Si, and soda-lime glass substrates by thermal pulsed chemical vapour deposition using a TFS 200 ALD system (Beneq Oy, Espoo, Finland). The deposition temperature and chamber pressure were 125 ◦C and 2 mbar, respectively, with 100, 200, 500, or 1000 deposition cycles. The precursor for copper was [Bis(trimethylsilyl)acetylene]-(hexafluoroacetylacetonato)copper(I) (Sigma-Aldrich, St. Louis, MO, USA), abbreviated to CuBTMSA for convenience heated to 80 ◦C, and the precursor for chlorine was pyridine hydrochloride (Sigma-Aldrich, ≥98%) heated to 50 ◦C. One CuCl deposition cycle was defined by the following sequence. Cu pulse (2 s) → N2 purge (4 s) → Cl pulse (3 s) → N2 purge (6 s). Nitrogen was also used as a carrier gas. Some samples were protected against oxidation and atmospheric moisture by an Al2O3 capping layer. This capping layer was deposited in situ by ALD from trimethylaluminium (TMA, Strem Chemicals, Newburyport, MA, USA; electronic grade) and ozone (ozone generator, BMT Messtechnik, Stahnsdorf, Germany; 8 g/h output) at 150 ◦C. One Al2O3 ALD deposition cycle was defined by the following sequence. TMA pulse (0.1 s) → N2 purge (3s) → O3 pulse (0.3 s) → N2 purge (3 s). Ozone was used as the oxidiser here because of the fear that

using water might hydrolyse the CuCl. The capping layer consisted of 47 ALD cycles giving an Al2O3 thickness of approximately 5 nm. The deposition chamber was connected to a nitrogen glove box with moisture and oxygen content values < 0.1 ppm. The samples were stored in the glove box and only removed for the time necessary for analyses except for those capped samples deliberately exposed to the atmosphere.

ALD is characterised by growth saturation with increasing precursor dose and the presence of an ALD temperature window. In the process described here, the full range of deposition parameters has not yet been explored with these precursors to definitively prove that the CuCl deposition was purely an ALD process. Therefore, the process can only strictly be called sequentially pulsed chemical vapour deposition. We consider, however, that in comparing the growth of different samples it is in order to refer to the number of "ALD cycles" as a shorthand descriptor rather than saying "sequentially pulsed precursor-purge cycles".

The CuCl was deposited on various types of substrate: (i) silicon wafers (100) (P/Boron doped, thickness = 525 ± 25 μm, *Ω* = 10–30 Ω·cm); (ii) soda-lime glass; and (iii) quartz glass Herasil 102 (Heraeus, Hanau, Germany). Before the deposition the samples of size approximately 1 × 1 cm2 were cleaned in an ultrasonic bath: 5 min in acetone, 5 min in isopropanol, and dried with nitrogen. Hydrofluoric acid was used to etch native oxide from the Si wafer surfaces. Transparent samples (soda-lime glass and quartz glass) were coated with Kapton tape on the back side to allow deposition only on one side. The Kapton tape was removed prior to optical measurements.

X-ray photoelectron spectroscopy (XPS) was carried out using an ESCALAB 250Xi apparatus (ThermoFisher Scientific, Waltham, MA, USA) with monochromated Al Kα radiation (1486.6 eV) to determine the chemical composition of CuCl thin films. All samples were measured at two spots at a takeoff angle of 90◦ in 10−<sup>8</sup> mbar vacuum at 20 ◦C. An electron flood gun was used to compensate for charges on sample surfaces. The spectra were referenced to C–C at 284.8 eV. The spot size was <sup>650</sup> × <sup>650</sup> <sup>μ</sup>m2 and the pass energy was 50 eV for the survey and 20 eV for the high resolution scan. The elemental composition was estimated from the survey spectra using Avantage ver. 5.938 software. The measurements on sample Q8 after atmospheric exposure were carried out using an Axis Supra instrument (Kratos Analytical, Manchester, UK) under slightly different conditions; spot size <sup>700</sup> × <sup>300</sup> <sup>μ</sup>m2 and the pass energy was 120 eV for the survey and 20 eV for the high resolution scan. The analysis was done by CasaXPS software ver. 2.3.19PR1.0. The high-resolution spectra were fitted using XPSPEAK software (ver. 4.1) with Shirley and/or linear background.

Grazing incidence X-ray diffractometry (GIXRD) was used to measure the crystalline properties of the CuCl. The measurements were taken with a SmartLab diffractometer (Rigaku, Tokyo, Japan) at grazing incidence angle α = 0.2◦ using a copper X-ray tube (1.542 Å).

Normal-incidence reflectance was measured in the Vis–UV range using an Avantes 2048 pixel spectrometer (Avantes BV, Apeldoorn, The Netherlands; spectral range from 1.9 to 5.1 eV), equipped with a fibre reflectance probe [23]. The angles of incidence covered the range of 0.2◦–1.7◦ and the nearly circular spot had a measured diameter of 0.8 mm for the 400 μm detection fibre. The small measurement spot allowed us to test the in-plane homogeneity of the samples. Bare substrates were used to collect reference signals.

The photoluminescence (PL) data were obtained with using a LabRam HR Evolution (HORIBA Scientific, Kyoto, Japan) 0.8 m focal length single-stage spectrometer with CW HeCd laser excitation at 325 nm (3.8 eV). The mirror lens collected signals from approximately a 1 μm circular spot; the excitation power was kept low enough to prevent heating of the layers (~200 W cm<sup>−</sup>2). The PL spectra were recorded with a Peltier-cooled back-illuminated UV-sensitive CCD detector Synapse 2048 × 512. The instrument spectral range sensitivity has been corrected by the Intensity Correction System (HORIBA Scientific, Kyoto, Japan) delivered with the system. All reflectance and PL measurements were performed at room temperature (300 K).
