Patina Formation and Aesthetic Durability of Architectural Copper and Copper Alloys in the Marine–Desert Environment of Dubai

Abstract

The use of copper and its alloys in architecture, especially in arid regions, is growing, driven by visual appeal, functional advantages, and sustainability. Changes in visual and colorimetric appearances and patina formation were evaluated for architectural Cu metal, brass (CuZn15), bronze (CuSn4), and a golden alloy (CuZn5Al5). Coupons were exposed over 4 years in Dubai, United Arab Emirates, at a test site located 2 km from the seashore under unsheltered conditions, and at various surface inclinations. Comparative exposures were conducted in Brest, France, at sites of increasing distance from the seashore. Visual appearance was assessed by colorimetry and optical imaging; patina cross-sections were characterized by means of scanning electron microscopy and elemental analysis (SEM/EDS), and crystalline phase identification was conducted by means of x-ray diffraction (XRD). All Dubai surfaces developed red-yellowish, heterogeneous patinas with embedded sand and dust, reducing lightness and visual appeal. Inclination had minor effect, although some extent of spallation occurred on downward-facing CuSn4. Even the corrosion-resistant CuZn5Al5 alloy lost its golden hue due to the incorporation of sand and dust into the patina. In Brest, appearance depended on the distance from the seashore, with green-blue patinas near the sea and red-yellowish farther inland, similar to Dubai. Cleaning may restore some luster, but the desert exposure generally reduced the long-term aesthetic performance of all materials.

1. Introduction

Copper and copper-based alloys are extensively used in architecture, cultural heritage, and engineering applications due to their durability, workability, and unique ability to develop visually distinctive patinated surfaces during atmospheric exposure. The formation of a patina, a complex layer of corrosion products that evolves naturally on exposed surfaces, is central to both the long-term performance and the aesthetic integration of copper materials into the built society. Patinas contribute to corrosion resistance by limiting further oxidation while simultaneously imparting the characteristic visual appearance ranging from brown and black oxides in the early stages to the well-known bluish-green hues of mature sulfate- and/or chloride-rich patinas [1,2,3,4,5,6].

The type of patina formation, its chemical composition, and visual expression are highly sensitive to the prevailing environmental conditions. In marine atmospheres, the combined action of high humidity, sea salt deposition, and cyclic wetting–drying processes promotes the formation of oxides (Cu₂O and CuO), as well as chloride-containing corrosion products such as atacamite/paratacamite (Cu₂(OH)₃Cl) [1,3,4,7]. These chloride-rich patinas, while often heterogeneous and porous, give rise to characteristic blue-greenish appearances that are aesthetically valued in coastal architecture [1]. In contrast, desert climates are generally marked by lower humidity, elevated temperatures, and frequent dust deposition, conditions that may limit the corrosion rate and constrain patina development to thinner patinas. The presence of sulfur-containing pollutants such as SO₂ and, e.g., (NH₄)₂SO₄, promotes the additional formation of basic copper sulfates such as posjnakite (Cu₄(SO₄)(OH)₆xH₂O) and brochantite (Cu₄(SO₄)(OH)₆), giving the patina a more greenish appearance [1,2,4,8,9,10].

Surface inclination further modulates patina formation by controlling water retention and runoff. Horizontal or upward-facing surfaces accumulate moisture and airborne particles, promoting thicker, more heterogeneous patinas with mottled coloration [3,11,12]. Conversely, vertical and inclined surfaces enhance drainage and drying, typically resulting in thinner, more uniform films [7,13]. Even within the same structure, microclimatic variations such as differences in orientation toward prevailing wind directions or sunlight can generate considerable diversity in both patina chemistry and visual appearance [7,13].

Another critical factor influencing the long-term behavior of copper patinas is their adhesion to the underlying metal. While relatively compact and protective patinas such as those rich in brochantite generally adhere well and provide durable protection, certain corrosion products, particularly chloride-rich phases like atacamite and paratacamite, may develop poor bonding to the substrate. This can lead to flaking or spallation of the patina under cyclic wetting and drying conditions, thermal expansion, or mechanical stresses [14,15]. Spallation not only exposes fresh metal to the environment, accelerating localized corrosion, but also introduces irregularities in the patina which change the visual appearance, producing patchy surfaces and uneven color development. In marine exposures, this phenomenon is more frequently observed due to the hygroscopic and expansive nature of chloride-containing corrosion products. In arid desert climates, abrasive dust and thermal cycling may still destabilize loosely adherent surface patina constituents.

While numerous studies have examined patina composition and atmospheric corrosion mechanisms of copper and copper-based alloys under diverse climatic conditions (e.g., both SO₂ and NaCl-rich environments), less attention has been paid to the influence of surface inclination and patina adhesion stability on the simultaneous evolution of protective function and aesthetic appearance in a marine–desert environment and similarities to marine conditions. This knowledge gap is particularly relevant in architectural and cultural heritage contexts, where both the durability and the visual integration of copper surfaces, such as copper-clad facades, are critical design considerations which become more and more popular.

These knowledge gaps are addressed in this study through systematic comparisons of the aesthetic evolution and spallation behavior of patinas formed on architectural Cu metal and Cu alloys (brass, bronze, and a golden alloy) exposed in a marine–desert environment, Dubai, United Arab Emirates, relative to marine conditions, Brest, France, at sites located progressively farther from the seashore. The aims of this study are the following: (i) to present changes in surface appearance, patina composition, and characteristics of copper metal (Cu metal) and copper alloy (CuZn15, CuSn4, CuZn5Al5) patinas formed at unsheltered exposure conditions (45° facing south) in a marine–desert atmosphere in Dubai compared to marine conditions with increasing distance from the seashore, i.e., decreasing influence of chloride deposition in Brest, France, up to 4 years; (ii) to assess the influence of surface inclination (inclined 45°, vertical 90°, and upside-down −180°) from an aesthetic uniformity and corrosion product spallation/flaking perspective; and (iii) to provide insights that bridge corrosion science and architectural practice.

2. Materials and Methods

Coupons, sized 5 × 10 cm, of commercial Cu sheet metal (Cu), brass (CuZn15–85 wt.% Cu, 15 wt.% Zn), bronze (CuSn4–96 wt.% Cu, 4 wt.%), and the golden alloy (CuZn5Al5–89wt.% Cu, 5 wt.% Zn, 5 wt.% Al, ≈1 wt.% Sn) were mounted on Plexiglas fixtures and freely exposed at three different inclinations (45°, 90°, and −180° from the horizontal) facing south at a marine–desert site in Dubai (in the Jebel Ali free zone), United Arab Emirates, located 2 km from the seashore for 1, 2, 3, and 4 years (2015–2019), following the ISO 8565 standard [16]. Exposures were also conducted for the same materials at four sites (45°, 90°-sites 1–4 (color measurements performed during November 2009–November 2013 [14]), and up-side down (−180°) at sites 1, 2, and 4 (2015–2019)) of increasing distance from the seashore, i.e., decreasing mean annual deposition rates of chlorides, in Brest, France. Site 1 is located at the military harbor <5 m from the seashore, not in the direct splash zone; site 2 is located 20–30 m from the seashore at St. Anne; site 3 is located in St Pierre, approximately 1.5 km from the seashore; and site 4 in Langonnet, approximately 40 km from the seashore. Previous measurements of mean monthly chloride deposition rates show 6–7 times lower rates at site 2 compared to site 1, and sites 3 and 4 are characterized by 30–45 and 40–70 times lower rates, respectively. More details are given in [14].

According to the available corrosivity and environmental data for 2015–2019 from Institut de la corrosion, Brest, providing the test sites, both Dubai and sites 1 and 2 in Brest were classified as corrosivity class C4 (high) with similar reported one-year corrosion rates of Cu metal varying between 16 and 26 g/m², year for site 1, between 15 and 21 g/m², year for site 2 in Brest, and 15 and 24 g/m², year in Dubai (data for an inclination of 45°, facing south) during 2015–2019, despite ≈ 10 times lower average annual mean monthly chloride deposition rates in Dubai (30–90 mg/m², day) compared to Brest (site 1 ≈ 1000 mg/m², day; site 2 ≈ 2100 mg/m²,day) for the given time period [17]. Both sites were characterized by extensive solar radiation (total annual sun irradiation of ≈4000 MJ/m² in Brest and 7000 MJ/m² in Dubai), whereas the annual rainfall quantity was largely different (Dubai: <25 mm/year, Brest: 900–1000 mm/year). The mean monthly temperatures were higher in Dubai (≈28 °C) compared to Brest (≈13 °C), whereas the mean monthly relative humidity levels were lower in Dubai (61%) compared to Brest (82%) for the exposure period.

Images of the Dubai test site and field exposure racks are presented in Figure 1. Details of the Brest exposure sites are presented elsewhere [14].

Surface appearance was evaluated by colorimetry (Minolta CM2500D spectrophotometer, Osaka, Japan) under D65 illumination in the CIELab color space, collecting ten measurements (covering the entire coupon) for each material, time of exposure, and surface inclination. Each color is defined by three coordinates: L* (lightness, from black to white), a* (chromatic axis from green to red), and b* (chromatic axis from blue to yellow). The overall color change (ΔE) is calculated as the distance between two points in this space, according to the CIE76 definition ΔE = √((ΔL*)² + (Δa*)² + (Δb*)²).

Scanning electron microscopy and energy dispersive spectroscopy, SEM/EDS analyses were performed to characterize the patina morphology and elemental composition of cross-sections using a Zeiss Supra 55 (equivalent) instrument (Carl Zeiss, Oberkochen, Germany), coupled with an Oxford Instrument X-Max 50 mm² SDD detector (Oxford Instruments plc, Abingdon, UK). Cross-sections of the exposed surfaces were prepared by embedding samples in conductive polymer, followed by polishing with 0.25 µm diamond to obtain a mirror-like surface. Crystalline phases within the patina were identified using an X’Pert PRO PANalytical x-ray diffractometer (Malvern Panalytical, Almelo, The Netherlands) equipped with an X-ray mirror and CuKα radiation.

3. Results and Discussion

3.1. Copper Alloy Patinas Under Marine–Desert (Dubai) and Marine (Brest) Exposure: Visual Appearance and Characteristics

The visual appearance and corresponding colorimetric characteristics of Cu metal and the three Cu-based alloys (CuZn15, CuSn4, and CuZn5Al5) exposed for 1, 2, 3, and 4 years in Dubai are presented in Figure 2 for surfaces inclined 45, 90, and −180° from the horizontal, facing south.

All materials and inclinations revealed a colorimetric appearance that ended up in the yellow-reddish quadrant (a*,b*) with considerably reduced lightness (L*). The reduced lightness was predominantly governed by the presence of sand firmly attached onto all exposed surfaces. The golden alloy, CuZn5Al5 (Figure 2d), showed a similar appearance as Cu metal (Figure 2a), CuZn15 (Figure 2b), and CuSn4 (Figure 2c), even though being slightly more yellowish for the surfaces exposed 45° from the horizontal. The yellowish appearance could be connected to the extensive presence of sand. All coupons exposed up-side down (−180°) revealed, independent on material, the largest change in lightness, compared to coupons exposed 45° and 90°. The lack of, or minor contribution of the blue and greenish components, illustrates the low levels of chloride deposition (annual average of 30–90 mg/m², day) and a relatively long distance of the test site from the seashore (2 km). The total color difference (ΔE) between the unexposed and the exposed surfaces was clearly visible, Figure 2e with ΔE values >> 3 defined as visual differences. Largest changes in appearance were observed for surfaces inclined 45° for CuZn15, closely followed by Cu metal ≈ CuSn4 >> CuZn5Al5. All surfaces inclined 90° and −180° showed small differences between the inclinations but larger changes in appearance compared to corresponding surfaces inclined 45°. Unexpectedly, the CuZn5Al5 inclined 90° and −180° showed a similar change in visual appearance as the Cu metal, CuZn15, and CuSn4 inclined 45°. This may be related to the presence of embedded sand and dust particles within the patina, as described below.

Aesthetic changes (and stereomicrograph images after 4 years), similar to those observed for Cu metal and the alloys in Dubai, have previously been reported by the authors for the same alloys when freely exposed at inclinations of 45° and 90° at sites located approximately 1.5 km (site 3) and 40 km (site 4) from the marine seashore of Brest, France [14]. The test sites located closer to the seashore (<5 m: site 1, and 20–30 m: site 2) of 10–100 times higher deposition rates of chlorides resulted in patinas of a more blue-greenish appearances for Cu metal, CuZn15, and CuSn4, whereas the color appearance of the CuZn5Al5 alloy remained in the red-yellow quadrant, Figure 3 [14,18]. The results illustrate the importance of the extent of chloride deposition and repeated dry and wet periods resulting in a patina of chloride-rich corrosion products on Cu metal, CuZn15, and CuSn4, giving them a blue-greenish appearance [1].

3.2. Field Exposure in the Marine–Desert Atmosphere of Dubai Compared to Marine Sites of Varying Chloride Deposition in Brest–Patina Formation and Composition

According to previous multianalytical characterizations of the patina formed at site 1, with the highest chloride deposition rate of the investigated marine sites [14,18], Cu metal developed a two-layered type patina with cuprite (Cu₂O) as the inner layer and atacamite (Cu₂(OH)₃Cl) as the outer layer, sometimes containing nantokite (CuCl). Alloy-specific products such as SnO₂ (Cu4Sn), Zn hydroxycarbonates (Cu15Zn), and Zn/Al oxides/hydroxides (CuZn5Al5) were also detected, with Zn- and Zn/Al-rich phases providing protective barrier effects. As a result, Cu15Zn and CuZn5Al5 showed greater resistance to chloride-induced corrosion than Cu metal and Cu4Sn. Over time, the outer layer compacted and adhered more strongly, reducing the corrosion rates. Spallation or flaking of corrosion products, mainly linked to CuCl-Cu₂(OH)₃Cl formation, was most severe for Cu metal and Cu4Sn, minor for Cu15Zn, and absent for CuZn5Al5. The patina thickness strongly influenced the visual appearance, with CuZn5Al5 retaining a lustrous golden finish after four years, while the other materials developed brown–red-bluish tones. This is attributed to the presence of Cu₂O, ZnO, Al₂O₃, and SnO₂ in the inner layer of the patina of the CuZn5Al5, along with the incorporation of SnO₂, Zn₅(CO₃)₂(OH)₆, Zn₆Al₂(OH)₁₆CO₃·4H₂O, and Zn₅(OH)₈Cl₂·H₂O in the outer patina layer, predominantly composed of Cu₂(OH)₃Cl, able to mitigate chloride-induced corrosion compared to pure Cu metal [15]. Similar patina compositions were observed at sites 2, 3, and 4, with the difference that the extent of chloride-containing corrosion products was reduced, though still present, with increasing distances from the seashore (i.e., reduced amounts of chloride deposition).

Analysis of corrosion products of the patinas formed on Cu metal, CuZn15, CuSn4, and CuZn5Al5 exposed at 45° from the horizontal in Dubai by means of SEM/EDS and XRD after 9 months and 1 year of exposure. Figure 4 revealed the presence of almost the same corrosion products as observed in the patinas formed at the marine site in Brest [15], as summarized above, though with considerably lower extent of chloride-rich corrosion products such as Cu₂(OH)₃Cl and CuCl.

Cross-sectional SEM images and compositional analysis of the patina on Cu metal after 1 year of exposure, Figure 5a, show a relatively thin (≈2–3 µm) two-layer patina with an inner layer of Cu–oxides with small amounts of Cl covered by a thin (<0.5 µm) outer layer containing chlorine (Cl) in addition to aluminum (Al) and silicon (Si), the latter most probably from deposited sand residues. With increasing exposure time, from 2 up to 4 years, the outer patina layer had evolved into varying thicknesses (12–30 µm) with elevated levels of Cl and sand/dust/construction-derived particles (Al, Mg, Fe, Ca, K, Si), embedded in the layer, contributing to its complex chemistry, heterogeneity, and varying thickness (Figure 5b–d).

The patina of the CuZn15 alloy was slightly thicker (≈3–4 µm) than observed for Cu metal after 1 year of exposure with an inner part of the patina composed of varying proportions of Zn- and Cu-rich constituents in striated layers and small amounts of oxygen (O) and Cl, Figure 6a. After 2–4 years, the patina thickness largely varied over the surface from typically 15 to 30 µm. The striped areas of Cu- and Zn-rich patina constituents formed an inner layer porous layer being covered by an outer highly heterogeneous layer also containing Zn- and Cu-rich corrosion products with embedded sand/dust particles rich in Ca, S, Na, Mg, Al, and Fe, Figure 6b–d.

Similar to previous findings at the marine sites in Brest, the bronze alloy, CuSn4, revealed after 1 year of exposure a relatively thick, ≈10–12 µm, and porous striated patina containing both Cu- and Sn-rich oxidized areas with small amounts of Cl integrated within the layers, as seen in Figure 7.

No Sn and slightly more Cl, combined with sand and construction particles, were present in the outermost top layer. Longer time periods, 2–4 years, as seen in Figure 7b–d, resulted in thicker but similarly striated oxidized Cu and Sn layers (up to ≈15–20 µm in thickness), with Cl integrated within the layer and a few µm thick top layers with embedded sand and dust particles. The presence of Cl within the layer is believed to be connected to the initial formation of CuCl, the corrosion product that readily converts to Cu₂(OH)₃Cl in the presence of moisture and oxygen, causing a volume expansion within the patina, which causes physical stress in the porous layer, which in turn leads to flaking/spallation of corrosion products [1,14,15]. From the XRD patterns, Figure 4, the presence of both CuCl and Cu₂(OH)₃Cl could be confirmed in the CuSn4 patina. Flaking/spallation has previously been reported to take place for both Cu metal and Cu4Sn exposed at near seashore marine conditions [14,15]. However, in the desert climate of Dubai, no visual flaking was observed on the CuSn4 samples exposed at 45°, which could be a result of the low deposition rates of chlorides. However, as discussed below, flaking was clearly taking place for the CuSn4 samples inclined upside-down (−180°). No evidence of either CuCl or flaking was observed for the Cu metal exposed in Dubai.

In agreement with previous observations for the CuZn5Al5 alloy at both urban and marine conditions, the patina formation was slower compared with Cu metal, CuZn15, and CuSn4 [8,18]. Cross-sectional investigations showed a two-layer patina with an inner layer with a thickness of ≈1.5 µm and an outer layer increasing in thickness with time to a total patina thickness of ≈5 µm after 4 years, respectively, as seen in Figure 8. The patina morphology with striated layers of Cu- and Zn-rich corrosion products observed for CuZn15 was also observed for the CuZn5Al5 alloy after 1 year of exposure including small amounts of Sn and Cl, as seen in Figure 8a. Sand and construction particles as well as Cl were mainly observed in the top layer. Different to CuZn15, the patina growth (corrosion rate) was considerably lower judged from the relatively thin patina after 4 years (≈5 µm compared to 15–30 µm), as seen in Figure 8c,d. This is connected to the presence of Al₂O₃, ZnO, and SnO₂ at the bulk/patina interface as described elsewhere [18,19] and the formation and intercalation of different Zn-, Al-, and Sn-rich corrosion products in addition to Cu-constituents acting as efficient barriers, thereby suppressing chloride-induced corrosion.

3.3. Effect of Surface Inclination on Patina Flaking/Spallation–Visual Observations

The literature findings show surface inclination to strongly influence patina formation on Cu metal and its alloys [20]. Upward-facing surfaces (e.g., 45°) in marine atmospheres generally experience longer wetting times and higher deposition of chlorides, aerosols, and dust, leading to thicker, more porous, chloride-rich outer patinas, whereas downward-facing surfaces develop thinner, denser, and more adherent films [21,22]. As already discussed above, spallation has been shown to be primarily driven by the formation of CuCl at the patina/metal interface, which transforms to Cu₂(OH)₃Clwith volume expansion, generating internal stresses [1,15]. These processes can explain the observation of spallation/flaking of corrosion products on Cu and Cu–Sn alloys at marine exposure conditions (Brest, France, near the seashore–site 2), in contrast to the more adherent patinas on Zn- and Al-containing alloys [14,15]. While multiple studies confirm the role of chloride deposition, patina heterogeneity, and environmental microclimates on spallation, field data directly comparing spallation as a function of inclination (e.g., 45° vs. upside-down) remain scarce.

The effect of chloride-driven spallation and surface inclination was therefore investigated during 3 years in the marine environment of Brest at three sites of largely different annual mean chloride deposition rates (sites 1, 2, and 4), as seen in Figure 9.

Like findings for surfaces inclined 45° [14], the extent of green-bluish patina formation was considerable for Cu metal and the alloys (CuZn15, CuSn4, and CuZn5Al5) also exposed up-side down (−180°) at site 1 located < 5 m from the seashore (not in the splash zone), with substantial flaking taking place on Cu metal, and even more pronounced on CuSn4, Figure 9a. No flaking was observed for the Zn-containing Cu alloys (CuZn15, CuZn5Al5). The underlying mechanisms for the lack of flaking/spallation have been reported to be related to the integration of corrosion products of Zn (Zn₅(CO₃)₂(OH)₆ and Zn₅(OH)₈Cl₂·H₂O) in the Cu-rich patina of CuZn15, and the additional presence of Cu₂O, ZnO, Al₂O₃, and SnO₂ in the inner layer of the patina of the CuZn5Al5 alloy and Zn- and Sn-rich corrosion products (SnO₂, Zn₅(CO₃)₂(OH)₆, Zn₆Al₂(OH)₁₆CO₃·4H₂O, and Zn₅(OH)₈Cl₂·H₂O in the outer patina) predominantly composed of Cu₂(OH)₃Cl, able to mitigate chloride-induced corrosion compared to pure Cu metal [15]. At site 2, 20–30 m from the seashore with 6–7 times lower deposition rates of chlorides than site 1, the effect of chloride-driven spallation was minor and not visually observed for any surface inclination (45°, −180°), as seen in Figure 9b. Similar findings were seen at site 4 with even considerably lower (40–70 times) mean annual chloride deposition rates than site 1, as seen in Figure 9c. The extent of flaking was in general more severe for surfaces inclined 90° compared to 45°, also consistent with differences in corrosion rates [14].

SEM/EDS cross-sectional analysis of the upside-down coupons (−180°) of Cu metal, CuZn15, CuSn4, and CuZn5Al5 is presented in Figure 10 after 2 and 3 years of exposure. Cu metal revealed a very thick patina with some sections being a two-layer type composed of a 25–30 µm thick inner layer mainly composed of Cu-oxides (Cu₂O) and small amounts of Cl, and a thick (≈40 µm) compact layer of chloride-rich Cu corrosion products (Cu₂(OH)₃Cl), Figure 10a,b. Other sections of the patina showed highly porous, thick (40–50 µm) layered patina structures rich in Cu-oxides alternating with Cu-, O, and Cl-rich corrosion products as well as sand and dust particles (Si, Al, Ca, Fe, Mg). The latter sections could most probably be easily flaked off.

The patina of the CuZn15 coupons was considerably generally thinner than the Cu metal patina consisting of sections with an inner, approximately 15 µm thick layer rich in mainly Cu and Zn-rich corrosion products and small amounts of Cl, and an outer layer (≈10 µm) consisting of Cu and Zn as well as larger amounts of Cl. Other sections showed a similar composition but was considerably thinner, ≈2 µm, Figure 10c,d. The CuSn4 patina was highly heterogeneous and largely varied in thickness from 5–6 µm to 50–60 µm, most probably because of flaked off corrosion products. The patina showed layered structures rich in Cu, Sn, O, and of varying amounts of Cl as well as embedded sand/dust-particles, Figure 10e,f. The CuZn5Al5 coupons showed a patina of Zn-, Cu-, Al-, and Sn-rich corrosion products being considerably thinner (2–3 µm) than the patinas on Cu metal, CuZn15, and CuSn4. Only very small amounts of Cl were observed in the patina, like the observations at 45°, showing the Zn- and Al-rich corrosion products able to hinder chloride-induced corrosion, explaining the thinner patina layers. Some sections of the patina revealed also for this alloy embedded sand/dust/construction particles, as seen in Figure 10g,h.

Exposure in Dubai for 4 years, a site characterized by considerably lower deposition rates of chlorides (30–45 times lower annual mean deposition rates) than sites 1 and 2 in Brest, did not result in any observed flaking/spallation of the Cu metal surfaces inclined either 45° or upside-down (−180°). As described above, this was attributed to no or minor amounts of CuCl within the patina, resulting in volume expansion during the formation of Cu₂(OH)₃Cl, whereas small amounts of both phases were observed for CuSn4, Figure 4. Nevertheless, no flaking was visually observed for surfaces inclined 45° from the horizontal, but clearly taking place on surfaces exposed up-side down (−180°), as seen in Figure 11.

The low deposition rates of chlorides measured at the Dubai site could not explain this flaking behavior as no flaking was observed for CuSn4 at the marine test sites of similar chloride deposition rates, i.e., sites 3 and 4. However, chlorine was evidently observed within sections of the upside-down patina of CuSn4 in Dubai, see Figure 10e, which possibly could be connected to CuCl and Cu₂(OH)₃Cl formation. The patinas were further locally considerably thicker, and the morphology different, Figure 10f, compared with the patina of coupons with a surface inclination of 45°, Figure 7. This could be an effect of condensation-forming droplets, local concentration of pollutants including Cl on the upside-down facing surfaces of the coupons due to large variations in daily temperatures which locally lead to longer wetting times with induced stresses, poor adhesion, and possible flaking of corrosion products. In combination with the incorporation of sand and construction particles into the patina, the patina may have become even more brittle. However, further analysis and investigations are required to properly understand these flaking processes. Like the findings at the marine sites, no flaking/spallation of corrosion products was observed for the Zn-containing alloys (CuZn15 and CuZn5Al5) for any surface inclination though being highly corroded, forming heterogeneous patinas with embedded sand and dust/construction particles. Changes in visual appearance of Cu metal, CuZn15, CuSn2, and CuZn5Al5 surfaces inclined 45° and upside-down (−180°) after 1, 2, 3, and 4 years of exposure are presented in Figures S1–S4 (Supporting Information).

4. Concluding Remarks

Coupons of architectural Cu metal, CuZn15, CuSn4, and CuZn5Al5, exposed 1.5–2 km from the seashore, up to 4 years in the marine–desert atmosphere in Dubai, United Arab Emirates, all, independent on material, time, and surface inclination (45°, 90°, and −180°), gained a red-yellowish appearance and heterogeneous patinas with embedded and well surface adherent sand and dust particles. Their presence resulted in a considerable reduction in lightness, providing all surfaces with a dull appearance. From an aesthetic perspective, this alters both the short- and long-term perceived color appearance uniformity and reduces the visual appeal of copper and copper alloy surfaces in marine–desert environments. The visual appearance of the same materials at the marine site in Brest was highly dependent on the distance from the seashore (i.e., chloride deposition rates) being green-bluish near the seashore and red-yellowish for the sites located farther from the seashore showing similar appearance as the marine–desert site. In agreement with findings for up-side down (−180°) samples exposed at the marine site 1 in Brest of high deposition rates of chlorides, exposures of the CuSn4 alloy in Dubai for 4 years resulted in evident spallation/flaking of corrosion products on the downward facing surfaces. Different from the findings at the marine site 1 of Brest, no flaking was observed on either Cu metal (no inclination) or on the CuSn4 surfaces inclined 45° or 90° in Dubai. Even though less pronounced as for the marine site 1 in Brest, next to the seashore, spallation clearly took place on the upside-down surfaces. This spallation is anticipated to become even more pronounced with prolonged exposure time, resulting in a highly patchy and dull surface appearance. As the surface aesthetics often is a key factor in design, materials that maintain higher reflectance and/or colorimetric stability under desert exposure conditions may therefore be preferred for long-term visual performance. Not even the CuZn5Al5 alloy showing a considerably thinner patina and no clear effect of chlorides due to the presence of Zn- and Al-rich corrosion products retained its golden appearance. This is primarily attributed to the incorporation of sand and dust particles into the patina. Nevertheless, the CuZn5Al5 alloy would be the best choice if a yellowish appearance is demanded. Upside-down surfaces should be avoided, if using Cu-Sn alloys for an architectural application. Cleaning strategies may possibly be a solution to maintain an even more lustrous appearance of the CuZn5Al5 alloy.