Introduction

Automotive conductive pastes are formulations consisting of an active metal component, a glass frit and an organic medium; the latter two facilitating the adhesion and application on to the target substrate, which is usually glass. Applications for these pastes include print-based in-car defogging systems, antennae and alarm circuits. The term ‘frit’ refers to ceramic mixtures that have been melted to form a glass and then crushed into a fine powder. Modern glass frits used in the silver pastes are lead-free to meet environmental regulations and are often predominantly bismuth-rich to effect good flow properties (1). Screen printing is the preferred mode of application for automotive pastes being cost-effective and scalable for industrial production.

Silver is the material of choice for automotive conductive paste due to its high electrical conductivity and oxidation resistance. In order to meet the industry defogger standards, the conductive tracks must operate with a 12 V line and the tracks need to have a resistivity of 6 mΩ sq^–1 at a thickness of 10 μm after being fired (2), which can only be met by using a noble metal such as silver. Typically, the silver loading in the paste is in the range of 60–88 wt% depending on the intended function. Table I summarises the automotive silver pastes produced by Johnson Matthey (3).

Although silver has excellent conductivity and stability, it is expensive. It is estimated that as much as 40% of the cost of the automotive silver paste is that of the silver content and between 2–3.4 g of silver is used per car for defogging purposes (2).

Any potential replacement for silver in automotive pastes should meet requirements of conductivity, solderability, adhesion, chemical resistance, optimal stability of resistivity over temperature and overall compatibility with automotive glass technology. Thrifting of silver with fillers is not uncommon in silver pastes in applications such as photovoltaic current collectors where fillers such as metals, silver coated metals and alloys have been attempted (4–6). However, stringent requirements for automotive silver pastes such as needing to pass environmental tests makes the choice of fillers rather limited. More than a decade ago, when the automotive industry considered the option of increasing the voltage in vehicles from 12 V to 24–42 V, some work was carried out to develop resistive silver pastes (to maintain the same power across the heating window) which involved thrifting silver pastes with base metals and their alloys (7). However, the industry abandoned the switch to high-voltage sources and no progress has been made in the area of thrifted automotive silver pastes since. Furthermore, in the past 10 years, the price of silver has decreased from its peak price of US$48.5 per troy ounce in April 2011 to US$17.9 per troy ounce in February 2017 (Figure 1) which has been a disincentive to the thrifting approach. However, there is still scope to further decrease the cost per kilogram of paste if a suitable replacement for some or all of the silver in automotive paste is found. This article provides an overview of the effects of reducing the silver content in 80% automotive silver paste using fillers which includes selected base metals, silver-coated base metals and base metal alloys.

Experimental

Silver pastes and thrifted silver pastes were fabricated with 80% metal loading by triple roll milling appropriate quantities of metal component, glass frit, additives and organic medium to achieve a viscosity suitable for screen printing. For the sake of comparison, the same bismuth(III) oxide (Bi₂O₃)-based glass frit was used in the present study the loading of which was typically 2–5%. Printed glass tiles were fired in a pre-heated fast-fire roller kiln, set to the target temperature, in air. Electrical resistivity of the tracks was measured by the Van der Pauw four-point probe method (8) on meander circuits. Colour measurements were carried out in a Datacolor spectrophotometer after calibrating using standard black trap and white tile. Temperature-programmed oxidation (TPO) studies were carried out in an Altamira AMI-200 instrument by heating ~200 mg of the powder to 600°C in 10% O₂/He gas.

Thrifting with Base Metals

A simple approach to thrifting silver is to replace with cheaper base metals. The closest conductivity match for silver (6.30 × 10⁷ S m^–1 at 20°C) is copper (5.96 × 10⁷ S m^–1 at 20°C) (9). However, copper suffers from poor oxidation resistance, with onset of oxidation around 140°C, making it an unlikely candidate for firing applications which involve temperatures in excess of 600°C. Other metals with a conductivity match close to silver are aluminium (3.50 × 10⁷ S m^–1 at 20°C) and zinc (1.69 × 10⁷ S m^–1 at 20°C). Zinc oxidises on firing to form zinc oxide (ZnO) whereas aluminium tends to either oxidise or alloy with silver (10), depending on the loading; all of which significantly increases the resistivity of the silver track. Although nickel is one of many carcinogenic metals known to be an environmental and occupational pollutant which requires extreme care during handling (11), it is a promising metal filler with conductivity of 1.43 × 10⁷ S m^–1 at 20°C and resists oxidation on firing. In order to test the effect of adding nickel filler to the silver paste, screen printable pastes were prepared by triple-roll mixing a mixture of appropriate quantities of silver, nickel (average particle size 10 μm, 8.4–32.4 wt% loading), glass frit, additives and organic medium. The paste was screen printed onto float glass substrates, dried around 100°C and subsequently fired in a roller kiln.

Figure 2 shows the variation of resistivity of 80% nickel-thrifted silver tracks fired at temperatures between 620–680°C in a fast-fire roller kiln. For comparison, resistivities of 80% Ag tracks fired at the same temperature are also plotted alongside. The resistivity of a pure silver track decreases with increasing firing temperature. For example, the resistivity of a silver track fired at 620°C was 2.6 μΩ cm which dropped to 1.9 μΩ cm after firing at 680°C, which is attributed to sintering of silver particles. As the nickel concentration in a silver track is increased, the resistivity of the track increases. The variation in resistivity (ΔR) with firing temperature between 620–680°C is between 3.8–13.7%, in the studied composition range. ΔR increases initially with increasing nickel concentration and then drops to lower values.

Apart from conductivity, addition of fillers will also have an impact on properties such as acid resistance and reverse colour. The acid resistance test involves evaluating the chemical resistance of silver tracks to withstand acid attack. This was carried out in 0.1 N sulfuric acid at 80°C followed by a tape test; the latter involving testing the adhesion of the tracks after the acid test using sticky tape. Acid resistances of silver pastes supplied by paste manufacturers vary significantly but most pastes pass a 2–4 h acid test. Acid resistance tests were carried out on silver tracks thrifted with 8 wt% nickel filler fired at 660°C. The resistances of the thrifted silver tracks were measured before and after the acid test as well as after tape tests and are summarised in Table II.

Acid resistance of nickel-thrifted silver tracks was poor with all tracks falling apart after the tape test. This is expected in the light of the fact that base metals vigorously react with acid which opens up channels for the acid to further penetrate into the track and attack the glass frit which binds the metal particles to the glass substrate, ultimately leading to delamination of the track during the tape test. Due to poor acid resistance, base metals are unlikely candidates for fillers in automotive silver pastes.

Thrifting with Silver Coated Fillers

A straightforward strategy to reduce the silver loading in silver paste is to use silver coated fillers such as silver coated copper and silver coated nickel, which in principle should meet the fine balance between conductivity and metal loading and may also offer better acid resistance due to the silver shell protecting the base metal core.

Investigations on the use of silver-coated copper powders as fillers for automotive silver pastes suggest that these Ag@Cu powders suffer the same fate as their pure copper counterparts (12). TPO studies carried out on two silver-coated copper powders with different silver loadings (10 wt% and 28 wt% silver) in 10% O₂/He ambient pointed that both samples oxidise on heating, irrespective of the silver loading, with two distinct oxidation steps involved (Figure 3). Increased silver loading on copper just increases the oxidation onset temperature. The onset of oxidation occurs between 200–300°C, which can be assigned to the formation of copper(I) oxide (Cu₂O) which on further heating oxidises to copper(II) oxide (CuO) between 375–425°C. The silver coating on copper delays the oxidation on heating by about 100°C however does not prevent it. This can be explained on the basis of non-uniform coverage of silver on copper and also due to the presence of pin holes in the silver shell which allows the penetration of oxygen, eventually resulting in the oxidation of copper (13). Oxidation of silver-coated copper powders at temperatures above 200°C makes them unsuitable for firing applications such as automotive pastes.

Silver-coated nickel (Ag@Ni) powders are also used extensively as filler material in inks and are claimed to have conductivity and chemical stability approaching pure silver (14). Hence, silver-coated nickel powder (40% Ag by weight) was tested as a filler for 80% silver paste. X-Ray diffraction (XRD) of 680°C fired silver prints containing Ag@Ni filler showed no evidence of nickel oxidation. Figure 4 shows the variation of resistivity as a function of firing temperature and Figure 5 shows the variation of resistivity as function of Ag@Ni loading in the silver paste. For comparison, the resistivity of 80% silver paste is also plotted alongside.

Resistivity of pure silver paste decreases with increasing firing temperature. Resistivity of silver tracks containing Ag@Ni also show similar behaviour, albeit linked to Ag@Ni loading (Figure 4). For example, at 8% Ag@Ni loading, the resistivity of a track fired at 620°C was 2.9 μΩ cm which dropped to 2.2 μΩ cm after firing at 680°C. However, at 32% Ag@Ni loading, the change of resistivity with increased firing temperature was almost negligible, remaining around 5 μΩ cm between 620–680°C. The resistivity of the Ag@Ni containing silver paste increases with increasing Ag@Ni loading, which is more pronounced at higher firing temperatures. Although the resistivity increases with increasing Ag@Ni loading, the latter promotes stable resistivity values over the typical firing range which is desired in automotive windscreen manufacturing.

Compatibility of thrifted silver pastes with automotive black enamel is critical as they are overprinted onto an enamel layer prior to firing and shaping of windscreen glass. Black enamels are made by pasting a mixture of black pigment and low melting glass (usually in the ratio of 1:4) and applied on to windscreens by either screen printing or inkjet printing techniques. Black pigments are usually derived from either copper chromite spinel, chrome iron nickel spinel or iron cobalt chromite spinel (15). The main purpose of the enamel layer is to protect the glue that holds the windscreen in place from degradation by ultraviolet (UV) light and also to hide the electrical connections. For enamel compatibility studies, thrifted silver pastes were printed onto the pre-printed and dried black enamel and fast-fired at 600°C, 620°C, 640°C, 660°C and 680°C.

The difference between two colour samples is often expressed as ΔE, which is the difference between the L*, a* and b* values of the reference and sample. Therefore, ΔE displays the difference between two samples as a single value for both colour and lightness. A detailed explanation of coordinates from Commission Internationale de l’Eclairage (CIE) Lab Colour model L*, a* and b* are explained elsewhere (16). Reverse colour values L*, a* and b* were measured on black enamel and black enamel under the silver print from which ΔE values were calculated using the formula (Equation (i)):

(i)

Table III shows the results of compatibility testing of 8 wt% Ag@Ni-containing silver track on a hiding black enamel 1T3015 manufactured by Johnson Matthey. The reverse colour data suggest that the Ag@Ni thrifted silver paste is compatible with 1T3015 enamel. Both ΔL and ΔE values are lowest around 660°C firing, which is similar to that observed for the pure silver paste. Although there is no general agreement on an ideal value; ΔE value of 2.3 has been proposed (17) as just noticeable difference (‘JND’), nevertheless, ΔE of 0.5 or less is preferred by windscreen manufacturers.

Acid resistance tests were carried out on 660°C fired, 8% Ag@Ni loaded silver tracks. Figure 6 shows the state of 8% Ag@Ni loaded silver tracks after 1 h, 2 h and 4 h acid test, respectively.

The resistance of the tracks was measured before and after acid tests as well as after tape tests (Table IV). Acid attack was evident in all tracks but 8 wt% Ag@Ni thrifted silver track passed the 1 h acid and tape test. This is an improvement compared to pure nickel fillers where delamination was observed in all samples exposed to acid.

Thrifting with Base Metal Alloys

Suitable cost-effective base metal alloys can also be used as fillers for silver paste as long as they are stable at high temperatures and have conductivity in the target range. Alloy powders including brass and stainless steel were tested in silver pastes. The XRD pattern of silver track containing 32 wt% brass filler (70:30 Cu:Zn), fired at 680°C is shown in Figure 7.

Major phases are silver and brass alloy along with minor amounts of CuO and ZnO, indicating that a small amount of brass alloy had been oxidised during fast-firing at 680°C. Typical resistivity of an 80% silver paste thrifted with 32 wt% brass filler and fired at 680°C was 12 μΩ cm. However, the variation in resistivity ΔR with increasing firing temperature (620–680°C) and increasing filler loading was >15% which could be related to the oxidation of brass filler, as evidenced from XRD. Base metal alloys that suffer from oxidation during firing will negatively impact the stability and compatibility of the silver paste with the other automotive glass components.

Even alloys that do not exhibit oxidation behaviour may still interact with either silver or glass components of the paste during firing, leading to undesired properties. XRD of silver tracks containing stainless steel (type 430-L) filler did not show evidence of oxidation on firing. Resistivity of 80% silver paste thrifted with 8–32 wt% stainless steel filler and fired at 680°C varied between 4–16 μΩ cm. ΔR with increasing firing temperature (620–680°C) was >15%. However, migration of silver into glass, commonly known as silver bleeding, was observed in fired silver tracks containing 430-L filler. Figure 8 shows the cross-sectional electron probe microanalysis (EPMA) image of silver track containing 24 wt% stainless steel powder fired at 640°C.

Both silver and stainless steel particles are homogeneously distributed in the track. The silver track is also less dense even after 640°C firing which could be due to poor packing of silver and stainless steel particles; d₉₀ of the latter being 25 μm. Glass frit distribution as seen from the silicon map is high at the interface between the glass substrate and the silver track which is expected due to glass melting and flowing under gravity. Overlap element maps point out that there is no evidence of detectable interaction between silver and stainless steel. On the other hand, silver bleeding into the glass substrate is evident. Silver bleeding occurs when silver ions are exchanged for sodium ions in glass (18), probably promoted by interaction with the filler. Stainless steel thrifted prints also showed poor acid resistance.

Summary

Attempts were made to reduce the silver content of the automotive silver paste by thrifting with selected base metals, base metal alloys and silver-coated base metals. Inevitably, the conductivity decreases with increasing filler loading which can be fine-tuned through the choice of fillers depending on the requirement. However, most fillers tested in this study show poor acid resistance which limits their use in automotive applications. The choice of the glass frit plays a significant part in improving the acid resistance which requires further research to fine tune the glass frit composition to the type of filler used, if the concept of thrifted silver pastes were to be a commercial success. Further considerations need to be made on the use of base metal fillers especially when the benefits of using them does not outweigh the cost.