Good morning, how are you doing?
If I had an option, I would rather fuse the sample and dissolve the glass disk in some mineral acid. This process removes the risk of HF.
Fusions are great dissolution processes and usually very complete and highly quantitative.
From the web site of http://www.claisse.com/borate-fusion-fluxes.php:
Fusion does not consist of heating the sample to its melting temperature, but rather having the oxidized samples dissolved into a solvent, generally a lithium borate flux. Therefore, the temperature required for preparing all samples should be slightly higher than the melting temperature of the flux but not exceed 1050 °C. Lithium metaborate (LiBO2, m.p. 850°C) reacts with acidic oxides, e.g. SiO2, and lithium tetraborate (Li2B4O7, m.p. 920°C) reacts with basic oxides, e.g. CaO. In practice, a flux is composed of various proportions of these two salts.
Choosing the correct fusion material will give best results.
On the ICP side - I prefer to use a cross flow nebuliser and Scott spray chamber. The cross flow nebuliser is typically constructed from inert materials like Pt-Ir and Sapphire, and the fluid path is fairly large in diameter (compared to pneumatic nebulisers). The Scott spray chamber has a very large internal volume and takes slightly longer to flush, but the large internal volume also reduces pulsation from the peristaltic pump.
On the operational side of the ICP, I would increase the coolant gas flow, as well as the auxiliary gas flow - the higher coolant gas will help to ensure the high sample load does not affect the sides of the torch too much, and ensures a slightly more stable plasma. The higher auxiliary gas flow ensures the plasma is lifted away from the injector tube and the high TDS does not cause a build-up at the injector tube's tip.
I would also add an internal standard. This is something of great debate, but a few simple rules - it must be stable in the solution, it must not be present (in significant concentrations) in the sample, and it must not cause interference on your analyte or be interfered with by the sample constituents.
Then lastly on ICP - do not rinse between samples - rather use the samples themselves to rinse out the previous samples. Rinsing with dilute acid reduces stability in the spray chamber (called the transient effect, look at a publication from J-L Todoli and J-M Mermet : "Effect of the spray chamber design on steady and transient acid interferences in ICPOES").