Summary. Borate-crosslinked guar or hydroxypropyl guar (HPG) polymer solutions have become increasingly popular as hydraulic fracturing fluids. These fluids are cheap and environmentally friendly, and they minimally impair a propped fracture while yielding maximum viscosity. The drawbacks, which have limited their use, are a restricted temperature range of applicability, relatively high tubing friction, and poor stability when prepared with seawater. This paper shows how these drawbacks can be eliminated by a fundamental understanding of the relation between fluid chemistry (as a function of borate crosslinker, pH, and polymer/crosslinker concentration) and its physical properties (proppant-carrying capacity, viscoelasticity, and the temperature stability of the resulting crosslinked structure). Introduction Borate-crosslinked HPG and guar solutions have become increasingly popular as hydraulic fracturing fluids in well-stimulation operations. Guar fluids are clean, compatible with resin-coated proppant, and in many cases self-breaking. The one major hindrance to their widespread application has been a perceived limited temperature range. Viscous properties and proppant-carrying capacity depend on the number and strength of crosslink bonds, which are controlled by the chemical equilibrium of the borate-fluid systems. This, in turn, is influenced by temperature and pH. Optimum hydraulic fracturing treatments require the borate fluid to have stable physical properties (e.g., viscosity). During a hydraulic fracturing treatment, the fluid is transported down the tubing into the formation. The subsequent temperature rise will alter the chemical equilibrium, changing the pH, the number of crosslink bonds, and therefore the fluid viscosity. This is of great concern because, if fluid viscosity becomes too low, the proppant settling rate may increase sufficiently to cause undesirable proppant distribution over the fracture (e.g., only the fracture bottom may be propped and communication with the perforated interval is lost). The decrease in pH with temperature increases also is a function of the composition of the water used to prepare the base gel. The preferred operational technique for massive hydraulic fracturing (MHF) treatments offshore is to use a liquid gel concentrate mixed on the fly with filtered seawater. A major incompatibility arises with the use of seawater because it contains multivalent metal ions. These may precipitate as hydroxides, which reduces the fluid pH and drastically affects fluid properties. This paper discusses the relation between the chemical equilibrium of the borate/polymer complex and gel viscoelastic properties. We describe a methodology to optimize these properties and present experimental data on the viscoelastic properties of borate-crosslinked gels. Steady-shear rheological measurements were carried out to investigate the borate/polymer complexes, as well as oscillatory shear and static proppant-settling measurements. The Appendix gives the underlying chemistry of the borate/HPG crosslinking. The concept described in the Appendix is combined with the experimental data developed in the following sections to furnish a fundamental understanding of the system. Experimental Equipment and Instrumentation. The linear viscoelastic properties of borate-crosslinked gels were studied with a controlled-stress rheometer, Type CS50, built by Carri-Med Ltd., England. The measurements were performed with a cone-plate geometry; the cone had a 6-cm diameter and 1.5 angle. During these tests, the fluids were subjected to an oscillatory shear with a small amplitude, from which the complex shear modulus, can be derived. is composed of an elastic component, the storage modulus, and a viscous component, the loss modulus, . The oscillatory shear measurements were executed at 0.8 Hz, an optimum frequency that allows both moduli to be observed. These measurements were combined with static proppant-settling tests performed in 7.5 × 30-cm glass cylinders to investigate whether static proppant settling is a function of the strength and density of crosslink bonds. These measurements were performed at a proppant concentration of 10 vol% 20/40-mesh Ottawa sand. This proppant concentration was chosen because the maximum settling rate is observed at this value. At lower values, proppant clustering is prevalent; at higher concentrations, hindered settling becomes important. The values reported here use the formation rate of clear fluid at the proppant bank top. The proppant-settling measurements are split into three regimes: nearly perfect proppant suspension (settling velocity, <0.5 cm/min), marginally acceptable proppant suspension (0. 5 5 cm/min), and poor or unacceptable suspension capacity (>5 cm/min). We have not yet investigated dynamic proppant-settling, although we expect that the applied shear influences proppant settling rates in viscoelastic fluids. Steady-shear rheological measurements of crosslinked gels are difficult in conventional rotational viscometers. Borate fluids crosslink rapidly and form viscous gels, so the fluid will not remain in the viscometric gap (the Weissenberg effect). Further, owing to the dynamic nature of the crosslinking process (which demands some time for equilibration), a long-pipe viscometer is required to measure the flow curve. Neither of these conventional rheometers is ideal. We used the helical screw viscometer (HSV), a practical instrument that can "characterize" borate fluid rheological properties. The HSV contains a helical screw impeller rotating in a draft tube, as Fig.1 shows (it is based on Kemblowski's original). The impeller keeps the fluid in motion continuously and prevents it from climbing out of the measuring gap. The impeller shears the fluid at a known rotational speed while its torque is measured. The shear and temperature regime can be adjusted easily. The HSV is quick and easy to operate. A pH meter, Type pH-196, with an E-56 glass electrode containing 3 mol KCl + AgCl electrolyte supplied by WTW G. H., was used to monitor the pH changes as a function of temperature changes and chemical additions to the borate-fluid sample. The pH meter was calibrated at room temperature with two buffer solutions at pH's of 7.00 and 10.00, respectively. The pH measurements at elevated temperature are compensated for by measurements made with an accurate NTC temperature sensor in the fluid. The HSV also can measure the rheology of fracturing-fluid slurries containing up to 50 vol% proppant particles because it has a wide measuring gap compared with the proppant particle diameter, and it keeps the fluid in motion continuously to prevent particle settling. Fluid and Test Conditions. We used HPG for the above measurements. We did not test guar in this study, although we expect only minor deviations from the results reported here. SPEPF P. 165^