POTASH basicS Later, "potash" became the term widely applied to naturally occurring potassium salts and the commercial product derived from them. The main potash salts are sylvite, carnalite, langbeinite, and polyhalite, mixed in varying concentrations with halite (rock salt). The main use of potash is as fertilizer.
Sylvinite is the most important ore for the production of potash
in North America. It is a mechanical mixture of sylvite (KCl, or
potassium chloride) and halite (NaCl, or sodium chloride). Most
Canadian operations mine sylvinite with proportions of about 31% KCl
and 66% NaCl with the balance being insoluble clays, anhydrite, and
in some locations carnallite. Sylvinite ores are beneficiated by
flotation, dissolutionrecrystallization, "heavies" separations,
or combinations of
The major source of potash in the world is from the Devonian Prairie Evaporite Formation in Saskatchewan, which provides 11 million tons per year. Russia is second at 6.9 million and the USA (mostly from New Mexico) at 1.2 million tons per year. A dozen other countries in Europe, Middle East, and South America produce potash from evaporite deposits. Potash can be mined mechanically
by underground machinery or by solution mining using ambient or
warmed water. Halite (salt) for human use or road deicing
can be mined the same ways. Potash ores contain halite as well, so
the byproduct of potash extraction is road salt. In earlier times,
salt was more valuable per ounce than gold, as it was essential to
human life. A person "worth his salt" was one who contributed his
fair share to the community. The word potash is used in several different contexts in the literature. Here are some variations:
PROPERTIES OF POTASH
For consistency, potash ore and fertilizer concentrations are rated by their equivalent K2O content. Some literature can be confusing because they rate the ore by its potassium content (K) or potassium chloride content (KCl), The table below lists the physical properties of potash minerals, including K and K2O values. The GR (API units) entry in the table do not seem to match any known correlation, so some caution is urged.
POTASH MINERAL
PROPERTIES
Formula
NaCl KCl KMgCl_{3}●6H_{2}O
Clay K_{2}SO_{4}Mg_{2}(SO_{4})_{2}
K_{2}SO_{4}Mg(SO_{4})(Ca(SO_{4}))_{2}●2H_{2}O
Actual sonic travel time in halite and sylvite may be slightly higher than shown above due to occluded water. The Vp/Vs ratio for most salts is close to 1.9 so shear travel time is close to 1.9 times compressional travel time.
Since potassium is radioactive, the K2O content can be derived from gamma ray logs, and this technique has been used since the 1960's. In 1964, I was stationed in Lanigan, Saskatchewan to run logs in potash exploration wells. While there, I scrounged a personal tour of the Esterhazy potash mine, then only two years old. This was the first and only time I have seen geological structure and stratigraphy from the "inside" of the rock. Truly amazing!
No direct calibration between GR and K2O had been developed up to that time, so I convinced a client to let me see his core assay data. After adjusting for hole size, mud weight, and bed thickness, a reasonable relationship was found, and was published as "Quantitative Log Evaluation of the Prairie Evaporite Formation of Saskatchewan" by E. R. Crain and W. B. Anderson, Journal of Canadian Petroleum Technology, Quebec City and Edmonton, JulySeptember, 1966.
The work was subsequently reprinted in five other papers by various authors, some included updates as tool technology evolved. The original GR correlation was unchanged, widely distributed, and was the standard for potash analysis from oilfield style logs run prior to the era of digital logs in the 1980's. Most analog oil field GR logs were nonlinear above about 300 API units due to dead time in the counting circuit. These older logs are still available in the well files and were recently used by Saskatchewan Industry and Resources to update their potash isopach and ore grade maps.
GAMMA RAY BOREHOLE CORRECTIONS
Where: WM = mud weight (lb/gal)
POTASH ORE GRADE FROM GAMMA RAY
The slope in the above equation can be determined by correlation to core assay data for other hole sizes or other tool types.
The nonlinear relationship must be honoured while analyzing these older logs for potash. The effect is negligible for conventional oil field applications. Modern digital tools are linear up to about 1000 API units so the discussion in this Section does not apply.
A 1967 paper showed a linear GR relationship up to 650 API units for the McCullough tool, but its use was not widespread in Canada. That graph showed 600 API units was equivalent to 45% K2O, identical to my original data, but the slope of the line at lower GR readings was different. No mud weight correction was implied but a bed thickness correction similar to mine was presented.
In
the analog era, GR logs were calibrated to a secondary standard
based on the API GR test pit in Houston which contained an
artificial radioactive formation defined as 200 API units in an
8 inch borehole filled with 10 lb/gal mud.
NONOILFIELD GAMMA RAY TOOLS
USING ANCIENT NEUTRON LOGS To quantify the relative amounts of carnallite and sylvite, the neutron response must be converted to porosity from count rates using the standard semilogarithmic relationship. A typical transform for a 1960's era Schlumberger tool is shown at the left. Charts for other tools can be found in ancient service company chart books. With the advent of the sidewall neutron log in 1969 and later the compensated neutron log, this transform was no longer required.
USING SONIC AND DENSITY LOGS
The
minerals sought are halite (rock salt), sylvite, carnallite, and
insolubles or clay. The only logs available on old wells are
resistivity, sonic, neutron, and total gamma ray. The
resistivity is not a helpful discriminator, except as a shale
bed indicator, so it is not used in the simultaneous solution.
These evaporite beds contain potassium and ore grade is measured
in units of potassium oxide (K2O). K20 is obtained from a gamma
ray log, corrected for borehole size and mud weight, using a
nonlinear transform derived from core assay data. In middle
aged wells, the density log is also helpful, and in modern wells
the PE curve can be added. Further, the gamma ray response is
linear on modern wells so the transform to K2O is not as
difficult to obtain. K2O is obtained, after borehole correcting the GR, from the equations and lookup table shown earlier, or from a fresh correlation based on specific data from the wells under study. Note that the chart and table given earlier are in percent K2O and this set of equations expects fractional units for K2O, neutron porosity, and all output volumes. Parameters in the sonic equation are in usec/ft. When
solved by algebraic means, these equations become: These equations were derived with DELT in usec/ft. All constants will be different if DELT is in us/m. To
convert from mineral fraction to K2O equivalent (K2O equivalent
is the way potash ores are rated), the final analysis follows:
EFFECT OF OCCLUDED WATER Where Vwtr = PHIN value in pure salt above the zone of interest. The
occluded water has zero gamma ray emission so the second equation
remains unchanged: The
porosity is read directly by the neutron log, hence, the third
equation becomes: The
sonic equation becomes: Where C = DELT in salt minus 67 usec/ft. Reduction
of these equations results in: Conversion to K20 equivalent remains the same as before. Note that mineral fractions are in volume fractions. To convert to weight fraction, one more step is needed. By using the density of each mineral times the volume fraction, summing these to get the total rock weight, then dividing each individual weight by the rock weight, we get weight fraction of each. This allows comparison to core assay data which are reported in weight fraction or percent. The same math is used in tar sands and coal analysis to allow comparison to lab data.
12: WTclay = Vclay * 2.35 Note that the densities in the above equations are the true density values, not the elecctron densities used in the original simultaneous equations. Mass
fraction or weight percent values are obtained b dividing individual
weights by WTrock. foer example:
If the density or PE equation were added, then the set would be exactly determined and the strategy of finding Vwtr and C in the pure salt bed would not be needed. This work was done in Saskatchewan before density logs were common, so the density equation was not used at that time. With a modern suite of calibrated logs, we can use conventional multimineral models to calculate a potash assay. With GR, neutron, sonic, density, and PE, we can solve for halite, sylvite, carnallite, clay (insolubles or shale stringers), and water (occluded in many salts as isolated pores). The potassium curve from a spectral gamma ray log might also prove useful, if the detector system is linear and does not saturate. Alternate mineral mpdels are quite possib;e in other potash areas of the world.
The mathematical methods are covered in the Lithology Chapters elsewhere in this Handbook. Matrix rock properties for the minerals were shown earlier in this article. Water is treated as a "mineral" so that it can be segregated from the water of hydration in carnallite.
Probabilistic analysis methods are also used with modern log suites. Here, the mineral mixture can be underdetermined, allowing the progarm to find the best mix at any particulat depth point.
The first step is to correct the gamma ray for borehole and mud weight effects, using the appropriate service company correction charts. The other logs seldom need much correction as the potash is not especial deep or hot. However, if a water based mud was used, it will have a high salinity, so a salinity correction for the neutron log may be required.
The second step is to confirm the GR to K2O correlation using any available potash core assay data. Since modern GR logs are more linear than older tools, the relationship should be a relatively straight line and can be extended beyond the available core data, as shown at the right.
SPECIAL CASES 1. an incomplete open hole logging suite 2. logs run through casing 3. logs run with GR in counts per second 4. logs run where thin beds predominate 5. combinations of the above.
Incomplete Logging Suite Here we must include fewer minerals in the model. Isolated water is easy to ignore, and insoluble clay comes next, although it is an important economic factor in the extraction process. In the worst case, we might need to settle for K2O from the gamma ray and a sylvite / carnalite discriminator based on the neutron log. This situation occurs most often when potash geologists are using logs in wells drilled originally for oil or gas, in which potash evaluation was not considered as a priority.
Through Casing Logs The most obvious problem will be to correct the gamma ray log for casing size and weight, cement sheath thickness, and borehole fluid weight using service company correction charts. Where core assay data is available from the well or from reasonably close offsets, the GR to K2O relationship can be confirmed. The second problem is usually an incomplete logging suite, as described above. If a through casing neutron log is available, scaled or not, a carnallite flag can be created.
GR in Counts per Second
Thin Bed Problems
Combinations of the Above
Ancient GR logs could be rescaled with a nonlinear transform
to make them respond similarly to modern logs. Once the
conversion is made, computer analysis is easier and cross
sections look better.


Page Views  Since 01 Jan 2015
Copyright 1978  2018 E. R. Crain, P.Eng. All Rights Reserved 