SPE-191461-MS

Superior Transport Capabilities of Neutrally Buoyant Proppants in

Slickwater Fluids Deliver Step-Change Increase in the Conductive Fracture
Area of Unconventional Wells

Harold Brannon, Sun Specialty Products

Copyright 2018, Society of Petroleum Engineers

This paper was prepared for presentation at the 2018 SPE Annual Technical Conference and Exhibition held in Dallas, Texas, 24-26 September 2018.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Technological advancements have recently been directed toward development and optimization of
horizontal completions in unconventional reservoirs, with the ultimate objective of increasing asset
performance and value. Unconventional plays are being completed with ever-longer laterals, tighter stage
spacing, and high rate slickwater applications designed with increasingly larger volumes of sand to create
increased reservoir contact area for greater hydrocarbon recovery. Success is predicated upon overcoming
the limited transport capabilities of slickwater. The benefit of higher injection rates employed to enhance
proppant transport is soon lost as the lateral velocity declines exponentially with distance from the wellbore,
allowing the sand to fall rapidly to the bottom of fractures, resulting in propping only a fraction of the
created fracture area. While there are advantages to the use of slickwater and sand for unconventional
applications, the transport characteristics inherent to slickwater/sand slurries suggest significant limitations
to step-changes in hydrocarbon recovery.
Near-neutrally buoyant, ultra-lightweight proppant is a proven solution to make productive the otherwise
non-propped area. Several previous studies in parallel plate slot flow models have shown ULWP-1.05 is
transported well in slickwater, whereas sand settles rapidly to form a dune even at high flow rates. Such
behavior is intuitive given the near-neutrally buoyant ULWP has an Apparent Specific Gravity of 1.05, in
contrast to the 2.65 ASG of sand and the 1.0 ASG of water.
Two new proppant transport models have recently been introduced, including a slot with multiple fracture
branches and, a 3D complex network flow model designed to imitate flow through a lateral wellbore into
a complex fracture network. In both, the ULWP-1.05 was observed to be transported near-homogeneously
with the fluid to the extremities of the apparatus. Conversely, small mesh sand tended to stay in the lower
sections of the models and to deposit prior to reaching the extremities.
As a prelude to ULWP-1.05 field application in Permian Basin extended length horizontal wells, proppant
transport and fracture conductivity data for the near-neutrally buoyant ULWP-1.05 were used in fracture
models to optimize proppant placement for maximizing conductive fracture area, with iterations to optimize
well performance in production simulations. A desired outcome of this endeavor is the development and
validation of an optimized stimulation design exhibiting materially enhanced well performance.
2 SPE-191461-MS

This paper includes analyses and observations from the proppant transport testing, fracture conductivity
testing, discussion of the subsequent fracture designs and production simulations, and comparison of the
production simulations with production experienced in field applications. Performance of slickwater fracs
with sand alone and, with both sand and near neutrally buoyant ULWP are compared. Lessons learned may
be used to substantially increase the conductive fracture area of unconventional wells, optimizing production
performance and stimulated reservoir recovery efficiency.

Introduction
The development and optimization of horizontal completions and hydraulic fracturing of unconventional
reservoirs has been the target of many recent technological advancements. The ultimate objective of these
efforts is increasing production rate, EUR, and asset value using an approach of pumping massive volumes
of slickwater and sand at high rate into unconventional wells completed with multi-stage horizontal laterals.
Many unconventional plays are in the ‘performance optimization’ phase of development with increasingly
longer laterals, tighter stage spacing, and high rate slickwater applications employing increasingly larger
volumes of sand to create the increased reservoir contact area needed for greater hydrocarbon recovery.
Current practices for development of unconventional multi-stage horizontal wells are typified by 5,000 –
10,000 ft laterals segmented into 25 to over 50 stages, each fracced with 400,000 gallons of fluid and 400,000
pounds of proppant. Some operators are exploring methodologies to further optimize well performance and
value, including laterals approaching 20,000 ft in length with over 100 discrete fracturing stages, and steady
increases in the fluid volume and proppant intensity per foot of lateral.
The motivation for these efforts is enhancing asset value and return on capital investment, reasoning
that drilling longer laterals completed with more stages and treated with larger fluid volumes will contact
greater reservoir volume and, that increased proppant volume will render more of that created fracture area
conductive; thereby, yielding greater and more rapid hydrocarbon recovery per well.
Low viscosity fluids are preferred due to tendencies to generate more fracture complexity, created fracture
length, and created area in unconventional reservoirs. However, successful fracture stimulation is predicated
upon overcoming the limited proppant transport capabilities of slickwater. Quartz sand is predominantly
used as the ‘proppant’ for unconventional fracturing applications, and has a Specific Gravity of 2.65,
compared to water having an SG of 1.0. The specific gravity differential dictates that the heavier material
will fall or sink, thus the sand falls to the bottom of the fractures. The result is that only a minority of the
created fracture area is actually propped.1
Techniques including increased fluid viscosity, increased injection rate, and/or smaller diameter
proppants have been employed to address these phenomena with marginal success. Proppant settling is
inversely proportional to fluid viscosity, but increasing viscosity poses risk for greater damage to the
reservoir and fracture conductivity, and typically inflates fluid costs. Proppant settling rate is a function of
the square of the particle diameter, and consequently a steady progression to smaller mesh sizes of sand
to combat settling (ie. 20/40 to 40/70 mesh) has been experienced. A smaller particle size is more easily
transported and may yield marginally greater propped fracture area, but that benefit comes at the cost of
compromise in the resultant fracture width and therefore, conductivity, which is orders of magnitude lower
than the larger proppant sizes. The benefit of high injection rates employed to enhance proppant transport
for yielding enhanced propped fracture area is limiteddue to the exponential velocity decline with distance
from the wellbore. Reduced velocity allows the sand to fall rapidly to the bottom of fractures, resulting in
propping of only a fraction of the created fracture area. The transport characteristics inherent to slickwater/
sand slurries suggest significant limitations to achieving step-changes in hydrocarbon recovery.
Proppant transport testing and fracture modeling suggest that sand coverage is typically on the order of
25% of the created fracture area, leaving the balance of the created fracture expanse unpropped. Unpropped
fracture area is may be effectively non-conductive, particularly post-drawdown.
SPE-191461-MS 3

Difficulties achieving effective proppant placement across the created fracture area are exacerbated by the
wellbore injection location relative to the fracture height in horizontal wells versus vertical wells. Fracture
height in many unconventional horizons will often approach three hundred feet; for example, having an
upper barrier at a depth of 9,200 ft, and a lower barrier at 9,500 ft. Vertical wellbore perforations are typically
placed across the majority of the payzone height, perhaps from 9,250 - 9450 ft, providing for injection of
the proppant slurry across the majority of the created fracture height. A horizontal lateral in the same zone
would typically be landed at about 9,250 ft. Fracturing slurry injection from a horizontal well can occur only
from the lateral, which is located in the lower section of the pay and has an effective perforated height of
less than one foot (ie. the diameter of the lateral). Given the relative geometries and slurry flow mechanics,
one could expect that with slickwater/sand fracturing slurries, sand will only be placed in the lower environs
of the created area in horizontal wells.
Several products and processes have been introduced to address the challenges of proppant placement
in unconventional wells.23,24 Many of those new developments are based on Goldberg-esque solutions to
mitigate the proppant settling difficulties by rendering the sand more buoyant for improved suspension.
One promotes coating the sand with a polymer, which when exposed to the water swells to increase the
effective diameter of the particle, serving to reduce the relative apparent specific gravity and increase the
fluid viscosity such that it mimics the buoyancy of a lower specific gravity particle. Although perhaps
beneficial to proppant suspension, increasing the apparent fluid viscosity serves to reduce fracture network
complexity compared to low viscosity fluids. The polymeric gel coatings may be continuously stripped from
the sand core by turbulence and other downhole interactions during the pumping process, serving to negate
desired transport benefits, and collaterally, creating residual polymeric damage risk to the proppant pack
conductivity. Other solutions promote attachment of gas bubbles to adhesive-coated sand particle surfaces to
generate a dirigible effect for enhanced suspension. Accordingly, one might question the adhesive tenacity
of the bubbles to remain attached to the sand, and furthermore, the degree of buoyancy afforded by the gas
bubbles in the elevated hydrostatic pressure environment downhole. Boyle's Law would suggest not much
benefit at the several thousand psi within an unconventional reservoir.
Near-neutrally buoyant, ultra-lightweight proppant is perhaps the simplest, most technically and
operationally sound solution to mitigate proppant settling concerns and make productive the otherwise non-
propped area. Several previous studies in parallel plate slot flow models have shown ULWP is transported
well in slickwater, whereas sand settles rapidly to form a dune even at high flow rates. Such behavior is
intuitive given the near-neutrally buoyant ULWP has an Apparent Specific Gravity of 1.05, in contrast to
the 2.65 ASG of sand and the 1.0 ASG of water.
An ultra-lightweight proppant exhibiting a Specific Gravity of 1.05 was introduced in 2007 by Brannon.16
ULWP-1.05 has been successfully applied in over 1,000 wells and 40 formations across the United States,
including the Wolfcamp, Spraberry, Woodford, Niobrara, Marcellus, and Mancos. Several published field
studies discussed application of ULWP-1.05 in predominantly low to moderate BHST, vertical, and by
current standards, relatively short horizontal unconventional wells, including the Devonian and Barnett
shales.7-9,11,13 The application techniques have evolved over time from 100% replacement of the sand with
ULWP-1.05 to supplementing the existing slickwater/sand treatment schedule with 2% to 5% ULWP-1.05
added at various stages within a treatment. Long-term cumulative production improvement with both
strategies has typically been observed in the 30-50% range for both oil and gas producing wells, with many
exceeding 200% production uplift. The enormity of the material volumes employed in the contemporary
unconventional shale operations dictated that additional ULWP-1.05 strategy refinement be explored
to maximize the return on investment. Continued advances of the near neutrally buoyant ULWP-1.05
composition and manufacturing process have resulted in the current commercial ULWP-1.05 products
which are suitable for application in reservoirs having BHSTs exceeding 300°F and closure stresses greater
than 10,000 psi.
4 SPE-191461-MS

This paper characterizes the fundamental differences in the fluid flow mechanics involved in the hydraulic
fracturing of conventional vertical wells and horizontal unconventional wells. Further, the paper addresses
limitations of slickwater/sand systems for optimizing horizontal unconventional well stimulation and, offers
proven processes for using near-neutrally buoyant proppant as an effective solution. Proppant transport
has been characterized via three physical methods, including a conventional parallel slot model, a new
intersecting branch slot model, and, a large 3D model with a multiple branched pipe device designed to
represent a complex fracture. A commercial fracture design model, MShale, was employed to contrast
various treatment alternatives, and a sophisticated production simulation model, MProd was used to predict
and compare well performance resulting from the design model outcomes.

Statement of Theory and Definition

Optimizing effective fracture area is often a critical consideration for efficient economic exploitation of
reservoir assets. Accordingly, maximizing the propped fracture area is a key parameter for achieving desired
stimulated well performance. Propped fracture area within the productive payzone is proportional to well
performance. Proppant transport is an essential consideration to effectively place proppant in the pay zone.
Efforts to improve effective fracture area have historically focused on the transport capability of the fluid/
proppant system employed. The transport of proppant in fractures during a hydraulic fracturing operation
and the suspension of proppant during the fracture and closure process can significantly impact the propped
fracture area. Poor proppant transport leads to excessive proppant settling into the lower regions of the
created fracture, yielding relatively short effective fracture lengths and insufficient coverage of the height
of the productive interval. The result can be reduced stimulation efficiency and a steeper production decline
than desired. Post-frac production analyses frequently illustrate that the effective fracture area is less than
that expected based upon the design, suggesting the proppant was not effectively placed across the created
fracture area.1-4

Proppant Settling. Proppant settling can be assessed by comparing single-particle static settling velocity
in a standing column of water using Stoke's Law. As shown in Equation 1, the terminal velocity, νt, is a
function of Dp, the median proppant diameter, μfluid, the Newtonian viscosity of the fluid, and the difference
between the specific gravity of the proppant and the fluid, γprop and γfluid, respectively:
(Equation 1)
Increasing the fluid viscosity and/or decreasing the proppant diameter have traditionally been used to
facilitate improved proppant transport in conventional hydraulic fracturing applications. In each instance,
the modification to positively impact proppant transport has a corresponding negative effect on fracture
conductivity. Reducing proppant diameter explicitly reduces proppant pack permeability and fracture width,
and increasing viscosity brings a greater likelihood of residual fluid damage.
The single variable remaining for substantive improvement of proppant transport and placement is
specific gravity of the proppant, γprop. Assuming all other variables to be fixed or constant, Stoke's Law
teaches that lighter particles settle more slowly than heavier particles. Decreasing the specific gravity of the
proppant to approach that of the fracturing fluid is the most practical solution to improve proppant placement
with low viscosity fluids because as neutral buoyancy is approached, the proppant settling velocity nears
zero.
Low viscosity fluids with smaller-sized sands are most typically employed in unconventional reservoir
applications to maximize fracture complexity and reservoir contact. Figure 1A and Figure 1B show setting
rate comparisons for small diameter proppants in water and slickwater, 1 cP and 2 cP viscosity, respectively.
Figure 1A is illustrative of a classical Stoke's Law comparison with non-viscosified water as the fluid
medium, and settling rates reported in units of feet per minute. Smaller diameter proppants provide lower
SPE-191461-MS 5

settling rates as shown in Figure 1A, where reducing median particle diameter of sand from 30/50 to 40/70
to 100 mesh, results in declining settling rates, from 17.2 ft/min to 9.6 ft/min to 2.4 ft/min, respectively.

Figure 1A and Figure 1B—Static single particle settling rates

Unconventional fracturing treatments commonly use (1) slickwater with 1 gpt friction reducing agent
yielding a viscosity of +/- 2 cP, (2) have a proppant slurry pumping duration in excess of one hour, and (3)
after the pumping treatment ceases, are not subject to fracture closure for confinement of proppant to arrest
settling for up to several hours. Figure 1B provides a real-world context for comparison. As shown in Figure
1B, 40/70 mesh sand is observed to fall at a rate of 144 feet per hour, suggesting that 40/70 sand placed
with 2 cP slickwater will fall as far as the fracture width and orientation will allow. Replacement with 100
mesh sand reduces the unhindered settling rate to 36 feet per hour. Settling of smaller sizes are successively
less apt to be benefited by fracture roughness or orientation. Owing to it's near neutral buoyancy, the 30/80
ULW-1.05 proppant is observed to settle in a static environment at a rate of 2.4 feet per hour, which is
roughly 1/60th of the rate of the 40/70 mesh sand, ie. in the time the ULW-1.05 proppant settles 1 foot, a
similarly sized sand will have fallen 60 feet.10,12,14,16
Throughout the course of a hydraulic fracturing treatment pumping operation, the proppant slurry is in
a dynamic flow environment in which the lateral flow velocity serves to transport proppant some distance
laterally to create propped fracture length. The lateral flow velocity's degree of benefit is however, quite
often significantly over estimated. Although a slurry injected at the perforations may have an instantaneous
lateral velocity of tens of feet per minute at the wellbore, that velocity declines exponentially with distance
from the wellbore. Hydraulic fractures in horizontal unconventional wells typically grow in a radial fashion.
In a radial geometry fracture, the rate declines at a rate proportional to the radius, or fracture half-length.
Simultaneously, particle settling continues to occur due to the downward forces of gravitational effects.
Figure 2 illustrates the decline rate of the lateral flow velocity versus distance from the wellbore for slurry
injection from 2 foot perforation clusters in a horizontal wellbore at 10 bpm, which is representative of an
interval with perforation clusters, and 80 BPM frac rate. In this example, the instantaneous lateral velocity
is 0.82 ft/sec (44.5 ft/min). Within 2 minutes, the leading edge of the slurry will be 20 feet from the wellbore
and the velocity will have declined to < 10 ft/min. After 20 minutes, the leading edge will have traveled 90
feet from the wellbore, and the velocity will be < 2 ft/min. One hour after being injected, the leading edge
is shown to be 175 feet from the wellbore, moving at 1.2 ft/min.14,16
6 SPE-191461-MS

Figure 2—Fluid Velocity Decline vs. Distance from the Wellbore

Proppant placement is an important consideration for effective stimulation, thus to estimate proppant
placement in a fracture, the proppant transport efficiency should be known. An estimate of single particle
transport can be derived from combining the information from Figure 1 and Figure 2. For example, one can
estimate the path of a sand particle in water with a 44 ft/min injection velocity exiting the wellbore. After
10 minutes, the particle will have traveled 48 feet laterally and 96 feet downward, and if the particle has
not found the floor within 30 minutes time, it will have moved 120 feet laterally and 288 feet downward.
Multiply the above by thousands of gallons of sand and slickwater slurry, and one could envision a dune-
shaped pile of sand having an apex near the height of the lateral wellbore and sloping laterally at the
angle of repose to the base of the fracture. Unconventional reservoir fractures are most likely not smooth,
unbounded parallel plates as are the conditions represented in the laboratory testing and mathematical
modeling represented above, but the physics involved and reported data reasonably represent, if not closely
approximate, behavior in a real subterranean fracture.
Fluid velocity in a fracture declines rapidly with distance from the injection location and transport
capability diminishes accordingly. Hydraulic fractures tend to grow radially until encountering boundaries,
and thereafter transition to elliptical geometries with increasingly higher length to height ratio. Vertical
wellbores have perforations distributed over most of the payzone height, thus the growing fracture may
encounter boundaries relatively early in the process, causing the fractures to become increasingly more
elliptical. Conversely, in horizontal wellbores, the injection height is limited to the diameter of the lateral,
and fracture development is predominantly radial throughout the fracturing process.
Effective proppant placement across the created fracture area is fundamentally more challenging in
horizontal wells than vertical wells due to the wellbore injection location relative to the fracture geometry.
Fracture height in many unconventional horizons often approaches four hundred feet, for example, the upper
barrier could be at a depth of 9,200 ft, and the lower barrier at 9,600 ft. Vertical wellbore perforations can
be placed across the majority of the payzone height, perhaps from 9,300 – 9,500 ft, providing for injection
of the proppant slurry across the majority of the fracture height. A lateral wellbore in the same zone would
typically be landed at 9,500 ft, with an injection height of about 9,499 – 9,500 ft. Fracturing slurry injection
from a horizontal well occurs only from the lateral, having an effective perforated height of less than one
foot (ie. the diameter of the lateral), which is located in the lower section of the pay. Given these distinctions
of the geometry and slurry flow mechanics, it is reasonable to expect that with slickwater/sand fracturing
slurries that sand will only be placed in the lower environs of the created area, and incapable of placement
in the upper and far field portions of the created area.
Unpropped fracture area is presumed to be effectively non-conductive, particularly post-drawdown. It is
generally recognized that propped fracture area is proportional to performance.5,21 Fracture conductivity, Cf
SPE-191461-MS 7

is the flow capacity of a fracture, and in order to be effective, should be orders of magnitude greater than the
conductivity of the reservoir. Classical hydraulic fracture engineering teaches that fracture conductivity is
infinite when it exceeds ten times that of the reservoir conductivity. Unpropped fractures must experience
slip or shear creating unmatched or misaligned fracture facies to exhibit conductivity greater than the
reservoir once the fracture closes, but often decline back to the reservoir conductivity as closure stress
increases with drawdown.
Ultra-lightweight proppant has proven to be an effective solution for mitigating proppant placement
difficulties and making productive what would otherwise be non-propped fracture area. A near neutrally
buoyant ultra-lightweight proppant having an Apparent Specific Gravity of 1.05 (ULWP-1.05), was
introduced in 2007 by Brannon etal.16The ULWP-1.05 is technically buoyant in water-based fluid with a
density of 8.80 lb/gal. Typically, field water for fracturing application ranges from is fresh (SG 1.0) up to
an SG 1.02. The ULW-1.05 is homogeneously suspended in the fluid initially and settles slowly, ie. it does
not float in water or light brine.
Several studies in parallel plate slot flow models have shown ULWP is transported well in slickwater,
whereas sand settles rapidly to form a dune even at high flow rates.10,12,14,16 Other studies have shown
that the ULWP-1.05 placed in partial monolayer distributions exhibit fracture conductivities equivalent to
packed fractures having greater than 10 proppant layers.18-20 Thus ULWP-1.05 proppant partial monolayers
equate to a relatively high conductivity fracture having much reduced particle volumes distributed across
a larger area than can be achieved with sands. Several case studies have discussed successful application
in predominantly low to moderate BHST, multizone horizontal unconventional wells, including the Barnett
and Devonian shales.7,8,11,13 Continued research and refinement of the near neutrally buoyant proppant
composition and manufacturing process resulted in the current ULWP-1.05, which is suitable for application
in reservoirs having BHSTs exceeding 300°F and closure stress greater than 10,000 psi.
Fundamental differences in the fluid flow mechanics involved in the hydraulic fracturing of conventional
vertical wells and horizontal unconventional wells have been characterized. The conductivity of fractures
propped with low concentrations of small mesh-sized sands is presented, and contrasted with the
conductivity, or lack thereof, of unpropped fractures, and the implications on unconventional well
performance. Further, the paper addresses the limitations of slickwater/sand systems for optimizing
horizontal unconventional well stimulation, and lastly offers proven processes for using near-neutrally
buoyant proppant as an effective solution. Proppant transport has been characterized via three physical
methods, including a conventional parallel slot model, a branch slot model, and, a large 3D complex
fracture network model with a multiple branched pipe device designed to represent a complex fracture. A
fracture design model was employed to contrast various treatment alternatives, and the associated production
simulation model was used to predict and compare well performance resulting from the design model
outcomes.

Description and Application of Equipment and Processes

Proppant Transport – Parallel Plate Slot Model
Proppant transport studies were conducted using a parallel plate slot model located in the Well Construction
Technology Center ("WCTC") at the University of Oklahoma. The WCTC parallel plate slot model has
dimensions of 16 feet in length, and 22 inches in internal height. The slot width is variable for 0.25″ to 1″,
and for the reported studies was held constant at 0.25″. The studies comparing transport with variables of
fluid viscosity (fresh water, slick water, 10 cP linear gel) with various proppant types (sand, LW ceramic,
and ultra-lightweight) and sizes (8 mesh to 100 mesh) were reported by Brannon, etal, including details
on the apparatus and testing procedures. Observations from the previous testing are included herein for
comparisons to the current studies.10,12,14,16
8 SPE-191461-MS

New proppant transport models have been introduced, including a slot with multiple fracture branches
and, a 3D complex network flow model designed to imitate flow through a lateral wellbore into a complex
fracture network.21,25-27

Proppant Transport - Branched Slot Model

A branched slot model (BSM) is operated by Premier Oilfield Laboratories located in Houston, TX.25 The
model is configured with a primary parallel slot having four tangential secondary and tertiary slots to
simulate a fracture network consisting of a primary frac, secondary and tertiary fractures. The branched slot
model is illustrated in Figure 3. The primary slot fracture has a width of ¼ in, four tangential secondary
fractures have a width of 1/8 in, with 1/16 in tertiary fracture branches from one of the secondary fractures
(upper left quadrant in the schematic). The model has an internal height of 12 inches. One or more of the
secondary slots may be optionally closed to flow to facilitate analysis of different operational parameters.
The slurry injection rate can be varied from 0.5 – 2.5 gallons per minute, providing the capability to observe
flow at velocities up to 0.5 ft/min, resulting in shear rates up to approximately 1001/sec in the primary slot.
The primary objective of the proppant transport model is to visualize proppant settling in a dynamic state
based on shear rate, proppant density, proppant size, and viscosity and elasticity of the polymer.

Figure 3—Branched Slot Flow Model. Courtesy of Oilfield Laboratory

A series of tests was conducted in the Branched Slot Model (BSM) with 1 gallon per thousand of anionic
friction reducing agent at injection rates of 1 and 2.5 gallons per minute to assess the transport of 1 ppa
40/70 sand, 0.2 ppa 30/80 ULWP-1.05, and a mixture of 0.5 ppa 40/70 sand and 0.2 ppa 30/80 ULWP-1.05.

Proppant Transport - 3D Complex Fracture Network Model

The 3-D Complex Fracture Network model (CFNM) is operated by PropTester in Cypress, Texas.26,27 The
PropTester Proppant Transport Research Council is an ongoing joint industry effort to develop means
to physically view proppant placement in various fluid and flow conditions within simulated complex
fracture networks. The Council sponsored a testing program for simulating flow through a pipe system,
with multiple proppant sizes, specific gravities, angularities, and special properties, with fresh water as the
fluid at relatively high shear rates. The ULWP-1.05 was included in the 2017 Council testing program as
a "Specialty Proppant."
The apparatus consisted of pipe-flow through clear pipe from 0.25″ to 0.5″ ID, arranged to mimic a
complex fracture system. This consisted of a 0.5″ feed-pipe conveying water/proppant slurry to the middle
of a vertically oriented ‘backbone’ with 13 pseudo-horizontal pipe arms extending laterally away from the
backbone, both above and below the main feed-pipe. The horizontal arms each reflect directional changes
up and down, as well as splits, and are located at different distances from the initial entry point. Pipe size
diminishes away from the backbone. The terminus of each pipe contained a screen which stopped proppant
but allowed fluid to proceed.26,27,31
SPE-191461-MS 9

The standard testing procedure involved creating a slurry of water and proppant, then pumping the
resultant slurry throughout the complex fracture system, allowing momentum, rate, flow regime and gravity
to control proppant distribution and segregation. The slurry was mixed in a 150 gallon blender tank, and
pumped at 16 gpm (gallons per minute) with a low pressure centrifugal pump. Each of the different proppants
were tested in the apparatus three times to ensure repeatability. The test was terminated after all the slurry
was pumped or in some cases when the system pressured out due to pipes or screens packing off. The
procedure included multiple pressure transducers, flow meters and video cameras recording data throughout
the test. Immediately after the test was concluded, multiple valves were shut to prohibit any post-test
proppant movement, and the apparatus was disassembled. Each tube has a screen at the end to capture the
proppant for purposes of mapping the relative mass of each material transported through the respective
pipes. Each section was analyzed according to proppant weight, particle size and resultant conductivity, as
well as visual anomalies at different locations in the arms. A schematic drawing showing the CFNM flow
pathways and a photograph of the model are shown in Figure 4.

Figure 4—Complex Fracture Network Model. Courtesy of PropTester, Inc.

Fracture Conductivity
Long term fracture conductivity testing was conducted at PropTester as per industry standard procedures
defined in International Organization for Standardization ISO-13503-5 Part 5 "Procedures for measuring the
long-term conductivity of proppants" with modifications as follows. Standard baseline testing specifies 2
lb/ft2 proppant loadings run between Ohio Sandstone cores for a duration of 50 hours at each stress. In these
tests, proppant loadings were representative of contemporary unconventional well fracturing applications
using small diameter sand, and for the ULWP-1.05, in partial monolayer concentrations.28,29,31

Fracture Design and Production Simulation Modeling

The MFrac™ Design and Evaluation Simulator and MShale™ discrete fracture network simulator were
used for predicting fracture propagation and extent in fractured reservoirs. Reservoir property inputs were
representative of a Wolfcamp formation in the Permian Basin. The same reservoir model was used for
all MFrac and MShale runs. A baseline treatment pumping schedule was developed and used consistently
throughout the study with variations as shown in Table 1. Proppant transport and fracture conductivity data
within the model were updated to reflect the most recent test data for both the sands and the ULWP-1.05.
10 SPE-191461-MS

Table 1—Model Treatment Design Summary

Data from the MFrac design and evaluation simulator was exported to the MProd™ single-phase
analytical production simulator to assess the production for the variety of treatment scenarios, as provided
in Table 3. The MProd production simulations were generated for a 5,000 ft lateral with 25 stages. The
Model A baseline production results were reviewed by Permian operators of Wolfcamp wells completed
with 5,000 ft laterals segmented with twenty five 200 ft stages, each treated with 400 Mgal slick water and
400 Mlb 40/70 sand. The consensus of those operators was that the Model A baseline production results
are representative of their experience.

Presentation of Data and Results

Proppant Transport - Parallel Plate Slot Model
Proppant transport studies were conducted using the parallel plate slot model located in the Well
Construction Technology Center ("WCTC") at the University of Oklahoma comparing transport with
variables of fluid viscosity (fresh water, slick water, linear gel) with various proppant types (sand, LW
ceramic, and ultra-lightweight) and sizes (8 mesh to 100 mesh) were reported by Brannon, etal.10,12,14,16
The slot data from over seventy tests in the WCTC slot was processed to calculate the terminal velocity
(vt), below which a particle will no longer transport. Analyses of the processed data confirmed that Stoke's
Law for static settling applies trendwise, to proppant transport flowing through a parallel slot. In other
words, for a given fluid and proppant type, terminal velocity is directly proportional to the difference in the
square of the median proppant diameter; a 4X difference in a vt for 2X proppant difference in diameter, D2p.
As the specific gravity of the proppant approached that of the fluid, the vt neared zero; ie. the ULWP-1.05
had relatively negligible settling, regardless of the diameter.
A mathematical model has also been generated for calculating the minimum flow velocity required for
proppant transport using the physical properties of fracturing slurry material.16 The mathematical model was
developed using data acquired from approximately 75 tests in the Parallel Plate Slot Flow Model, several
fluid types and viscosities, with several proppant types and sizes. Using the physical property parameters
for 5 cP water-based fluid and 40/70 sand in the mathematical model, the minimum horizontal velocity for
slurry transport (MHVST) was calculated to be 0.33 ft/sec (20 ft/min). Figure 2 illustrates the velocity decline
for radial fracture development for a 10 BPM injection rate. The lateral velocity at the injection declines
from approximately 45 feet per minute at the wellbore injection location to 20 feet per minute within 10 feet,
SPE-191461-MS 11

and further to about 2 feet per minute at 100 feet. Once the fluid velocity falls below the MHVST, proppant
movement is limited bed-load transport, meaning further transport laterally will occur only by rolling over
the top of the settled proppant bed.

Proppant Transport - Branched Slot Model

A series of tests were conducted in the Branched Slot Model with 1 gallon per thousand of anionic friction
reducing agent at injection rates of 1 and 2.5 gallons per minute to assess the transport of 1 ppa 40/70 sand,
0.2 ppa 30/80 ULWP-1.05, and a mixture. Note, that 0.2 ppa ULWP-1.05 was used rather than a typical
application concentration of 0.05 ppa in order for it to be better observed.24, 30
Test BSM-1: 1 gpt Anionic FR, 1 ppa 40/70 sand, 1 gpm. The majority of the 40/70 sand settled in the
primary slot (~90%). The settled dune in the primary slot had a maximum height of 8″, occurring nearest
the inlet, a minimum height of 3″, occurring at the far end of the primary slot, and sloping in between.
Approximately 10% of the sand entered the tangential secondary slots and settled within 4″ of the junction.
No sand was observed in the tertiary slots, and very little sand was observed to transport sufficiently for it
to be discharged through the exit ports of the slot into the catch pan.
Test BSM-2: 1 gpt Anionic FR, 0.2 ppa 30/80 ULWP-1.05, 1 gpm. Suspension transport of the 30/80 mesh
ULWP-1.05 proppant was observed throughout all open slots, and visually appeared to be suspended at
about the same concentration in all slots. A thin layer (< ¼″) of settled ULWP-1.05 proppant was observed
in the primary slot, with very little settled in the secondary and tertiary slots (<1/8″). Visually, an estimated
70% of the ULWP-1.05 proppant was discharged out of the slot exit ports into the catch pan.
Test BSM-3: 1 gpt Anionic FR, 1 ppa 40/70 sand, 0.2 ppa 30/80 ULWP-1.05, 1 gpm. Separation of the
sand and ULWP-1.05 was near instantaneous upon injection into the primary slot. At the conclusion, <
1% entrainment of ULWP-1.05 in the sand dune was observed. The sand was distributed very similarly to
that observed in Test 1 with an estimated 90% settled in the primary slot. The remaining 10% of the sand
entered the tangential secondary slots and settled within 4″ of the junction. No sand was observed to enter
the tertiary slots. The ULWP-1.05 was observed to be distributed near homogeneously in all slots, including
the tertiary slots. A thin layer of ULWP-1.05, <1/8″, was observed on top of the sand dune in the primary
slot, and a few particles were observed entrained within the sand dune.
Test BSM-4: 1 gpt Anionic FR, 1 ppa 40/70 sand, 0.2 ppa ULWP-1.05, 2.5 gpm, all slots open. Significant
turbulence was observed in the first two feet of the primary slot, presumably due to the higher injection
rate. At the conclusion, the sand was, as observed previously, a dune with a height of 4″ across the entire
length of the primary and secondary slots. Very little sand was observed to have entered in the tertiary slots.
As observed in the previous tests, the ULWP-1.05 was distributed in relatively homogeneous suspension
throughout the slot apparatus. A few ULWP-1.05 particles were observed to be entrained within the sand,
which was attributed to the chaos observed near the injection ports.

Proppant Transport - 3D Complex Fracture Network Model

Test CFNM-1: 1 ppa 40/70 mesh sand. The 40/70 mesh sand was observed to initially flow exclusively
into the lines near the height of the lateral and below. Sand was not observed to enter into the upper lines
until after four minutes into pumping, which coincidentally was the same time flow into the lower level lines
slowed and then ceased due to sand settling and bridging to create a diversion effect. With the exception
of the lower-most line, the 40/70 mesh sand failed to effectively traverse changes in flow path direction
and created a bridge to block further flow. The post test location of the sand vertically was near evenly
distributed, which is at least partly due to the diversion from bottom to top occurring as the lower lines
bridged off. Approximately 45% of the sand was located in the upper half of the model and 55% in the
12 SPE-191461-MS

lower half. Near half of the 40/70 mesh sand which entered the model was caught in the terminal screens,
the other half settled out in pipes along the way.
Test CFNM-2: 1 ppa 100 mesh sand. The 100 mesh sand was observed to be transported from the lateral
line into most all segments within one minute after entry into the system, as diversion from bottom to top
was much reduced relative to observations the 40/70 mesh sand test. Once having exited the primary lateral
line and entering the secondary flow lines, the 100 mesh sand was observed to exhibit settling and duning
behavior. The 100 mesh sand distributed fairly evenly vertically, with 47% of the mass deposited in the
top half of the apparatus and 54% in the lower. The horizontal distribution was much less even, with 16%
of the 100 mesh sand transported to the terminal screens and 84% settled in the tubes between the central
line entrance and the ends.
Test CFNM-3: 0.2 ppa 30/80 ULWP-1.05. The ULWP-1.05 was observed to flow homogeneously
throughout the entire apparatus. The ULWP-1.05 was observed to flow with the fluid through every pathway.
The testing matrix included various sizes of sand, ceramic, and specialty proppants. It was reported that
ULWP-1.05 was the only proppant tested which was observed to flow throughout the upper most pipes to
the terminal screen. The mass of ULWP-1.05 which was captured at the terminal screens in each pipe were
measured and the distribution was relatively consistent. Very little ULWP-1.05 settled in the pipes prior to
reaching the terminal screens.
Test CFNM-4 – Mixture of 0.5 ppa 40/70 mesh sand and 0.2 ppa 30/80 mesh ULWP-1.05. Separation
of the 40/70 mesh and ULWP-1.05 was near instantaneous upon exiting the lateral line. A total mass of
1298 grams of sand recovered post-test, of which only 7% (90 grams) were transported to the screens. The
remaining 1208 grams were settled in the tubes, predominantly in the lower regions of the apparatus, as
shown in Figure 5A. The ULWP-1.05 masses measured in the lower-most tubes ranged from 12 to 102
grams each, with an average of 45 grams per tube, the majority of which was at the respective terminal
screens, as shown in Figures 5B and 5C. In the upper-most quadrant, the individual captured masses per
tube ranged from 26 to 33 grams, with an average of 29 grams each. 48% of the ULWP-1.05 went to the
upper half of the model, and 52% to the lower half.

Figure 5—Post-test photographs of proppant in the Complex Fracture Network Model. Courtesy PropTester,
Inc. 5A. 40/70 mesh sand settled in lower pipes. 5B. ULW-1.05 packed at an upper screen, and 5C. lower screen.

Fracture Conductivity Testing

Conductivity of Unpropped Unconventional Reservoir Fractures. Rtudies have reported the unpropped
conductivity of shale fractures to be less than 0.1 mDft at closure and declines with increasing stress upon
drawdown to 0.01-0.001 mDft.22 Propped fractures with 20/40 mesh sand at 1 lb/ft2 loading nitrogen gas
flow exhibited fracture conductivities ranging from 100 mD-ft at 1,000 psi, declining to 60 mdft at 5,000
psi.15 Studies using split Eagle Ford and Utica shale cores in a Hassler sleeve testing cell have been used ot
SPE-191461-MS 13

show conductivities exhibited by propped and unpropped fractures at closure stresses from 1,000 to 5,000
psi.22 The fracture conductivity of unpropped fractures using nitrogen gas for flow were observed to exhibit
fracture conductivity of 3 mD-ft at 1,000 psi closure stress, declining to 0.3 mDft at 5,000 psi. Lower
conductivities were measured using brine flow. Flowing a neutral pH brine with 1,000 psi closure stress,
the unpropped fracture was observed to have a conductivity of 1 mD ft, declining to < 0.1 mDft at 5,000
psi. These data are provided numerically in Table 2, and graphically illustrated in Figure 6. Previous studies
reported estimates for unpropped conductivity in tight gas and unconventional reservoir environments to
be of a similar magnitude as reported by Wu, etal.5,15,22

Table 2—Fracture Conductivity

Table 3—MFrac and MProd Data

14 SPE-191461-MS

Figure 6—Unpropped vs. Propped Fracture Conductivity.

Conductivity of Unconventional Reservoir Fractures with Low Small Mesh Sand Loading. The
proppants used for horizontal unconventional well stimulations today are of predominantly smaller size and
used at lower loadings than the customary 20/40 sand at 2 lb/ft2. Perhaps most common are 40/70 or 100
mesh sand placed at areal loadings of 0.25 lb/ft2, and often, lower. The data used in frac models for these
sizes and loading are typically extrapolated from proppant databases which have only limited measured
data at 0.5 lb/ft2 or greater. An effort was undertaken to acquire standardized fracture conductivity data for
40/70 mesh and 100 mesh sands at 0.25 lb/ft2 loadings. Tier 1 quality white sand was used for the fracture
conductivity testing. The data are shown in Table 2 and graphically in Figure 7. The 40/70 mesh sand at a
0.5 lb/ft2 loading was observed to exhibit a baseline conductivity of 69 mD-ft after 24 hours at 250°F and
6,000 psi closure stress. 100 mesh sand 0.25 lb/ft2 exhibited 35 mD-ft at similar conditions.29

Figure 7—Conductivity of small mesh sands and ULWP-1.05.

Conductivity of Fractures with Partial Monolayer Loadings of ULWP-1.05. As discussed in previous

papers, the ULWP-1.05 is best applied in partial proppant monolayers, the loading concentration for which
is dictated by the median particle diameter.31 The optimal concentration for the 30/80 mesh ULWP-1.05 is
0.01 lb/ft2. Long-term conductivities for the ULWP-1.05 at temperatures up to 275°F and closure stresses
to 8,000 psi are provided in Table 2 and graphically compared to the low loading sand conductivities at
similar conditions in Figure 7. The 0.01 lb/ft2 ULWP-1.05 exhibited a conductivity of 293 mD-ft at 250°F
and 6,000 psi closure stress.
SPE-191461-MS 15

Fracture Conductivity – Summary and Discussion

Proppant transport testing and fracture modeling suggest that the propped area in unconventional wells
treated with sand slickwater and typically on the order of 25% of the created fracture area, leaving the other
75% of the created fracture unpropped. Unpropped fracture area is effectively non-conductive, particularly
post-drawdown, thus not significantly contributive to ultimate stimulated well recovery. Previous studies
cited herein show unpropped fracture conductivity is in the range of three orders of magnitude lower than
the fracture conductivity of propped fractures in unconventional wells. Quantification of unpropped fracture
conductivity using split Eagle Ford core in a Hassler sleeve device by Wu, etal. reported it to be < 0.1 mDft
at 5,000 psi closure, and appearing to be trending to < 0.01 mDft as the closure stress increases to 6,000 psi.
The conductivities of smaller mesh sands at lower than ‘conventional’ loading has become very common
in unconventional well stimulation. While a 20/40 sand @ 2 lb/ft2 loading was common in the past,
unconventional horizontal well fracs more likely apply 40/70 mesh sand at 0.25 lb/ft2 loadings. Published
data using standardized testing procedures on small mesh sands at low loadings are rare. Data for a 0.5 lb/ft2
loading of 40/70 sand and a 0.25 lb/ft2 loading of 100 mesh sand are provided in Table 2. Comparison of the
low loading/small mesh sand data to the unpropped conductivity data confirms the "3 orders of magnitude
lower than propped fractures" guidance of the previous studies.
The optimum conductivity using ULWP-1.05 is achieved with partial monolayer concentrations, which
for 30/80 mesh ULWP-1.05 is achieved with a 0.01 lb/ft2 loading. The ULWP-1.05 at a loading of 0.01 lb/ft2
exhibited fracture conductivity an order of magnitude greater than the 0.5 and 0.25 lb/ft2 loadings of 40/70
or 100 mesh sands, and, 3 to 4 orders of magnitude greater than an unpropped fracture.

Fracture Modeling and Production Simulation

Horizontal vs. Vertical Wellbore Completion. Fracture models were developed with MFrac™ to compare
fracture development in a vertical well versus that in a horizontal well to illustrate the proppant placement
challenges of horizontal wells versus verticals. Identical reservoir properties, treatment volumes, treatment
staging, and pumping metrics were used, as outlined in Table 1 for Model A and Model B. Model A,
shown in Figure 8, is representative of a typical slick water sand treatment employed in unconventional
horizontal well applications and was used throughout the production simulation studies as the baseline case
for comparison. Model A and Model B wells were configured with identical completions – 200 ft stage
spacing, 8 clusters per stage, six shots per cluster - other than one being vertical and the other horizontal.
The vertical well perforations in Model B (Figure 9) were distributed across the middle of the payzone
height, and the horizontal lateral in Model A was landed at 25% of the height above the lower boundary.

Figure 8—Frac Model A, Baseline Horizontal Wolfcamp, 400 Mgal Slickwater, 400 Mlb 40/70, 80 BPM
16 SPE-191461-MS

Figure 9—Frac Model B, Wolfcamp, Vertical Orientation, 400 Mgal Slickwater, 400 Mlb 40/70 sand, 80 BPM

Fracture models showing proppant placement with identical treatments in a vertical and a horizontal well
have significantly different propped fracture geometries, as shown in Figures 8 and 9. The same fluid and
sand volumes were pumped at the same rates in each, but the created fracture area for the vertical well frac
is 6% larger than the horizontal well. The propped fracture area of the vertical well is 32%, the propped
fracture height is 42% more than the horizontal, and the propped fracture length is 7% shorter. One could
surmise from this data that the vertical well frac will provide more access to the payzone height, and the
horizontal well frac more access to the hydrocarbons more distant laterally, or farfield from the well.
Figure 10 provides a comparison of the production simulations from the models of the vertical and the
horizontal wells with the same fracturing treatment. Intuitively, the 32% greater propped area would be
expected to favor the vertical well case. The one year cumulative production for the horizontal and vertical
wells were almost the same, at 255 Mbbl and 252 Mbbl, respectively, but the 10 year cumulative production
(EUR) favored the horizontal well by 9%, 612 Mbbl vs 539 Mbbl, perhaps suggesting that propped length
has more influence than propped height.

Figure 10—Production Simulations: Wolfcamp, Model A (Horizontal Baseline) and Model B (Vertical)

Effects of Injection Rate. The effects of reducing injection rate were evaluated, as shown in Figure 11
for Model C. Data inputs were identical to those in Model A except for the injection rate reduction from
the standard 80 bpm to 60 bpm for Model C. The injection rate reduction to 60 bpm resulted in a 25%
reduction in propped fracture area, and a 15% reduction in the propped to created fracture area ratio. The
SPE-191461-MS 17

slurry injection velocity reduced with the pump rate, which allowed for settling closer to the wellbore, and
the longer pumping time required allowed the settling to occur over a longer period. As a result, a 4%
reduction was observed in the one year cumulative production for the lower pump rate, and increased to
an 8% reduction in the EUR; 612 Mbbls in Model A pumped at 80 bpm, to 564 Mbbls for Model C, as
shown in Figure 13.

Figure 11—Frac Model C, Horizontal Wolfcamp, 400 Mgal Slickwater, 400 Mlb 40/70 sand, 60 BPM

Effects of Increased Fluid Viscosity. Model D was developed to evaluate the effects of fluid viscosity
increased from slick water with 2 cP to a 10 cP linear gel, as shown in Figure 12. Fluids more viscous than
slick water are often used in hybrid fracs to facilitate transport and placement of larger mesh size proppants.
All metrics were the same as Model A except for the increase in viscosity. The viscosity increase to 10 cP
resulted in a 12% reduction in propped fracture area, marginally greater propped height (+7%), and reduced
propped length (-19%) as compared to Model A. The 1 year cumulative production for the 10 cP model
was marginally lower than for the 2 cP model, at 249 Mbbl and 255 Mbbl, respectively. The delta increased
significantly for the 10-year cumulative EUR, as for Model D it was observed to be 493 Mbbl, or 20%
less than Model A.

Figure 12—Frac Model D, Horizontal Wolfcamp, 400 Mgal 10 cP linear gel, 400 Mlb 40/70 sand, 80 BPM
18 SPE-191461-MS

Figure 13—Production Simulation. Variables Impact. Model A (Baseline horizontal) vs Model

C (60 BPM), Model D (10 cP linear gel), Model E (30/50 sand), and Model F (100 mesh sand).

Effects of Proppant Size. The 40/70 mesh sand used in the Model A baseline case was replaced with similar
volumes of 30/50 mesh sand in Model E, and 100 mesh sand in Model F to evaluate the respective effects
on proppant placement and well performance. The Mfrac models for Model E and Model F are provided as
Figures 14 and 15, respectively. Both models used the same fluid volumes, pump rates, etc., and all yielded
the same created fracture area. Model E with the 30/50 mesh sand displayed the greatest propped fracture
area and highest fracture conductivity, which translated to a 290 Mbbl cumulative production after 1 year,
or 14% better than the 40/70 baseline, as shown in Figure 13. However, the EUR for the 30/50 case was
about the same as Model A, with 611 Mbbl compared to Model A with 612 Mbbl.

Figure 14—Frac Model E, Horizontal Wolfcamp, Slickwater, 400 Mlb 100 mesh sand, 80 BPM
SPE-191461-MS 19

Figure 15—Frac Model F, Horizontal Wolfcamp, 400 Mgal 400 Mgal Slickwater, 400 Mlb 30/50 sand, 80 BPM

Model F with 100 mesh sand in lieu of the 40/70 mesh baseline exhibited 3% greater propped fracture
area, and had a significantly lower average fracture conductivity than Model A (0.81 mDft vs 1.59 mDft).
The 1 year cumulative production for 100 mesh Model F was 210 Mbbl or 21% less than observed for the
40/70 baseline. However, over the life of the well the Model F production closed the early time gap, with an
EUR only 4% less than observed for the 40/70 mesh model, at 590 Mbbl vs. 612 Mbbl for the 40/70 baseline.
Effects of Sand Intensity. Proppant intensity refers to the pounds of sand pumped per lateral foot of
horizontal treated. The Model A baseline uses 400,000 lbs of sand to treat a 200 ft lateral stage, thus the
so-called proppant intensity is 2,000 lbs/ft. Many operators have sought to optimize the proppant intensity
as a means to optimize well performance. From discussions with a variety of Permian operators it appears
that a 2,000 lb/ft intensity is typical. Model G and Model H were generated to explore the effects of sand
intensity on the propped fracture area and well performance (Figures 16 and 17, and Figure 18).

Figure 16—Frac Model G, Horizontal Wolfcamp, 400 Mgal Slickwater, 320 Mlb 40/70 sand, 80 BPM
20 SPE-191461-MS

Figure 17—Frac Model H, Horizontal Wolfcamp, 400 Mgal Slickwater, 480 Mlb 40/70 sand, 80 BPM

Figure 18—Production Simulation. Effects of sand intensity (lb/ft). Frac Model A (Baseline, 400 Mlb,
40/70 mesh sand) vs. Model G (480 Mlb, 40/70 mesh sand) and Model H (320 Mlb, 40/70 mesh sand).

Model G included 20% more sand than Model A, thus a 2,400 lb/ft intensity. Model H used 20% less
proppant than Model A, ie. 1,600 lb/ft. The created area metrics for the Models A, G, and H were nearly
identical, as would be expected since all used the same fluid volumes and the same rates. The propped
fracture lengths were also identical. The propped fracture height and propped fracture area of Model G were
~10% greater than observed for than Model A, and the 1 year cumulative production was 6% greater (270
Mbbl compared to 255 Mbbl). The EUR for Model G with 20% more sand than the baseline was observed
to be 1% less than that of the Model A baseline. Model H, with 20% less sand than the baseline Model
A, exhibited a propped fracture height and propped area 9% less than Model A. The 1 year cumulative
production was 240 MMlb, or 6% less than the baseline. The EUR for Model H was 601 Mbbl, which is
about 2% less than the baseline which had 20% more sand.
Effects of Adding Ultra-lightweight Proppant, ULWP-1.05 to Slick Water Fracs. In Model I and Model
J, ultra-lightweight proppant was added to the slick water/sand slurry of baseline Model A at 3% and 4% by
weight of sand to assess the effects of improved proppant transport and propped fracture area metrics and
production improvement. The 3-4% bwop were selected to provide the desired partial monolayer proppant
concentration of the ULWP-1.05 across most of the created fracture area.
SPE-191461-MS 21

Model I included 16 Mlb of ULWP-1.05 distributed throughout the slurry stages with the 400 Mlb
baseline sand volume (4% by weight of proppant, bwop), as shown in Figure 19. The MFrac plot in Figure
21 showing the distribution of fracture conductivity readily illustrates the increase in propped fracture area
afforded by the inclusion of the ULWP-1.05. Compared to Model A, the propped fracture area increased
by 4.4X, and the propped to created fracture area increased to 84%, compared to the 19% observed for the
baseline model. The MProd production simulation data for all the models with ULWP-1.05 is provided as
Figure 20. The production simulation for Model I shows a 49% increase in the 1 year cumulative production,
from the 255 Mbbl projected for Model A, to 380 Mbbl for Model I with 4% bwop ULWP-1.05. The EUR
for Model I was 762 Mbbl, 25% higher than the Model A baseline.

Figure 19—Frac Model I, Hz Wolfcamp, 400 Mgal SW 400 Mlb 40/70 sand, 400 Mlb 40/70 sand, 16 Mlb ULWP-1.05, 80 BPM

Figure 20—Frac Model J, Hz Wolfcamp, 400 Mgal SW, 400 Mlb 40/70 sand, 12 Mlb ULWP-1.05, 80 BPM
22 SPE-191461-MS

Figure 21—Production Simulation. Assessing the impact of adding ULWP-1.05. Frac Model A (Baseline)
vs. Model I (16 Mlb ULWP-1.05 added (4% bwop)), Model J (12 Mlb ULWP-1.05 added (3% bwop), Model
K (400 Mgal, 360 Mlb sand, 12 Mlb ULWP-1.05), and Model L (360 Mgal, 320 Mlb sand, 12 ULWP-1.05).

Model J was similar to the previous Model I, except that the ULWP-1.05 addition was reduced to 12
Mlb, or 3% bwop. Model J's MFrac fracture conductivity graphic, Figure 20, looks very similar to that for
4% bwop in Model I. The propped fracture area was about 4% greater than observed in Model I, for the
4.8 times the propped area of the Model baseline without ULWP-1.05). The Model J propped to created
fracture area ratio is 87.9%, compared to 84.8% for Model J compared to the 19% observed for the baseline
model. The 1 year cumulative production for Model J was 365 Mbbl, or 4% less than Model I. The EUR
cumulative production for Model J is effectively the same as observed for Model I, at 763 Mbbl.
The 40/70 sand volume was reduced by 10% to 360 Mlb for Model K. Otherwise, Model K is identical
to Model J with 12 Mlb ULWP-1.05. The Model K results from MFrac and MProd, as shown in Table 3,
Figure 22, and Figure 21, were almost identical to the Model data. The only difference was the Average
Fracture Conductivity in Model K was about 1% lower than was observed for Model J. As configured, the
MFrac and Mprod show no ill effects from reducing the sand by 10%.

Figure 22—Frac Model K, Hz Wolfcamp, 400 Mgal SW, 360 Mlb 40/70 sand, 12 Mlb ULWP-1.05, 80 BPM.

The sand volume for Model L was reduced by 20% and the fluid volume was reduced 10% from the
amounts used in the Model A baseline, ie. to 360 Mgal fluid and 320 Mlb sand. As with Models J and K,
SPE-191461-MS 23

Model L included 12 Mlb of ULWP-1.05 distributed throughout the slurry stages. The graphic for Model is
shown in Figure 23. Due to the reduced fluid volume, the created fracture area was 5% lower than observed
in models I, J, and K. The propped fracture area was very similar to Model I; about 4% less than observed
in Models J and K. The propped to created area, at 88.4% was the highest observed for any of the models
and 4.6 times relative to the Model A observation. The one year cumulative production for Model L is 362
Mbbl which is slightly lower than observed the other runs with 12 Mlb ULW-1.05. The EUR cumulative
production for Model L is 739 Mbbl, or 3% less than Models I, J, or K, yet 21% greater than the non-
ULWP-1.05 containing baseline Model A.

Figure 23—Frac Model L, Hzl Wolfcamp, 360 Mgal SW, 320 Mlb 30/50 sand, 12 Mlb ULWP-1.05, 80 BPM

Discussion
New proppant transport models have recently been introduced, including a slot with multiple fracture
branches and, a 3D complex network flow model designed to imitate flow through a lateral wellbore into a
complex fracture network. Ultra-lightweight proppant with a near neutral specific gravity of 1.05 has been
shown via proppant transport testing as capable of overcoming the sand proppant placement difficulties
using slick water. Transport of near-neutrally buoyant ULWP-1.05 in low viscosity fluid is significantly
superior to 40/70 or 100 mesh sand. In all of the testing device configurations, sand was observed to settle
rapidly and dune within no more than a few feet of the entrance into the simulated fracture. ULWP-1.05
was transported near-homogeneously with the slick water throughout the created fracture area, a capability
suggesting that it can provide for much improved proppant placement and the creation of much greater
conductive fracture area. The near neutrally buoyant ULWP-1.05 was observed to be transported near-
homogeneously with the fluid to the extremities of the apparatus. Conversely, sand, even in small mesh sizes,
tended to distribute within the lower sections of the models and to deposit prior to reaching the extremities.
Based upon the observations from these tests, the transport capability of ULWP-1.05 in low viscosity fluids
is significantly superior to 40/70 or 100 mesh sand.
The poor placement of sand proppants with slickwater fluids throughout the created fractures is common
cause of unconventional horizontal well underperformance relative to potential. Multiple studies suggest
that 15% to 30% of the created fracture area is propped using sand/slickwater treatment systems. The
shortcomings of sand transport with slickwater in multizone horizontal well applications preclude step-
change improvement in hydrocarbon recovery. Conversely, the abilities of ULWP-1.05 to be transported
near-homogeneously with the slickwater throughout created fracture area provide for much improved
proppant placement and conductive fracture area.
24 SPE-191461-MS

The baseline treatment design Model A is representative of current materials, volumes, and field practices
used in fracturing Wolfcamp horizontal wells in the Permian Basin. The MFrac for this case showed the
propped fracture area was 19% of the created fracture, which is consistent with the previously reported
15-30% range. The MProd production simulation for Model A was vetted with actual Wolfcamp multi-zone
horizontal well production. The simulator data was fairly close to the acknowledged operator community
experience, lending more confidence to the models.
A series for fracture models were run to compare the effects on production expected from common
variable changes in the standard or baseline slick water and sand treatments. Decreased pump rate or use of
a 10 cP fracturing fluid in lieu of the 2 cP slick water both resulted in reduced production. Increasing the
sand size from 40/70 mesh to 30/50 mesh provided no discernible effect on the well performance. Reducing
the sand size to 100 mesh resulted in a marginal reduction in well performance.
A model to compare an increase in the sand intensity by 20%, to 2,400 lbs/ft from the baseline of 2,000
lbs/ft, was shown to provide a 6% production improvement at 1 year, but <1% at 10 years. Conversely,
reducing the sand intensity by 20%, to 1,600 lbs/ft, was shown to reduce the performance relative to the
2,000 lb/ft case by 6% after one year, but only 2% at 10 years. The data suggest users should closely weigh
the value of increasing the sand intensity against the economic and time horizons driving those decisions.
Neutrally buoyant ULWP-1.05 complimenting the slickwater/sand treatment designs in the fracture
modeling provided step-change improvement in propped fracture area and simulated well performance.
The addition of 30/80 mesh ULWP-1.05 at 3 - 4% by weight of sand to the baseline treatment design
resulted in a four-fold increase in the propped fracture area and a nearly 50% increase in the first year
cumulative production, equating to an additional 125,000 barrels of oil for a 5,000 ft. lateral with 25 stages
after one year and an additional 150,000 barrels in the 10 years. The production uplift observed in the
models containing ULWP-1.05 was compared to case histories of wells using ULWP-1.05 in reservoirs
having similar characteristics but being vertical wells or horizontals having relatively short lateral lengths.
A cross-section of case histories from the hundreds of ULWP-1.05 applications in unconventional reservoirs
shows an average cumulative production uplift of over 40% for both oil and gas wells, for periods of 12
months to 5 years. For example, a Delaware Basin operator who was an ‘early adopter’ of the ULWP-1.05
technology/product in the Cherry Canyon oil wells has reported an average uplift of over 200% for over 100
wells completed during 2010 – 2015 compared to wells treated without ULWP-1.05 but similarly otherwise.
The simulations for the ULWP-1.05 models included in this study predict production uplift in the 30%
range lending confidence in the models as the production uplift predictions are conservative relative to field
experience, and encouragement that it suggests that additional optimization is possible.
Furthermore, designs including the ULWP-1.05 with reduced sand intensity and fluid volume suggest
the step-change well performance benefits derived from the ULWP-1.05 addition can be maintained while
managing total stimulation costs. Whereas the ULW-1.05 contributed over 65% additional propped fracture
area and a corresponding 40% production uplift, reducing the sand and fluid volumes by 20% resulted in
only 3% reduction in propped fracture area and a 6% reduction of the production associated with the sand
proppant. As reflected in Model L with 12 Mlb ULWP-1.05 and, a 25% reduction in sand volume, the one
year cumulative production was reduced < 1% compared to Model I without the sand reduction.

Conclusions
Poor placement of sand proppants in created fractures with slickwater fluids is a common cause of
unconventional well underperformance relative to potential. Only 15% to 30% of the created fracture area
is typically propped using sand/slickwater treatment systems.
Unpropped fractures are effectively non-productive fractures. At fracture closure, the conductivity of
unpropped unconventional reservoirs fractures is on the order of 10-1 mD-ft, and after drawdown, 10-2 mD-
ft or lower.
SPE-191461-MS 25

The areal concentration of sand in horizontal unconventional well fractures is typically 0.1 – 0.5 lb/ft2.
Long-term reference conductivity data generated by use of standard testing procedures are provided for
the 40/70 mesh sand at 0.5 lb/ft2 and 100 mesh sand at 0.25 lb/ft2. The conductivites of the low loading,
small mesh sized sands was observed to be two orders of magnitude greater than measured for unpropped
fractures.
The conductivity of 30/80 ULWP-1.05 at a partial monolayer concentration of 0.01 lb/ft2 is an order of
magnitude greater than the 40/70 mesh or 100 mesh sands, and greater than four orders of magnitude greater
than the unpropped fracture conductivity.
Proppant settling in low viscosity fluids prior to fracture closure or proppant confinement can be
significantly contributive to lower than desired propped fracture height or area, particularly in environments
requiring several hours for closure. Using a 2 cP fluid, 40/70 sand falls at 144 ft/hr. In contrast, the similarly
sized 30/80 ULWP-1.05 settles at an acceptable rate of 2.4 ft/hr.
Modeling of slickwater fracturing treatments with sand proppant in an unconventional horizontal well
illustrated about 25% of the created fracture is propped and productive. Neither increased fluid viscosity or
decreased pump rate were observed to produce improved proppant placement or well performance.
Increase in proppant size to 30/50 mesh provided the greatest benefit in the first year showing cumulative
production of +14%. However, no benefit was observed in comparison with the 40/70 mesh baseline at 10
years. Decreasing the proppant size to 100 mesh resulted in a 21% reduction in the one year cumulative
production, but only a 3.5% reduction was observed at 10 years.
An evaluation of the effects of the sand intensity suggests caution for consideration of the cost/
performance efficiencies. A 20% increase in sand injected per lateral foot to 2,400 lb/ft resulted in a year one
cumulative production uplift of only about 6%, declining to a 1% uplift at 10 years. Conversely, a reduction
of 20% in the sand intensity resulted in a 6% reduction in the first year and -1% at 10 years.
Near-neutrally buoyant ultra-lightweight proppant can be utilized to overcome the sand proppant
placement difficulties using slick water. Transport of near-neutrally buoyant ULWP-1.05 in low viscosity
fluid is significantly superior to 40/70 or 100 mesh sand. In all model configurations, sand was observed
to settle rapidly and dune near of entry into the simulated fracture. ULWP-1.05 was transported near-
homogeneously throughout the created fracture area to deliver much improved proppant placement and
conductive fracture area.
ULWP-1.05 added to contemporary slickwater/sand treatment designs has been shown to provide
step-change improvement in propped fracture area and production performance of multistage horizontal
unconventional wells. The addition of 3 - 4% 30/80 mesh ULWP-1.05 resulted in a four-fold increase in
propped fracture area and a 50% increase in the cumulative production. Furthermore, designs including
the ULWP-1.05 with significantly reduced sand intensity and fluid volume suggest that the incremental
production benefits derived from the ULWP-1.05 addition can be realized while managing total stimulation
costs.
Incorporation of ULWP-1.05 into hydraulic fracturing treatments can deliver a step-change increase in the
conductive fracture area of unconventional wells, and thereby, unconventional well production performance.

Acknowledgements
The author acknowledges and expresses appreciation to Sun Specialty Products, Inc for their permission and
support for this paper, to Bazan Consulting for their fracture modeling and production simulation efforts,
and to the contributors at PropTester Inc and Premier Oilfield Laboratories for the proppant testing.

Nomenclature
ASG = Apparent Specific Gravity
Bbls. = barrels
26 SPE-191461-MS

BHST = bottom hole static temperature (°F)

BPM = barrels per minute
bwop = by weight of proppant
cP = centipoise, fluid viscosity
Cu. Ft. = cubic feet
D2prop = proppant diameter squared
°F = Fahrenheit
frac = fracture
ft = feet
ft/min = feet per minute
gals. = gallons
gpm = gallons per minute
gpt = gallons per thousand gallons
kf = permeability
kfw = fracture conductivity
lbs. = pounds
lb./ft2 = pounds per square feet
mD-ft = millidarcy feet
Mbbl = thousands of barrels
Mbo = thousand barrels of oil
Mgal = thousand gallons
MHVST = minimum horizontal velocity for slurry transport
Mlb = thousand pounds
ppa = pounds per gallon added
ppg = pounds per gallon
psi = pounds per square inch
q = flow rate (gal/min)
sec = seconds
spf = shots per foot
SG = Specific Gravity
ULWP-1.05 = ultra-lightweight proppant, SG = 1.05
V = average fluid velocity (ft/sec)
w = fracture width
ρ = fluid density (lb/gal)
μ = fluid viscosity (cP)
γprop = SG of proppant
γfluid = SG of fluid
vt = terminal velocity

References
1. Economides and Martin; "Modern Fracturing, 1st Edition," 2007.
2. Darin, S.R. and Huitt, J.L.: "Effect of Partial Monolayer of Propping Agent on Fracture Flow
Capacity", Paper 6813, presented at the Annual Fall Technical Conference of the SPE, Dallas,
October 4-7, 1959.
3. Novotny, E.J.: "Proppant Transport", Paper 6813, presented at the Annual Fall Technical
Conference of the SPE, Denver, October 9-12, 1977.
SPE-191461-MS 27

4. Biot, M.A. and Medlin, W.L.: "Sand Transport in Thin Fluids," paper SPE-1291G, presented at
the 1985 SPE Annual Technical Conference and Exhibition, Las Vegas, September 22-25.
5. Fredd, C.N., McConnell, S.B., and England, K.L.: "Experimental Study of Fracture Conductivity
for Water-fracturing and Conventional Fracturing Applications" paper SPE-74138, presented at
the 2001 Annual Technical Conference and Exhibition, New Orleans, September 30 – October 3.
6. Rickards, A.R., Brannon, H.D., Wood, W.D., and Stephenson, C.J.: "High Strength, Ultra-
Lightweight Proppant Lends New Dimensions to Hydraulic Fracturing Applications," paper
SPE-84308, presented at the 2003 SPE Annual Technical Conference and Exhibition, Denver,
October 5-8.
7. Schein, G.W., Carr, P.D., Canaan, P.A., and Richey, R.: "Ultra Lightweight Proppants: Their
Use and Application in the Barnett Shale," paper SPE-90838, presented at the 2004 SPE Annual
Technical Conference and Exhibition, Houston, September 26-29.
8. Edgeman, R., Gertsner, M., Nelson, S., and Malone, M.: "Lightweight Proppant. A New
Innovation in Hydraulic Fracturing": paper presented at the 51st Annual Southwest Petroleum
Short Course, Lubbock, Texas, April 21-22, 2004.
9. Brannon, H.D., Malone, M.R., Rickards, A.R., Wood, W.D., Edgeman, J.R., and Bryant, J.L.:
"Maximizing Fracture Conductivity with Proppant Partial Monolayers: Theoretical Curiosity of
Highly Productive Reality" paper SPE-90698, presented at the SPE Annual Technical Conference
and Exhibition, Houston, September 26-29, 2004.
10. Brannon, H.D., Wood, W.D., and Wheeler, R.S.: "Large-Scale Laboratory Investigation of the
Effects of Proppant and Fracturing Fluid properties on Transport," paper SPE-98005, presented at
the 2005 SPE Formation Damage Symposium, Lafayette, Louisiana, U.S.A, February 15-17.
11. Kendrick, D.E. Puskar, M.P., Schlotterbeck, S.T.: "Ultralightweight Proppants: A Field Study in
the Big Sandy Field of Eastern Kentucky," paper SPE-98006, presented at the 2005 SPE Eastern
Regional Meeting, Morgantown, West Virginia, September 14-16.
12. Brannon, H.D., Wood, W.D., and Wheeler, R.S.: "The Quest for Improved Proppant Transport:
Investigation of the Effects of Proppant Slurry Component Properties on Transport" paper
SPE-95675, presented at the SPE Annual Technical Conference and Exhibition, Dallas, October
9-11, 2005.
13. Chambers, R, and Meise, K.: Comparison of Fracture Geometries Utilizing Ultralightweight
Proppants Provide Evidence That Partial Monolayers can Be Created: A Case History," paper
SPE 96181, presented at the 2005 SPE Annual Technical Conference and Exhibition, Dallas,
October 9-11.
14. Brannon, H.D., Wood, W.D., and Wheeler, R.S.: "Improved Understanding of Proppant Transport
Yields New Insight to the Design and Placement of Fracturing Treatments," paper SPE-102758,
presented at the 2006 SPE Annual Technical Conference and Exhibition, San Antonio, Texas,
September 24-27.
15. Palish, T., Duenkel, R., Bazan, L. Heidt, H., and Turk, G.: "Determining Realistic Fracture
Conductivity and Understanding Its Impact On Well–Theory and Field Examples," SPE-106301,
presented at the 2007 Hydraulic Fracturing Technology Conference, January 29-31, College
Station, TX.
16. Brannon, H.D., Wood, W.D., and Wheeler, R.S.: "A New Correlation for Relating the Physical
Properties of Fracturing Slurries to the Minimum Flow Velocity Required for Transport," paper
SPE-106312, presented at the 2007 SPE Hydraulic Fracturing Technology Conference, College
Station, Texas, U.S.A, Jan 29-31.
17. Warpinski, N.R., Mayerhofer, M.J., Vincent, M.C., Cipolla, C.L. Lolon, E.L.: "Stimulating
Unconventional Reservoir: Maximizing Network Growth while Optimizing Fracture
28 SPE-191461-MS

Conductivity," SPE-114173, presented at the SPE Unconventional Reservoirs Conference,

Keystone, CO, February 10-12, 2008.
18. Brannon, H.D., and Starks, T.R.: "The Impact of Effective Fracture Area and Conductivity on
Fracture Deliverability and Stimulation Value", paper SPE-116057, presented at the SPE Annual
Technical Conference and Exhibition, Denver, September 21-24, 2008.
19. Brannon, H.D., and Starks, T.R.: "Maximizing Return on Fracturing Investment by Using Ultra-
Lightweight Proppants to Optimize Effective Fracture Area: Can Less really Deliver More?",
paper SPE-119385, presented at the 2009 SPE Hydraulic Fracturing Technology Conference, The
Woodlands, Texas, U.S.A, Jan 19-21.
20. Brannon, H.D., Kendrick, D.E., Luckey, E., and Stipetich, A.: "Multistage Fracturing of
Horizontal Shale gas Wells Using >90% Foam Provides Improved Production," paper
SPE-124767, presented at the 2009 SPE Eastern Regional Meeting, Charleston, West Virginia,
September 23-25.
21. Alotabi, M.A. and Miskimins, J.L.: "Slickwater Proppant Transport in Hydraulic Fractures:
New Experimental Findings and Scalable Correlation," paper Spe-174828, presented at the SPE
Annual Technical Conference and Exhibition, Houston, TX, September 28-30, 2017.
22. Wu, W., Kakkar, P., Zhou, J., Russell, R., and Sharma, M.M.: "An Experimental Investigation
of the Conductivity of Unpropped Fractures in Shale," paper SPE-184858, presented at the 2017
SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, U.S.A, Jan 24-26.
23. Jackson, K. and Orekha, O.: "Low Density Proppant in Slickwater Applications Imp, roves
Reservoir Contact and Complexity: A Permian Basin Case History", paper SPE-184858,
presented SPE Liquids-Rich Basins Conference, Midland, Tx, September 13-14, 2017.
24. Radwan, A.: "Multifunctional Proppant: A Review of Over 30 Field Trials In Low Permeability
Formations," paper SPE-187329, presented SPE Annual Technical Conference and Exhibition,
San Antonio, TX, October 9-11, 2017.
25. Premier Oilfield Laboratories Technical Services Report TSR201707-193, "3D Slot Testing for
Sun Drilling." July 27, 2017.
26. PropTester Proppant Transport Research Council: "2017 Proppant Transport Research Council
Mid-Year Report," presented August 17, 2017, Houston, Texas.
27. PropTester Proppant Transport Research Council: "2017 Proppant Transport Research Council
Annual Report," presented January 25, 2018, The Woodlands, Texas.
28. ISO 13503-5: Petroleum and natural gas industries - Completion fluids and materials - Part 5:
Procedures for measuring the long-term conductivity of proppants. International Organization for
Standardization, July 31, 2006.
29. PropTester: "Conductivity Analysis, 100 mesh Tier 1 Sand," Report Number 400-18-01-63-03-D,
February 5, 2018.
30. Personal communication with Frank Zamora, May 2, 2018.
31. PropTester: "Conductivity Analysis, 30/80 mesh FracBlack HT," Report Number
400-18-04-15-09-J, May 7, 2018.
32. Personal communication with Ross Rucker, May 23, 2018.